Tumors in Domestic Animals (5th edition)

SAVE OFFLINE

Tumors in Domestic Animals

This book is dedicated to mentors. They taught us and built the foundation of veterinary pathology. We are all indebted to their hard work, and their willingness to teach and create new information.

Tumors in Domestic Animals Fifth Edition Edited by

Donald J. Meuten College of Veterinary Medicine North Carolina State University Raleigh, NC, USA

This edition first published 2017 © 2017 by John Wiley & Sons, Inc. First edition, 1961 © The Regents of the University of California Second edition, 1978 © The Regents of the University of California Third edition, revised and expanded, 1990 © The Regents of the University of California Fourth edition, 2002 © Iowa State Press Editorial Offices 1606 Golden Aspen Drive, Suites 103 and 104, Ames, Iowa 50010, USA The Atrium, Southern Gate, Chichester, West Sussex, P019 8SQ, UK 9600 Garsington Road, Oxford, OX4 2DQ, UK For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley‐blackwell. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Blackwell Publishing, provided that the base fee is paid directly to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license by CCC, a separate system of payments has been arranged. The fee codes for users of the Transactional Reporting Service are ISBN‐13: 978–0‐8138–2179–5/2017. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting a specific method, diagnosis, or treatment by health science practitioners for any particular patient. The publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. In view of ongoing research, equipment modifications, changes in ­governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. Readers should consult with a specialist where appropriate. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom. Library of Congress Cataloging‐in‐Publication Data Names: Meuten, Donald J., editor. Title: Tumors in domestic animals / edited by Donald J. Meuten. Description: Fifth edition. | Ames, Iowa : John Wiley & Sons Inc., 2017. | Includes bibliographical references and index. Identifiers: LCCN 2016007018 | ISBN 9780813821795 (cloth) | ISBN 9781119181194 (Adobe PDF) | ISBN 9781119181187 (epub) Subjects: LCSH: Tumors in animals. | Veterinary oncology. | MESH: Neoplasms–veterinary | Animals, Domestic Classification: LCC SF910.T8 M6 2017 | NLM SF 910.T8 | DDC 636.089/6992–dc23 LC record available at https://lccn.loc.gov/2016007018 A catalogue record for this book is available from the British Library. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Cover image: bottom row left – courtesy of Dr S. Shapiro Set in 9/11pt Minion Pro by SPi Global, Pondicherry, India

1 2017

Contents

List of Contributors, vi Preface, vii 1 An Overview of Molecular Cancer Pathogenesis, Prognosis,

and Diagnosis, 1 John M. Cullen and Matthew Breen

15 Tumors of the Urinary System, 632

Donald J. Meuten and Travis L.K. Meuten

16 Tumors of the Genital Systems, 689

Dalen W. Agnew and N. James MacLachlan

17 Tumors of the Mammary Gland, 723

Michael H. Goldschmidt, Laura Peña, and Valentina Zappulli

2 Trimming Tumors for Diagnosis and Prognosis, 27

18 Tumors of the Endocrine Glands, 766

3 Immunohistochemistry: Fundamentals and Applications

19 Tumors of the Nervous System, 834

Paul C. Stromberg and Donald J. Meuten in Oncology, 44 José A. Ramos‐Vara and Luke B. Borst

4 Epithelial and Melanocytic Tumors of the Skin, 88

Michael H. Goldschmidt and Kyle H. Goldschmidt

5 Mesenchymal Tumors of the Skin and Soft Tissues, 142

Mattie J. Hendrick

6 Mast Cell Tumors, 176

Matti Kiupel

7 Tumors of the Hemolymphatic System, 203

Victor E. Valli, Dorothee Bienzle, Donald J. Meuten, and Keith E. Linder

8 Canine and Feline Histiocytic Diseases, 322

Peter F. Moore

9 Tumors of Joints, 337

Linden E. Craig and Keith G. Thompson

10 Tumors of Bone, 356

Keith G. Thompson and Keren E. Dittmer

11 Tumors of Muscle, 425

Barry J. Cooper and Beth A. Valentine

12 Tumors of the Respiratory Tract, 467

Dennis W. Wilson

13 Tumors of the Alimentary Tract, 499

John S. Munday, Christiane V. Löhr, and Matti Kiupel

14 Tumors of the Liver and Gallbladder, 602

John M. Cullen

Thomas J. Rosol and Donald J. Meuten Robert J. Higgins, Andrew W. Bollen, Peter J. Dickinson, and Sílvia Sisó-Llonch

20 Tumors of the Eye, 892

Richard R. Dubielzig

21 Tumors of the Ear, 923

Bradley L. Njaa

Appendix: Diagnostic Schemes and Algorithms, 942 Introduction, 942 Mitotic count, 944 Canine melanomas and melanocytic neoplasms, 946 Histologic grading of canine cutaneous mast cell tumors, 948 Prognosis of canine cutaneous mast cell tumors, 949 Canine subcutaneous mast cell tumors, 951 Cytologic grading of canine cutaneous mast cell tumors, 952 Evaluation of regional lymph node metastasis in canine cutaneous mast cell tumors, 953 Canine oral perioral mast cell tumors, 955 Feline cutaneous mast cell tumors, 956 Canine soft tissue mesenchymal tumor (sarcoma), 957 Canine soft tissue mesenchymal tumor: Future?, 959 Joint tumors in dogs, 960 Lymphoma and lymphoid leukemia prognoses, 961 Enlarged lymph node evaluation in dogs, 965 Bone marrow evaluation, 966 PARR (PCR for antigen receptor rearrangement), 967 Canine and feline nasal tumors, 969 Scoring system and prognosis for canine lung tumors, 970 Histologic grading and prognosis for feline lung tumors, 971 Mammary, 972 Urothelial (transitional) cell carcinoma (UC), 973 Skin masses, 975 Canine breed predispositions for epidermal and melanocytic tumors, 976 Index, 979

v

List of Contributors

Dalen W. Agnew

Michael H. Goldschmidt

Laura Peña

Dorothee Bienzle

Mattie J. Hendrick

José A. Ramos‐Vara

Andrew W. Bollen

Robert J. Higgins

Thomas J. Rosol

Luke B. Borst

Matti Kiupel

Sílvia Sisó-Llonch

Matthew Breen

Keith E. Linder

Paul C. Stromberg

Barry J. Cooper

Christiane V. Löhr

Keith G. Thompson

Linden E. Craig

N. James MacLachlan

Beth A. Valentine

John M. Cullen

Donald J. Meuten

Victor E. Valli

Peter J. Dickinson

Travis L.K. Meuten

Dennis W. Wilson

Keren E. Dittmer

Peter F. Moore

Valentina Zappulli

Richard R. Dubielzig

John S. Munday

Kyle H. Goldschmidt

Bradley L. Njaa

Michigan State University, USA

University of Guelph, Ontario Veterinary College, Canada

University of California, San Francisco, USA

North Carolina State University, USA

North Carolina State University, USA

Cornell University, USA

University of Tennessee, USA

North Carolina State University, USA

University of California, Davis, USA

Massey University, New Zealand

University of Wisconsin‐Madison, USA

University of St. Thomas, USA

vi

University of Pennsylvania, USA

Marshfield Labs, Pennsylvania, USA

University of California, Davis, USA

Michigan State University, USA

North Carolina State University, USA

Oregon State University, USA

University of California, Davis, USA

North Carolina State University, USA

Colorado State University, USA

University of California, Davis, USA

Massey University, New Zealand

Kansas State University, USA

Complutense University of Madrid, Spain

Purdue University, USA

The Ohio State University, USA

BioMarin Pharmaceutical Inc, USA

The Ohio State University, USA

Massey University, New Zealand

Oregon State University, USA

University of Illinois, USA

University of California, Davis, USA

University of Padua, Italy

Preface

Dr. Jack Moulton would be very pleased that his book is in its fifth edition. He would welcome the return of color: it was present in the first edition in 1961. The quality of images between editions is not to be compared, but it is interesting and enlightening to compare old color and B&W images of the same tumors and leukemias with current images. The names of many tumors have changed and subtypes have emerged as we uncovered more about their true identity and behavior, at least with present day techniques. This will happen in every edition and is a credit to diagnosticians and investigators around the world. Past and present authors are part of the group that produced these discoveries. The present authors have summarized and detailed this body of information for pathologists, oncologists, molecular biologists, cancer biology investigators, veterinarians, and students of veterinary medicine. The authors have provided exquisite color images or created collages that pictorialize diagnostic features, pathogenesis, and techniques. Some images correlate radiographic, MRI, or CT studies with gross and histologic lesions, while others depict immunohistochemical, molecular, or cytological characteristics. Images provided by colleagues from around the world greatly improved this book, and their contributions are credited in legends. I hope readers note these contributors and realize how much the authors and I appreciate their willingness to share their expertise in pathology and photography. Images are an integral part of this book and pathology; they summarize and capture what words attempt to describe. Technology and its application are responsible for much of the new information in all disciplines. The authors have done an excellent job of blending core, basic pathological tenets that will remain constant with new information uncovered by evolving technology. There are new authors, new chapters, and a tremendous volume of new information that is cited, reviewed, and summarized in a reader‐friendly and practical manner. However, an evaluation of contents is the job of readers and reviewers, not of an editor proud of the book and friend of the contributors. I am proud of the contents because of the quality of work that authors so willingly gave for so little material reward. When you see them at meetings please thank them – it will mean more coming from you than me. The value of this book is due to their hard work in each chapter and throughout their productive careers. I am greatly indebted and thankful to them, as shall be our colleagues and patients in our clinics. The book remains focused on diagnostics, biologic behavior of animal cancers, and oncogenesis. An overview of all the tumors within each body system is provided via text, tables, and images. This feature is why books still have value in the age of on‐line everything. Authors also provided insights on how they differentiate tumors that appear histologically similar. Because of the desire to treat cancer in pets it is a reality that journal publications and a large component of this book are now focused on dogs and cats. Accurate morphological diagnoses and determination of prognostic parameters are required for the selection among numerous

treatment modalities currently available. Although accurate tumor diagnosis and classification often requires multiple techniques, histopathology and cytology remain the foundation of diagnostics. Authors have integrated the growing body of discovery with the practical components of our work. Providing prognostic information based on histological findings is an expected part of pathologists’ assessments but this text has not delved into treatment options as clinical oncology rewrites these approaches weekly. Often, clinicians value margin assessment as much or more than the diagnosis. There is a chapter dedicated to this topic and trimming biopsy specimens for histologic examination. Tumor cytology has been added and will continue to expand in subsequent editions. For many tumors, cytology is as accurate as histopathology, can be used at patient side with minimal invasiveness, and can yield specimens for molecular studies. Grading schemes based on cytology and molecular profiles will be included in future editions. Cytological evaluation can replace biopsy for some osseous tumors, and if treatments beyond excision or palliative measures are not a consideration, then cytology can provide a final diagnosis for many solid tumors and most leukemias. Reference texts in the tissue and cytological diagnostic arenas are largely divorced, and integration of these can benefit morphological assessments of cancer disease. Furthermore, cytological assessment permits rapid turnaround time and will become same day when representative cytological images are sent from patient side to pathologists. Immunohistochemistry is now a routine part of our diagnostic evaluation and a chapter is dedicated to this topic. Authors emphasize it is only one part of the diagnostic puzzle. IHC, PARR, molecular signatures, and other ancillary tests must be integrated with all the other data available; they are not stand-alone tests. We need these tests most when the interpretation of light microscopy is not straightforward, yet that is when ancillary tests may also be ambiguous. We should convince clinicians of this and we need standardization of methods. The tsunami of molecular diagnostics has not yet washed away H&E. Molecular tools to create signatures that differentiate tumors, detect cancers at the earliest possible intervention, elucidate oncogenesis, or become independent prognosticators are exciting developments. In addition to the molecular characteristics of tumors, the next generation of prognostic tools will look at the host’s ability to respond to cancers. It is likely that the host’s response will be as or more predictive of biologic behavior for many tumors than is assessment of margins, mitotic counts or immunohistochemical profiles. It will be terrific when tests identify whether a patient is likely to develop cancer, which cancer is likely, which treatment is best, and how innovative techniques (e.g., CRISPR) can be directed to treat the cancer. Cost accountability will always be a component of veterinary diagnostics and care. I doubt reviewers will find much controversy in this book. There may be some diagnoses or a pathogenesis that others may not accept, but not many. If anything is controversial it may be in the

vii

viii   Preface

appendices, where I tried to summarize information so it can be easily found and used. However, without standards for the diagnostic parameters to predict behavior or standardization of the techniques used, it is difficult to compare studies or merge data from different reports. Users are encouraged to read the original references for the many excellent details they provide. We need standardization of diagnoses, techniques, and our follow-up assessments in veterinary oncology and oncologic ­ pathology. Some would say this book is the standard with which to diagnose cancers in domestic animals. However, the techniques used to provide diagnoses need standardization, such as antibodies, primers, margins, areas in which mitotic figures are counted, and flow cytometry. Without standards for the diagnosis and techniques employed we cannot reliably compare results between studies and contradictions of prevalence, biological behavior, and diagnosis will continue. Standardization of outcome assessments is imperative, yet it may be the weakest link in the connection of h ­istologic, cytologic, and molecular parameters with prognosis. Documentation of tumor recurrence with histopathological confirmation and autopsy data is essential to assess a tumor’s biological behavior. Unfortunately the number of published case series reported with autopsy and histopathology is abysmally small. If veterinarians want to rely on information from investigative studies to help owners and their pets, they need to help collect accurate follow-up data. We need to train our new veterinarians that, in addition to being competent surgeons, clinicians, and caregivers, they have a responsibility to serve the profession as clinical scientists too.

All of us are indebted to mentors. They taught us, they nurtured our passion for pathology and were role models for each new generation of veterinary pathologists. The authors and readers of this book double‐scoped and learned from some of the founders of veterinary pathology, and I dedicate this book to mentors. Many are icons in our disciplines but many of us benefitted enormously from hard‐working and committed teachers who were not so widely known, but were essential to us. If I listed names I would forget some and offend others so please take a moment to think about those individuals who shaped you and your career. We all remain responsible for passing on this mantle. We have not defeated cancer. That stated goal from many years ago did not acknowledge the complex biology of cancer. No one could have predicted the information about cancer would grow to the enormity it has. Despite all the time, money and great minds that have investigated cancer it remains a leading cause of death in animals and people. The discipline of oncology will expand in human and veterinary medicine and professionals from both disciplines should work together to understand how cancers develop and how to better fight them. Veterinarians, physicians, and researchers need a book like the one Dr. Moulton envisioned. This book would not be a reality without Wiley and the executive editor, Erica Judisch, and the freelance project manager Nik Prowse. An unseen person who helped throughout was Laura Cullins; she solved the many problems that arose and made my job easier. Many wanted this edition published sooner, none more so than my wife; thank you Nicki. Don Meuten

1

An Overview of Molecular Cancer Pathogenesis, Prognosis, and Diagnosis John M. Cullen and Matthew Breen North Carolina State University, USA

Fundamentals of cancer biology

Cancer is a disease of the genome, arising from DNA alterations that dysregulate gene structure or function.1,2 Damage to the ­cellular genome or altered expression of genes is a common feature for v­ irtually all neoplasms.3 Given that there is an inherent error rate in DNA replication, all multicellular organisms face the near certainty of developing a neoplasm if they survive long enough, as essential mutations for neoplastic transformation will eventually develop. Many mutations may be inconsequential, but cancer can develop when nonlethal mutations occur in a small subset of the coding and noncoding regions of the genome, perhaps affecting even just a few of the ~20,000 genes thought to comprise the mammalian genome.4 There are many agents, in addition to deficiencies in DNA replication fidelity and error repair, that drive tumor formation. These agents include viruses, mutagenic chemicals, and radiation. Unraveling the pathogenesis of cancer has not only helped us understand how a cell transforms into a tumor but has also ­promoted molecular tests that now help diagnose and provide prognoses for a variety of cancers in humans and animals. Genetic injury Genetic damage is a universal component of the pathogenesis of neoplasia, with somatic mutations in genes identified in 90% of cases, and germ line mutations identified in 20% of human ­neoplasia and both features found in a small percentage of neoplasms.1,2,4 In some cases a single base pair mutation is sufficient to

alter a critical amino acid, leading to altered protein function and an increased risk for neoplastic transformation. Other types of mutations involve insertions, deletions, or duplications of gene segments. Structural chromosomal changes, such as translocations, which lead to chimeric transcripts or deregulation of gene expression through movement of promoters and enhancer regions adjacent to relevant genes can also drive malignant transformation. In addition, gene copy number increases or decreases (gene dosage) can also occur. Epigenetics DNA sequence mutations are not the only route to neoplasia.5 Epigenetic mechanisms regulate gene expression without causing structural changes to the genome and also play a role in malignant transformation.6 Epigenetic changes are reversible, heritable alterations of gene expression without mutation of the genome. Three main forms of epigenetic gene regulation include DNA methylation, histone acetylation, and microRNA expression. Gene expression can be silenced, diminished, or increased by altering methylation patterns in the DNA. Aberrant methylation patterns, such as hypermethylation and hypomethylation, are common in a variety of neoplasms and are linked to abnormal gene expression levels. In particular, methylation of tumor suppressor genes leading to their suppression is recognized in a number of human cancers, including breast, colon, and renal carcinomas. Histone proteins serve as spools that are wound with DNA strands to package cellular DNA into nucleosomes, which when

Tumors in Domestic Animals, Fifth Edition. Edited by Donald J. Meuten. © 2017 John Wiley & Sons, Inc. Published 2017 by John Wiley & Sons, Inc.

1

2    Tumors in Domestic Animals

compacted constitute the main components of chromatin. Gene expression can be altered by loosening or tightening the binding of the DNA strands to the histone proteins. Tightly wound DNA is either not transcribed or transcribed at lower levels than when it is more loosely associated with the histone proteins. Histone proteins that have been altered, often by acetylation, have a more relaxed binding pattern with their associated DNA and this facilitates gene expression. MicroRNAs (miRNAs) are small, nontranscribed RNA molecules, approximately 22 nucleotides in length, that contribute to a complex termed the RNA‐silencing complex, which binds to  specific sequences in messenger RNA strands and directs gene silencing.7 MiRNAs may regulate up to 30% of genes via posttranscriptional control. Amplifications and deletions of miRNAs are common in various human cancers and more work  is needed to evaluate animal neoplasms.8 Consequently, miRNAs can participate in tumor formation, as increased ­ expression of miRNAs that target tumor suppressor genes leads to an increased tumor risk, as does decreased expression of miRNAs that target oncogenes. Although none of these mechanisms alter the structure of the ­cellular genome they can significantly alter gene expression and have all been shown to be involved in neoplastic transformation.8 A  combination of mutation and epigenetic mechanisms is involved in the multistep process that leads to the emergence of a population of cells with a malignant phenotype known as malignant transformation. Epigenetic perturbations may offer an explanation for some types of tumors associated with chronic inflammation or the presence of foreign material but no clear pattern of mutation, such as the fibrosarcomas and osteosarcomas that can develop in dogs infected with Spirocirca lupi or rare reports of sarcomas forming following metallic implants. Injection site sarcomas in cats seem to be associated with the inflammation induced by the vaccine adjuvant, although a subset of the affected cats have mutations in the tumor suppressor gene p53.9–11 Tumor heterogeneity, tumor progression, and clonal evolution There are many pathways that can lead an individual cell to the malignant phenotype, but they all involve multiple genetic and ­epigenetic alterations. Tumor progression is a process by which cells that have developed a neoplastic phenotype acquire more characteristics that lead to malignancy and metastasis. Tumor growth starts clonally with a single cell that has undergone neoplastic transformation and the incipient tumor develops by clonal expansion of that cell. When a population of cells is identified as clonal it is a strong indicator that the population is neoplastic. However, this is not universally true, as clonal lymphocyte populations can be identified in some inflammatory conditions such as cases of feline infectious peritonitis and infections with Ehrlichia sp. Initially, all cells in the neoplastic mass are identical, but due to the genetic instability the tumor cells acquire genetic and epigenetic changes that give rise to tumor heterogeneity (Figure 1.1).12 In some instances this genetic change can be dramatic and sudden, as seen in chromothripsis, a phenomenon in which hundreds to thousands of genetic rearrangements can occur in one or a few chromosomes during a single event. Processes frequently affected include various types of DNA repair, telomere maintenance, DNA replication, and chromosome segregation. Some genetic changes are lethal to the

affected cells, whereas others confer new functions and phenotypes, giving inherent growth advantages. Over time, the developing tumor mass becomes composed of a heterogeneous cell population and the tumor accumulates characteristics that make them more dangerous to the host.12 As a neoplastic cell replicates, subclones emerge that are more locally aggressive, more likely to metastasize, and less responsive to therapy. Tumor progression has been attributed to a greater level of genomic instability in affected cells, which explains why early detection of the neoplasm is associated with improved prognosis. However, early detection is challenging in most clinical settings. By the time most malignant neoplasms are detected, using contemporary imaging methods, they most likely comprise a heterogeneous cell population, the neoplasms having completed the greater part of their growth. A single transformed cell must undergo at least 30 doublings to form a 1 g mass, an approximate cut‐off for clinical detection, but only approximately 10 more doublings are needed to form a 1 kg mass (Figure  1.2). Since a 1 kg mass is regarded as a lethal tumor burden for a human, it is likely that fewer doublings are needed to form lethal cancers in small domestic animals. Cancer is a multistep process and in some types of epithelial cancers there is a histologic phenotype that is characteristic of the different steps, including hyperplasia, dysplasia, and adenoma and carcinoma formation (Figure 1.3). Progressive accumulation of various mutations and epigenetic disturbances accompany these different histologic phenotypes. However, in some circumstances activation of an oncogene in an otherwise normal cell can lead to cell senescence and inhibit tumor formation.13 Paradoxically, expression of an activated oncogene can lead to an exit from the cell cycle and termination of cell growth. Oncogene‐induced senescence is consequently considered an authentic tumor suppressor mechanism in vivo. Ultimately, genetic and epigenetic alterations lead to a common pattern of features, or hallmarks, that distinguish neoplastic cell populations from normal cells. These hallmarks of cancer were initially proposed in 200014 and an updated review has been recently published.15 Each of these hallmarks and their relevance regarding animal carcinogenesis, diagnosis, prognosis and therapy will be discussed in this chapter.

The hallmarks of malignancy

The six key elements of malignancies are shown in Figure 1.4. Sustaining proliferative signaling: ­proto‐oncogenes and oncogenes Normal tissues are often capable of responding to injury or tissue loss by proliferation. Proliferation is driven by growth factors that bind specific cellular receptors, often tyrosine kinases, causing them to become activated and propagating a cascade of intracellular signals that culminate in mitosis. However, proliferation is ­controlled and limited, retaining normal structure and function. In cancer, proliferation is persistent and unregulated. Cell proliferation and maturation are regulated by a subset of cellular genes. Proto‐oncogenes are normal genes that encode ­ ­proteins participating in one or more signal transduction pathways associated with important regulatory pathways.16 Because of their central role in the life cycle of the cell, proto‐oncogenes have been conserved throughout evolution and their DNA sequences vary little from yeast to humans. Disturbances in gene structure or expression can alter the cellular function of a proto‐oncogene,

An Overview of Molecular Cancer    3

Neoplastic transformation Clonal expansion

Normal cell

Tumor cell

Proliferation of genetically unstable cells

Variants

Progression

Heterogeneous cell population

Metastatic Require fewer growth factors Lethal mutations Invasive Non antigenic

Metastasis

Tumor cell heterogeneity. Although tumors arise from a single cell, the inherent genetic instability in tumor cells gives rise to additional mutaFigure 1.1  tions and a heterogeneous population of cells with different genetic characteristics. Some mutations are lethal to developing cell lines and they die, but other mutations provide various features that facilitate the emergence of viable cell lines which may contain malignant characteristics including the ability to metastasize.

causing it to stimulate tumor formation. Once this alteration has occurred, the proto‐oncogene is termed an oncogene. More than 100 oncogenes have been identified, and their numbers increase with continued genetic analyses of neoplasms. Typically the genes are referred to using a three‐letter nomenclature. Many oncogenes were initially identified as part of the genome of retroviruses that caused cancer and they were named for the virus from which they were originally identified. For example, the proto‐oncogene myc was originally isolated from the avian myelocytomatosis virus, and ERB A and ERB B were isolated from avian erythroblastosis virus. It should be remembered that it is not the oncogene, but the encoded protein that leads to cell transformation. The proteins encoded by oncogenes are referred to as oncoproteins.

Classification of oncogenes

Oncogenes can be grouped into five categories based on the types of oncoproteins they encode. These categories include growth factors, growth factor receptors, intracellular signal transducers, nuclear regulatory proteins (transcription factors), and cyclins. The proto‐ oncogene sis encodes the beta chain of platelet‐derived growth factor (PDGF). When fibroblasts are infected with simian sarcoma virus, a retrovirus that contains the oncogene sis, there is an excess of sis oncoprotein produced. This protein leads to overstimulation of PDGF receptors on the cell surface in an autocrine fashion and can drive fibroblasts towards malignant transformation. In this circumstance the oncoprotein has a normal amino acid sequence but is produced in an abnormal, deregulated amount. Mutant forms of

4    Tumors in Domestic Animals

Transformation Normal cell

Transformation via chemicals, viruses, radiation or inherited predisposition

Single tumor cell

Clinically undetectable

30 doublings

Earliest detectable mass ~1 g (109 cells)

~10 doublings

~up to 1kg (1012 cells)

Lethal tumor burden Figure 1.2 Tumor doubling. Tumor growth starts with a single cell that

expands clonally. It takes approximately 30 doublings to form a 1 g mass, at which time most lesions can be detected clinically. Only 10 more doublings are needed to form a 1 kg mass, considered to be a lethal burden in humans. Likely, a smaller mass would be lethal in dogs or cats.

growth factors also occur, and they may inappropriately stimulate receptors by binding to them in an abnormal fashion. Oncogenes may encode growth factor receptors. A typical growth factor receptor has three components: an extracellular growth factor binding domain, a transmembranous segment, and a cytoplasmic domain with kinase activity. Oncogene‐encoded growth factor receptors, such as ERB B, are often truncated into a form that no longer has the extracellular receptor portion of the normal ­protein. These abnormal receptors do not require growth factor binding to be stimulated and are constitutively activated. The intracellular signal transducers are located in the cytosol (e.g., ABL, RAF) or are membrane associated (e.g., RAS, SRC).

Typically, these molecules are enzymes in the tyrosine kinase family. Point mutations or more gross structural alterations can constitutively activate these proteins, producing a level of activity that in turn leads to uncontrolled cell proliferation. Tyrosine kinase receptor activity is abnormal in several animal cancers, including mammary carcinomas and mast cell tumors, and several inhibitors of tyrosine kinase activity are being investigated, or are currently in use, as therapy.17 Transcription factors are nuclear proteins that regulate gene expression. They bind to selected sites on DNA in a complex with other proteins to facilitate gene expression. The oncoproteins encoded by MYC, JUN, and FOS are transcription factors that stimulate expression of genes necessary for cell division. Abnormal levels of expression, or mutations that alter the function of these proteins, can compromise growth control. Cyclins are a series of proteins that precisely regulate movement through the cell cycle.18 Individual cyclins are expressed for brief intervals at appropriate points in the cell cycle. The cyclins interact with and activate enzymes termed cyclin‐dependent kinases (CDKs). The CDKs, in turn, activate proteins that are essential for progression through the cell cycle. Disruption in the function of cyclins leads to dysregulated control of cell replication. Several types of tumors in humans have been described with mutations in the genes that encode cyclins or CDKs.19 Abnormal cyclin expression has been documented in canine and feline neoplasms, including cyclin A in mammary carcinomas and cyclin D1 in mammary tumors, squamous cell carcinomas, and, to a lesser extent, basal cell tumors.20,21

Alterations of gene structure (function)

Proto‐oncogenes can be transformed into oncogenes following damage to their structure. Structural alterations can occur by mutation of individual nucleotides or alterations that may occur during changes to karyotype organization, such as chromosomal translocation events. Damage to individual nucleotides (i.e., point mutation) is the most common mutation sustained by proto‐oncogenes. Chemical carcinogens and some forms of radiation exert their influence this way. Mutation of a single nucleotide can lead to the incorporation of a novel amino acid into a protein, and, if appropriately localized, the activity of the protein can be profoundly altered. One of the better characterized signal transduction pathways affected by mutation involves the RAS (derived from rat sarcoma virus) signaling pathway (Figure 1.5). All mammalian cells express three related RAS proteins, designated K‐RAS, N‐RAS, and H‐RAS. Each of these proteins has a similar function, acting as an enzyme (GTPase) that phosphorylates GTP and acts as a switch regulating cell proliferation and survival. Any RAS family member can drive tumor development when they are mutated in specific codons. Signaling via RAS begins when growth factors bind to specific cell surface receptors. This induces the receptors to dimerize, autophosphorylate, and undergo a conformational change. As a result of the conformational change, the receptors can interact with an associated bridging protein complex, which in turn transfers activation to the RAS protein located on the cytoplasmic surface of the cell membrane. Normally, the RAS protein is inactive and is bound to guanine diphosphate (GDP). When the RAS protein is stimulated it exchanges GDP for guanine triphosphate (GTP) and becomes activated. RAS protein is negatively regulated by GTPase‐activating protein (GAP), a protein that enhances the hydrolysis of RAS‐bound GTP to GDP. Activated RAS attracts a serine/threonine kinase, termed RAF (derived from rapidly accelerated fibrosarcoma), to the

An Overview of Molecular Cancer    5

Oncogene-induced senescence

Normal mucosa

Hyperproliferation

Adenoma

Carcinoma Benign

Malignant

Figure 1.3  Histologic evolution of a carcinoma. The cellular development of cancer is a multistep process in most cases. There are several phenotypic steps in the evolution of colonic carcinoma including areas of hyperproliferation/hyperplasia, then dysplasia, followed by adenoma and, in a subset of these, carcinomas. The distribution of different phenotypes is not uniform throughout individual lesions and regions with different phenotypes may be seen when a lesion is sampled. Spontaneous growth arrest or resolution of tumors may occur, as overexpression of oncogenes can drive cellular senescence in some circumstances.

Growth factor

Proliferative signaling sustained

Growth suppressors evaded Invasion and metastasis activated

Cell death resisted Angiogenesis induced

Replicative immortality enabled

Figure 1.4 Hallmarks of cancer. These features are key elements of malignancies.

inner aspect of the cell membrane, where RAF is phosphorylated by membrane‐associated kinases. Activated RAF in turn phosphorylates mitogen‐activated protein (MAP) kinases, and these kinases migrate to the nucleus, where they stimulate the synthesis of nuclear transcription factors, such as MYC. These transcription factors stimulate the expression of genes that cause resting cells either to enter the cell cycle and divide or to alter their differentiation or synthesis patterns. Control of cell signaling is fine‐tuned by the balance of a matrix of stimulatory and inhibitory influences. Consequently, cell proliferation can be driven not only by stimulatory events, but also by the disruption of inhibitory pathways. The RAS gene offers a

P

RAS GDP Inactive

RAS GAP

GTP

Active

Activation of transcription

Figure 1.5 RAS oncogene. An example of proto‐oncogene activation is shown in this overview diagram of the typical RAS signaling cascade. When a growth factor binds to its transmembrane receptor the receptor becomes activated. Receptor binding triggers activation of RAS via a bridging protein. Inactive RAS, which is bound to guanosine diphosphate (GDP), becomes activated via an exchange (red arrow) for guanosine triphosphate (GTP). Activated RAS acts through intermediary proteins to activate mitogen‐activated protein kinases (MAP kinases) that lead to altered nuclear signal transduction and cell mitosis. In normal cells GTPase‐activating protein (GAP) stimulates dephosphorylation of activated RAS to an inactive form that curtails signaling (blue arrow). Mutant RAS does not interact with GAP normally and consequently stimulates cell proliferation in an unchecked fashion.

good example of this abnormality as well, as mutations in RAS typically impair the ability of the GAP protein to dephosphorylate RAS, causing it to remain in the active state, leading to a constitutive activation of RAS driving cell proliferation.

6    Tumors in Domestic Animals

Mutations in RAS are common in human tumors. In some s­ urveys 16–30% of all tumors are reported to bear RAS mutations.22 Specific sites of human tumors bearing RAS mutations in 20% or more cases include the biliary tree, large intestine, small intestine, pancreas, and the skin. Evaluation of specific RAS isoforms reveals that they selectively appear in different tumor types.23 Specific isoforms can be quite common in individual tumor types. Mutations in K‐RAS are found in 22% of all human tumors with an incidence of 8% for N‐RAS and 3% for H‐RAS. N‐RAS is mutated in human melanoma, acute myelogenous leukemia, thyroid neoplasia, and multiple myeloma. K‐RAS, involved in pancreatic, colorectal, ­thyroid, and lung carcinomas and acute myelogenous leukemia among others, is the most frequently mutated RAS gene in human neoplasms. Approximately 61–90% of human pancreatic carcinomas contain a mutation in K‐RAS. Mutations in RAS are not nearly as well documented in animals as humans, as most veterinary studies are limited by small sample sizes, but RAS mutations are reported to be less frequent in a variety of sarcomas of dogs and cats than they are in humans.24–32 RAS mutations have been noted in several tumor types, including K‐RAS mutations in approximately 15–25% canine lung tumors of different types33–35 and N‐RAS mutations in canine leukemia, where up to 25% of cases of acute myeloid leukemia or lymphoid leukemia have mutations.36,37 Altered RAS signaling is one pathway among several alterations in signal transduction that lead to cellular hyperproliferation. The ability of malignant cells to proliferate in a sustained fashion can also derive from the impact of an excess of growth factors. Mitogenic signals can be generated in an autocrine fashion in circumstances where malignant cells release growth factors that bind their own receptors and initiate signaling that leads to proliferation. In a more complicated paracrine fashion, malignant cells can signal nearby stromal cells, causing them to release mitogenic factors that in turn stimulate the tumor cells. In some cases, mutations lead to multiple copies of growth factor receptors on an individual cell, leading to excessive intracellular signaling in the face of a normal amount of growth factors. Alternatively, malignant cells may become independent of growth factors when they have mutated growth factor receptors or signal transducers such as RAF that are constitutively activated stimulating the downstream signaling cascade. Another more clinically relevant example of abnormal proliferation caused by a mutation is found in canine and feline mast cell tumors. The proto‐oncogene c‐KIT (also referred to as CD117 and stem cell factor receptor) encodes the transmembrane tyrosine kinase receptor KIT. Following binding of its ligand, stem cell factor, the resulting signal transduction is involved in survival, maturation, migration, and proliferation of mast cells and other hematopoietic cells. Mutations in the c‐KIT gene, in particular small internal tandem repeats in exons 11 and 12, lead to an abnormal receptor that does need ligand binding as a prerequisite for activation, and consequently it constitutively stimulates signal transduction.38 There are several other mutations that also can activate c‐KIT.39 Abnormal localization of the gene product in the cytoplasm in neoplastic mast cells is also associated with increased proliferation of the affected cells and likely a poorer prognosis for dogs bearing these genetic alterations than those without such changes.40 The presence of internal tandem repeats (ITRs) can be detected by a PCR‐based assay, which identifies neoplastic mast cells in approximately 30% of malignant canine cutaneous mast cell tumors.41 Identification of the c‐KIT ITR mutations can be used for diagnostic purposes, to assess prognosis and to monitor response to

therapy. Not all malignant mast cell tumors bear this mutation so its role is not entirely clear. However, targeting the tyrosine kinase activity of c‐KIT is recognized as a useful approach in a variety of human and veterinary applications. There are new chemotherapeutic agents that inhibit the tyrosine kinase activity of c‐KIT now available in veterinary medicine, Palladia® (Zoetis) and Kinavet® (AB Science) that are used to treat cutaneous mast cell tumors in dogs as well as other malignancies with aberrant c‐KIT activity.

Chromosomal translocation

Chromosome translocation results in the movement of one chromosome to another chromosome, or exchange of segments between different chromosomes in reciprocal translocation events. This process can deregulate transcription by bringing in close ­juxtaposition active cellular promoters and proto‐oncogenes. One example occurs in both humans and mice: the proto‐oncogene MYC is overexpressed in lymphomas of B‐cell lineage due to translocation of an active cellular promoter from the immunoglobulin gene to another chromosome that contains MYC. In some circumstances the functions of proto‐oncogenes are altered by chromosome translocation. A well‐characterized example of this process occurs in the distinctive translocation that produces the Philadelphia chromosome found in up to 95% of cases of human chronic myelogenous leukemia (CML), but it is not specific, as it is found in approximately 25% of cases of human acute lymphoblastic leukemia and rarely in acute myelogenous leukemia.42 This rearrangement involves an exchange of chromosomal segments between the distal ends of human chromosomes 9q and 22q, resulting in a derivative chromosome 22 in which a fragment of the proto‐­ oncogene (abl) on human chromosome 9 becomes juxtaposed to the breakpoint cluster region (bcr) on human chromosome 22. This fusion yields an abnormal hybrid gene that encodes a chimeric messenger RNA containing information from both genes. When the message is translated, a hybrid protein, termed a fusion protein, results. In this circumstance the fusion protein is an active oncoprotein that results in elevated tyrosine kinase activity, which is crucial to its oncogenic potential. To halt progression of the leukemia the BCR‐ABL kinase antagonist imatinib mesylate (Gleevec®) is used in therapy. As a competitive inhibitor of the tyrosine kinase activity, the drug serves to block the proliferative signal given by the BCR‐ABL protein, preventing the formation of new abnormal cells. While this effect does not apply to all patients with CML it is sufficiently effective that Gleevec therapy is now considered standard of care for patients with the Philadelphia chromosome. Although CML is rare in domestic dogs, numerous cases have been shown to present with evolutionarily conserved cytogenetic change (structural chromosomal changes) resembling the Philadelphia translocation in human cases. In the canine aberration, termed the “Raleigh” chromosome, the canine genes BCR (dog chromosome 9) and ABL (dog chromosome 26) are juxtaposed and produce a fusion protein43,44 (Figure  1.6). These data suggest that treatment with Gleevec or other tyrosine kinase inhibitors could be an option for therapy of BCR‐ABL‐positive canine cases. The Raleigh chromosome has since been identified in additional canine leukemias, including chronic monocytic and acute myeloblastic cases.43,45 More study will be needed to determine the frequency of this translocation in canine CML and other leukemias. Cytogenetic analysis has been used to monitor response to therapy and revealed a marked reduction of circulating neoplastic cells with the translocation following therapy (vincristine and prednisolone) in one case of canine chronic monomyelocytic leukemia.46

An Overview of Molecular Cancer    7

Human chromosomes: Philadelphia translocation/chromosome der 9

HSA 9

Ph1 HSA 22 22q12 22q13

9q33

bcr locus

22q12 22q13

bcr locus

9q33

c-abl locus

BCR-ABL hybrid gene

9q34

Tyrosine kinase

c-abl locus

9q34

A

Canine chromosome composition HSA 17 HSA 12 HSA 10

HSA 9

CFA 9

CFA 26 HSA 22 q genes

c-abl locus

9q33 9q34

22q12 22q13

bcr locus

9q34 9q33

B Canine chromosomes: Raleigh translocation/chromosome der 9

CFA 9

Ph1

CFA 26 bcr locus 9q33 22q12 22q13

c-abl locus

BCR-ABL hybrid gene

bcr locus

Tyrosine kinase 9q33

c-abl locus

C Figure 1.6  Raleigh chromosome. In human leukemias, a characteristic chromosome is the Philadelphia chromosome (Ph). This derivative chromosome, also referred to as the Philadelphia translocation, is the result of reciprocal translocation between human chromosomes 9 and 22, bringing the genes BCR and ABL (panel A) together to create activation of the tyrosine kinase of c‐ABL. The evolutionarily conserved translocation (panel B) has been detected in canine leukemias, the result of a reciprocal translocation between regions of dog chromosomes 9 and 26 (shown in panel C). The canine event is referred to as the Raleigh chromosome and has been detected in chronic myelogenous leukemia and chronic myelomonocytic leukemia. Within these patients, the frequency of cells with the Raleigh chromosome has been shown to decrease in response to tyrosine kinase inhibitor treatment, indicating that its presence may be used to monitor cytogenetic remission.

8    Tumors in Domestic Animals

Alterations of gene expression

Gene expression can be altered via gene amplification or deletion, promoter insertion, gene translocation, and regulatory miRNA. Each of these genetic mechanisms can lead to the deregulated ­synthesis of normal (i.e., wild type) proto‐oncogene proteins. Given that many proteins encoded by proto‐oncogenes function to stimulate cell proliferation, it is obvious that their overexpression would have the potential to lead to cancer formation. For reasons that are not well understood, tumor cells often ­sustain excessive rounds of localized DNA replication that can result in the formation of multiple copies (hence the term gene amplification) of the same gene or genes. The duplicated genes (or amplicon) may be found in small chromosome‐like structures termed double minutes, or may form concatenated (i.e., like beads on a string) structures within a chromosome that can be identified as homogeneously staining regions (HSRs). HSRs are portions of chromosomes that lack the characteristic banding pattern found in normal chromosomes. In general, gene amplification leads to the overproduction of the products encoded by the genes within the amplicon, increasing the potential for neoplastic transformation. MicroRNAs An additional and relatively new mechanism for influencing gene expression involves miRNA.47 There are more than 1000 types of miRNA expressed in essentially all cells. Primarily they bind to message RNA and promote degradation of messenger RNA, thereby preventing translation and influencing gene expression. Specific miRNAs are associated with some human neoplasms, particularly colorectal cancer and chronic lymphocytic leukemia, but more work is needed to evaluate the role of miRNAs in neoplasia of veterinary interest. Oncogenic viruses In veterinary medicine there are several important oncogenic retroviruses. These include feline leukemia virus, bovine leukemia virus, and avian leukosis virus. When certain oncogenic retroviruses, known as nonacute retroviruses, insert their genome into cellular DNA, the regulatory elements normally controlling viral gene expression also affect the expression of nearby cellular host genes. Viruses and cells have two major types of these regulatory elements, enhancers and promoters. Both elements stimulate gene expression, but differ in their functional attributes. Promoters stimulate adjacent genes but must be properly oriented (upstream of the gene) to facilitate expression. Enhancers stimulate promoter activity, but unlike promoters, their capacity to stimulate transcription is orientation independent. Since, in general viral promoters and enhancers are more potent than their cellular counterparts, they can significantly increase and thus dysregulate cellular gene expression. When a retrovirus integrates within a region of genomic DNA flanking a proto‐oncogene, transcription of the proto‐oncogene can be deregulated, leading to cell transformation. In most ­circumstances, viral insertion events affect the regulation of gene expression, not the function of the gene or genes affected. There is also a second type of oncogenic retrovirus, called acute transforming retroviruses, that are typically replication defective, but carry an oncogene derived from a host’s proto‐oncogene in their genome and rapidly transform infected cells. Feline sarcoma virus is an example of this type of virus. Oncogenic DNA viruses generally differ from oncogenic retroviruses in that they contain authentic viral genes that encode oncoproteins capable of transforming infected cells. These viral proteins

often act by interfering with the proteins encoded by tumor suppressor genes. Bovine papillomavirus and several primate herpes viruses are examples of oncogenic DNA viruses of ­veterinary importance. Tumor proliferation The phenotypic manifestation of the first hallmark of cancer (Sustaining Proliferative Signaling) is typically recognized as an increase in mitotic rate. Many tumor types have an increased rate of mitosis, enumerated by counting the number of mitotic figures observed in a specific number of high‐power (40 × objective) microscopic fields (the mitotic count in 2.37 mm2; see pp. 944–945) although there are several other immunohistochemical methods to assess cell proliferation, including staining for proliferating cell nuclear antigen (PCNA) or Ki67 that may be more readily interpreted than counting mitoses. The presence of an increased number of mitoses is considered in the diagnosis of benign or malignant forms of various tumors. For example, the mitotic count of soft tissue sarcoma in dogs is currently used as one of the principal criteria in the assessment of malignancy and to predict the likelihood of recurrence and or metastasis. However, increased mitoses alone are not necessarily an indication of malignancy since the tissue of origin requires consideration. For example, canine histocytomas are characterized by a high mitotic count yet these neoplasms typically regress spontaneously. A similar observation can be made for transmissible venereal tumors of dogs. There also are examples of tumors with a low proliferation index that behave aggressively, such as maxillary fibrosarcoma in dogs. In some cases, a high mitotic count may not correlate with the cellular proliferation of a mass. This occurs when there is arrest of mitosis, leaving elevated numbers of mitotic ­figures at a given point in time, but in the absence of completed cell ­division, no increase or a limited increase in cell population. Evading growth suppressors A second hallmark of cancer involves the ability to bypass potent growth inhibitory signaling.48 The major agents of growth inhibition are a group of 25 or more tumor suppressor genes. Tumor suppressor genes play a critical role in the control of normal cell growth. They serve as the “brakes” to cell replication. When tumor suppressor genes are inactivated, cells lose regulatory control of cell proliferation. A single, intact, functional copy of a tumor suppressor gene is sufficient to maintain control of cell proliferation. When both alleles are lost or damaged the affected cell has a high risk of neoplastic transformation. The discovery of tumor suppressor genes arose from the study of certain human families that presented with a significantly increased incidence of specific tumor types. Genetic analysis of these “cancer families” revealed that some family members were born with one mutated allele of a critical gene, and when a second mutation in the functional allele occurred spontaneously the affected individual was at a very high risk to develop neoplasia.9 The first tumor suppressor gene to be discovered this way was the retinoblastoma or RB gene. Loss of both alleles led to the development of retinoblastomas in affected children. Loss of function of both alleles of another tumor suppressor gene, TP53, was identified in other kindreds and termed Li–Fraumeni syndrome. These individuals are at elevated risk for a variety of mesenchymal neoplasms, but mutated TP53 is frequently identified in many human malignancies.49,50 At least one heritable cancer syndrome (renal carcinoma and nodular dermatofibrosis, or RCND, of German shepherd dogs) has been described in dogs with an autosomal dominant inheritance.51,52

An Overview of Molecular Cancer    9

The heritable factor for this syndrome maps to dog chromosome 5 (CFA 5), and specifically to the tumor suppressor gene folliculin (FLCN, previously BDH gene). This region in the dog chromosome overlaps a corresponding region in the human chromosome that was recently described as the heritable factor for the corresponding human disease (Birt–Hogg–Dubé syndrome). Inactivation of this tumor suppressor gene is critical to the development of this syndrome. It is probable that other comparable syndromes to those that are described in humans will eventually be identified in companion and laboratory animals, but it is unlikely these will account for more than 5–10% of all cancers in animal cases. There are a few examples in dogs in which a specific cancer (melanoma, histiocytic sarcoma) is associated with the absence or decrease of tumor suppressor gene(s). It is likely there are inherited susceptibilities to cancer in many breeds. Examples are lymphoma in the golden retriever53 and parathyroid neoplasia in keeshonds.54 Ongoing cytogenetic studies are being conducted to assess the ­genetic basis of breed‐related alterations in tumor risk. To understand the relevance of tumor suppressor gene inactivation in tumorigenesis, a brief review of the normal cell cycle and how it differs from that in neoplastic cells is warranted. The cell cycle ­consists of a series of biochemically distinct temporal periods that prepare the cell for division.55 Following mitosis, a cell may withdraw from the cell cycle and enter a quiescent stage (G0 phase) or continue to proliferate. In most instances, cells in G0 can be recruited into the cell cycle when necessary by interactions with one or more growth factors. The first growth phase of the cell cycle is termed G1, for the gap in time between mitosis and the next round of DNA synthesis. The duration of this phase of the cell cycle is more variable than the duration of other phases, ranging from 6 to 12 hours. During G1, RNA and proteins are synthesized but no DNA is formed. Synthesis of DNA occurs in the S phase, during which the DNA content of the cell increases from diploid to tetraploid. The duration of the S phase is similar in all cells and takes from 3 to 8 hours. The S phase is ­followed by the G2 phase, a pause of about 3–4 hours that precedes mitosis. During the G2 phase the cell has two complete sets of ­diploid chromosomes. Mitosis, or the M phase, takes no more than an hour to complete in normal cells. The ability of cells to restrict or slow their movement through the cell cycle is regulated. This can be observed when normal cells in P

P

Hyperphosphorylated pRB

Hypophosphorylated pRB

P

P P

tissue culture are damaged, for example in irradiation‐induced g­ enetic damage.56 Irradiated cells in the early stages of the cell cycle respond by halting their progress prior to the S phase; this pause in the cell cycle has been termed the G1/S checkpoint. During the pause, DNA that has been damaged by irradiation can be repaired before mutations are passed on to the genomes of daughter cells. In cells in which tumor suppressor genes are absent or not functioning properly, genetic damage is left unrepaired, which often leads to ­genetic instability in the daughter cells and additional oncogenic events. A similar checkpoint is present at the transition between the G2 and M phases of the cell cycle. The best characterized of the tumor suppressor genes are TP53 gene, activated only in cases of genetic damage or hypoxia, and the retinoblastoma (RB) gene, which is constitutively involved in the cell cycle.57,58 Both of these genes encode nuclear phosphoproteins that regulate cell cycle progression. When the RB protein (pRB) is in its hypophosphorylated form it inhibits entry of the cell into the S phase of the cell cycle by binding a transcription factor transcription factor E2 promoter‐binding‐protein (E2F) that stimulates mitosis‐promoting genes (Figure 1.7). When a cell is stimulated to divide, pRB is hyperphosphorylated by cyclins, causing it to release E2F, which enables cells to enter the S phase. Following the S phase, pRB is dephosphorylated and is, once again, able to bind E2F and inhibit entry of the cell into the S phase. In tumor cells, the ability of pRB to bind E2F is disrupted and the checkpoint is eliminated. For example, oncogenic DNA viruses (discussed later) can disrupt cell cycle control by synthesizing viral proteins that block the uptake of transcription factors by pRb protein. The TP53 gene encodes a nuclear phosphoprotein that can ­regulate movement of the cell through the cell cycle. Although this phosphoprotein (p53) is not involved in regulation of the normal cell cycle, it plays an important role in cells that have sustained ­genetic damage or in conditions of hypoxia. In the absence of functional p53 these genetically damaged cells may undergo neoplastic transformation. Through mechanisms that are not well understood, p53 can detect when a cell sustains genetic damage by UV light, irradiation, or carcinogenic chemicals and then arrests the entry of the cell into the S phase from the G1 phase of the cell cycle to allow time for the repair of cellular DNA damage via growth arrest and DNA damage‐inducible protein (GADD45), which allows for DNA repair and cyclin‐dependent kinase inhibitor

E2F

E2F

P

P

P Histone methyltransferase

Histone deacetylase

P

E2F

E2F

P

S-phase gene Activation of transcription

Block of transcription

Figure 1.7  Tumor suppressor protein pRB. When pRB is hyperphosphorylated by cyclin‐dependent kinases it releases members of the transcription factor

E2F family that then bind to DNA and stimulate progress from G1 into the S phase of the cell cycle. When pRb is hypophosphorylated it binds E2F and interacts with histone‐modifying proteins, histone deacetylase and histone methyltransferase, inhibiting progress through the cell cycle. When the ability of pRB to bind E2F is disrupted by mutations or viruses, the checkpoint is eliminated and cells may then proliferate in an uncontrolled fashion.

10    Tumors in Domestic Animals

1  (CDKN1 or p21) that inhibits phosphorylation of cell c­ycle‐ related kinases. If the extent of DNA damage is too excessive, p53 can ­promote cellular apoptosis. Although normally a short‐lived protein, after genetic damage, p53 is modified in a way that causes it to have a significantly longer half‐life. P53 will then accumulate in the nucleus, leading to cell cycle arrest by activating transcription of genes that inhibit specific cyclin‐dependent kinases and prevent the phosphorylation of the RB protein. Other effects include expression of genes involved in DNA repair or apoptosis. Cells carrying mutated TP53 genes or cells infected with oncogenic DNA viruses that alter the function of p53 do not arrest before entering the S phase of the cell cycle and are less likely

to undergo apoptosis (Figure  1.8). Affected cells can continue to replicate with damaged DNA, and those that do not develop lethal genetic changes are at risk for acquiring additional genetic damage, leading to neoplastic transformation. The canine DNA sequence for TP53 is 87% identical to the human sequence and has a similar intracellular role.59.60 Because mutations in TP53 occur in a high proportion (approximately 50%) of some types of human neoplasms, the frequency of TP53 mutations in animals has been examined. Mutations in TP53 of dogs have been detected most often in osteosarcomas, where they can be detected in approximately 40% of cases.61 In canine melanomas mutations have been identified in several tumor suppressor genes,

Normal cell (p53 normal)

Cell with p53 loss

DNA damage Activation of p53

No DNA repair

miRNA Transcription Mutation and expansion

p53 binds to DNA CDKN1A (p21) BAX GADD45

Quiescence/ Senescence

Successful DNA repair

Failed DNA repairapoptosis

Malignant tumor

Figure 1.8  Tumor suppressor protein, p53. The tumor suppressor gene (TP53) encodes a protein, p53, which is crucial for repair or apoptosis of genetically damaged cells. Signaling is mediated through growth arrest and DNA damage‐inducible protein (GADD45) that allows for DNA repair and cyclin‐ dependent kinase inhibitor 1 (CDKN1 or p21) that inhibits phosphorylation of cell cycle‐related kinases and arrests progression through the cell cycle. When genetic damage is too severe to be repaired p53 can initiate apoptosis via activation of the apoptosis‐stimulating gene BAX. Alternatively, activation of p53 in severely damaged cells can also trigger transcription of microRNAs (miRNA) that drive cell senescence. When TP53 is damaged by chemicals, radiation, viruses, or inherited defects, p53 production may be abrogated or a mutant p53 protein produced. Mutant p53 does not function normally and affected cells with damaged DNA do not arrest the cell cycle to enable DNA repair. Mutated cells are able to progress though the cell cycle giving rise to daughter cells with mutations and eventual tumor formation. Thus, when the gene TP53 is damaged or absent, tumor suppression is compromised.

An Overview of Molecular Cancer    11

including TP53.62 TP53 is also found to have altered expression or to be mutated infrequently in several types of canine and feline neoplasms, including canine mammary tumor,63 canine and ­ feline  squamous cell carcinoma,21,64 and canine mastocytoma.65 Cytogenetic analysis has demonstrated loss of several tumor suppressor genes in canine histiocytic sarcoma.66,67 Resisting cell death Genes that control programmed cell death play a significant role in tumor development when they fail to function normally.68 B‐cell lymphomas serve as examples of the importance of the genes that control apoptosis. These tumors are characterized by an increased expression of the gene BCL2 (derived from B‐cell lymphoma 2), which blocks apoptosis. BCL2 is only one of a family of genes that participate in the regulation of apoptosis. The ability of oncoproteins such as BCL2 to block cell death pathways may enable cells that have sustained genetic damage to escape mechanisms that would stimulate normal cells to undergo programmed cell death. Alternatively, neoplastic lymphocytes that overexpress BCL2 can persist and slowly form lymphoid masses, unlike normal lymphoid cells that have a finite lifespan. Consequently, cells eluding ­apoptosis could multiply and are at risk to accumulate additional genetic damage that can heighten malignancy. Overexpression of BCL2 has been demonstrated in feline lymphoma, but was not associated with prognosis.69 Enabling reproductive immortality Essentially unlimited replicative potential is a key feature in the formation of malignancies.48,70 While normal cells are capable of no more than 60–70 doublings before they become senescent and die,

malignant cells must be free of these growth constraints in order to continually grow and expand. The principal mechanism by which cells replicate without entering senescence involves the enzyme telomerase. Telomerase is typically inactive in somatic cells but is active in stem cells, germ cells, and cancer cells. Activation of telomerase is recognized in about 85–95% of human cancers.70,71 Telomerase is a ribonucleoprotein complex that includes a reverse transcriptase, an RNA template, and additional proteins (Figure 1.9). This complex is responsible for adding back short sections of DNA of 50–200 nucleotides that are lost from the chromosomal ­telomeres (specialized nucleoprotein structures at the ends of chromosomes) during normal DNA replication cycles. Continued loss of the ends of chromosomes in normal cells will eventually trigger senescence and apoptosis by activating tumor suppressor genes encoding TP53 and pRb. Cells that lack normal tumor suppressor gene activity do not arrest at appropriate cell cycle checkpoints, leading to acquisition of various mutations. Telomerase activity has not been extensively studied in veterinary oncology. There is evidence from one canine study that nearly all lymph nodes (97%) with a histologically confirmed diagnosis of lymphoma (various subtypes) had detectable telomerase activity and activity was greater than that seen in normal lymph nodes.72 More work in this area is needed. Inducing angiogenesis Solid neoplasms depend on the blood vessels and supporting stroma that they recruit from adjacent tissue for their survival and growth.48,73 Without vascularization, growing tumor masses are limited to about a 1–2 mm diameter. In normal tissue and in ­neoplastic masses angiogenesis is regulated by competing pro‐ and anti‐angiogenic signaling. The transition to a pro‐angiogenic status

Metaphase

Bridge-fusion-breakage cycle

Inactive telomerase

Shortened telomerase

Anaphase

–Telomerase

+Telomerase

Mitotic catastrophe

Senescent cell

CANCER

Figure 1.9  Telomerase. Telomerase is an enzyme that enables cells to replicate in an unlimited fashion. Cells with repressed telomerase activity such as somatic cells eventually reach senescence after a finite number of mitoses and undergo cellular senescence, a permanent growth arrest state, or apoptosis. Telomerase adds back short sections of DNA that were lost from the chromosomal telomeres (repetitive nucleoprotein sequences at the ends of chromosomes) during normal DNA replication cycles. Cells with telomerase activity such as stem cells, germ cells and cancer cells can potentially proliferate indefinitely and are potentially immortal.

12    Tumors in Domestic Animals

occurs when anti‐angiogenic signaling is overwhelmed. Tumor cells secrete growth factors such as vascular endothelial growth factor A (VEGF‐A) and various types of fibroblastic growth factors (FGF) or stimulate other cells to release angiogenic factors that stimulate the vessels and supporting stroma in tumors (Figure 1.10). There are many cell types that participate in angiogenesis. In addition to the tumor cells and adjacent stromal elements, bone marrow–derived cells, primarily cells of the innate immune system, macrophages, neutrophils, and mast cells, and also myeloid precursors infiltrating at the margins of neoplastic lesions release angiogenic factors contributing to the ingrowth of new vessels.73 Angiogenesis was once thought to be significant primarily when robust tumor growth was occurring, yet is now known to begin early in the process of tumorigenesis and is evident in preneoplastic and benign lesions.74 The processes of angiogenesis and stroma formation are similar in tumors and wound healing, leading to the conceptual description of tumors as nonhealing wounds. There are some distinct differences in the structure and function of the vessels that are formed during each process.75 In tumors, the blood vessels are poorly differentiated and are not distributed uniformly through the tumor. Tumor blood vessels tend to be more tortuous and dilated than normal vessels, with gaps in the endothelium rendering them persistently permeable, unlike vessels in healing wounds that have a transient phase of permeability.76 Since most tumor cell entry into the bloodstream occurs between gaps in the endothelial cells it is likely that metastasis is facilitated by these abnormal vessels. Increased interstitial pressure due to the permeable vessels and the lack of lymphatics to carry away the leaked fluid lead to edema formation. This edema and the resultant interstitial fluid pressure tend to collapse the vessels within the tumor, thus obstructing local blood flow. The density of vascular supply to tumors is frequently minimally adequate and is deficient in 1

arteriolar supply, in particular. As a result, irregular blood flow and ­ erfusion cause localized areas of hypoxia and anoxia, leading to p apoptosis or necrosis. Without angiogenesis, tumors have to rely on cellular diffusion to provide needed nutrients and eliminate waste products. Angiogenesis plays an essential role in sustained tumor growth, as well as metastasis. Recruited endothelial cells do more than provide perfusion, as endothelial cells also secrete growth factors that can stimulate tumor cell growth. Angiogenesis, measured as the density of the microvasculature within a tumor, has been shown to be a significant prognostic indicator for some human neoplasms such as those of the lung and breast.77,78 Because of this powerful effect on tumor growth, angiogenesis is an area of particular interest in tumor biology. Angiogenesis by itself, however, is not an indication of malignancy as even benign neoplasms have the ability to stimulate vascular growth. Tumor stroma is composed of non‐neoplastic connective tissue, blood vessels, and inflammatory cells.48 While the vasculature is an essential component of stroma formation because of its nutrient support of the neoplasm, the greatest proportion of the tumor stroma is nonvascular. The noncellular components of the stroma include collagen types I, III, and V, glycosaminoglycans, proteoglycans, fibronectin, fibrin, and plasma proteins. Fibroblasts, endothelial cells, and inflammatory cells are the principal cellular constituents. Initially, the tumor stroma resembles granulation tissue with a high density of blood vessels and smaller numbers of fibroblasts. The persistent permeability of tumor vessels allows a continued leakage of macromolecules, engendering a perivascular deposition of fibrin that serves as scaffolding for migration of host stromal cells and tumor stroma formation. As this tissue matures, collagenous stroma predominates and vascularity diminishes, creating a desmoplastic or scirrhous response. For reasons that are unclear, the amount of

2

3

Growing tumor Malignant tumor Angiogenic factors Metastasis Angiogenesis

Nutrients

Figure 1.10  Angiogenesis. Tumor angiogenesis is a critical step for the growth of the primary mass as well as metastatic masses. Tumor cells release angio-

genic factors that stimulate budding of new vessels that deliver oxygen and nutrients to the growing tumor cells and provide venous drainage to remove waste products. New vessels also provide an avenue for vascular metastasis. Tumor size is limited to approximately 1 mm in diameter without supporting blood vessels.

An Overview of Molecular Cancer    13

stroma produced by different neoplasms varies considerably. Certain carcinomas such as gastric, urothelial, and mammary carcinomas are more prone to develop desmoplasia (scirrhous response) than other neoplasms. The resultant masses are very firm to the touch, and the stroma can comprise a larger proportion of the mass than the tumor cells do. A newly emerging understanding of epithelial–mesenchymal interactions is clarifying the role of fibroblasts and other stromal elements in tumor growth and desmoplasia. Fibroblasts or myofibroblasts adjacent to carcinomas, termed cancer‐associated fibroblasts, have a fetal‐like phenotype that differs from fibroblasts in other parts of the body.79 In some cases tumor cells secrete extracellular vesicles containing genetic information that can re‐program mesenchymal stem cells to produce extracellular matrix.80 Tumor‐ associated fibroblasts stimulate tumor cell proliferation via release of growth factors and proteases in response to cytokines signaling by neoplastic epithelial cells and can also facilitate angiogenesis, invasion, and metastasis. Overall, there are multiple interactions between cancer cells and adjacent stromal elements and inflammatory cells that can facilitate tumor growth and metastasis in complex patterns (Figure 1.11). Activating invasion and metastasis Metastasis is an inefficient multistep process, and only a very small proportion of cells are able to complete the process.81–84 Once a malignancy develops, a metastatic subclone may arise within the tumor through the process of tumor progression. In epithelial tumors a common initial step supporting invasion is the loss of intercellular adhesion due to impaired activity of cell adhesion factors such as E‐cadherins. Conversely, cell surface adhesion molecules such as N‐cadherin, associated with cell migration during development, may be re‐expressed. A series of steps occur during the transition from noninvasive (in situ) carcinoma to a metastatic carcinoma (Figure  1.12). Initially, metastatic cells penetrate the basement membrane in a two‐step process. First, metastatic cells attach to the basement membrane via laminin and fibronectin

EGF

receptors among others; subsequently, they secrete hydrolytic enzymes (proteases) that degrade the basement membrane. The next step involves locomotion. Tumor cells migrate into the extracellular matrix facilitated by the release of products secreted by the tumor cells and host inflammatory cells, particularly macrophages. Connective tissues are unequally susceptible to invasive processes. Hyaline cartilage, for example, contains inhibitors of matrix degrading enzymes and is highly resistant to invasion. Eventually metastatic cells encounter a blood or lymph vessel. Entry of tumor cells into the bloodstream or lymphatics, termed intravasation, is only possible after attachment of tumor cells to the basement membrane of the vessel and degradation of this barrier. Extravasion is facilitated by the increased permeability of the new, but abnormal, blood vessels formed within tumors compared to vessels in normal tissue. Tumor cells can then pass through the junctions between adjacent endothelial cells or pass directly through the intact endothelium. Lymphatic vessels pose less of a barrier to entry than blood vessels because lymphatic vessels lack a basement membrane. The mere presence of tumor cells in vessels does not ensure that those cells will eventually give rise to metastatic populations. Once tumor cells enter the vasculature, they encounter the array of host cells involved in immune‐mediated killing of tumor cells. To ­survive, the tumor cells must evade intense scrutiny by the host immune response. One way tumor cells evade host defenses is by interacting with blood components, such as platelets and fibrin, to form thrombi. When the tumor cells are enclosed by fibrin, they may be protected from recognition by the immune system and have a better chance to survive in the hostile environment of the blood. Extravasation of surviving tumor cells may occur in a directed, nonrandom fashion.

Epithelial–mesenchymal transition

Recent research has revealed a process termed the epithelial–­ mesenchymal transition (EMT)85 that regulates the acquisition of capabilities needed to facilitate invasion and metastasis. In keeping with the view that cancer cells do not possess unique behaviors or

Cancer cells

CSF-1

Proteases VEGF

PDGF

HGF

TGF-β

Extracellular matrix

Proteases Tumor promoting inflammatory cells

TGF-β Basement membrane/ sequestered growth factors

Cancer-associated fibroblasts

VEGF FGF2

Endothelial cells

ANG-1 PDGF

Pericytes

Figure 1.11  Tumor cell and stromal interactions. Interactions between tumor cells and the adjacent stroma play a key role in many facets of tumor evolution.

Multiple interactions between the tumor cells and stromal and inflammatory cells mediate tumor growth, differentiation, and metastasis, as well as host tissue responses.

14    Tumors in Domestic Animals

Tumor cell Clonal expansion Primary tumor

Metastatic subclones Lymphocyte Intravasation

Interaction with host lymphocyte

Tumor cell embolus

Platelets

Extravasation

Adhesion

Metastatic tumor

Angiogenesis

Growth Figure 1.12  Metastasis. Invasion and metastasis are hallmarks of malignant tumors. Each step in the process of metastasis can involve progressive histological changes and/or molecular alterations, some of which are ­illustrated here.

capabilities, but rather co‐opt normal cellular processes for the ­detriment of the host, EMT involves the emergence of capabilities normally only expressed in embryogenesis or wound healing. Malignant epithelial cells may stably or transiently acquire the ability to invade, resist apoptosis, and invade locally through this process. In addition to the altered behaviors, a phenotypic alteration can also be observed. Typical polygonal epithelial cells can be changed to spindle‐shaped fibroblast‐like cells with the ability to locomote, resist apoptosis, and secrete enzymes that digest the local stroma. The mechanisms that facilitate the gene expression driving these changes are not well characterized, but studies suggest that interaction with other cells in the local environment enable or facilitate the transition. Local environmental factors are likely to play an

important role. Histologic and immunohistochemical examination of the invasive margins of some carcinomas reveals that EMT occurs only at the leading edge of the neoplasm and not in the center of the mass. Malignancies have been likened to villages, rather than monotypic masses of proliferating cells, because of the important interactions between stromal and inflammatory cells and the tumor cells.48 Cell– cell interactions include secretion of factors by mesenchymal stem cells in response to signals released by the tumor cells that enhance invasion of the tumor cells. Inflammatory cells, particularly those of the innate immune system, can also facilitate tumor development. Macrophages have been shown to facilitate breakdown of the extracellular matrix to enhance invasion. In an experimental model of mammary carcinoma, tumor‐associated macrophages secrete epithelial growth factor to support mammary carcinoma growth and the tumor cells secrete CSF‐1 to stimulate the macrophages. This interaction also facilitates intravascular invasion and metastasis.86 Once malignant cells have disseminated into the bloodstream, lymph flow, or other spaces they still have to undergo a series of steps to establish a viable mass at a new site. Recent studies have elucidated the predilection for certain tumors to metastasize to particular organs. Paget’s 1889 theory of “seed and soil,” which explains why certain tumors tend to metastasize to a particular set of organs, holds true today, although the mechanisms are now becoming clear. Some tumor cells are guided to particular organs because they bind to tissue‐specific endothelial cell surface markers. In other tumor types, the cells bear receptors to specific chemokines and home towards organs that release these chemokines; they are less likely to be found in organs that do not release these chemokines.87 The newly extravasated tumor clone must next acquire a blood supply. A new vascular network is needed not only to provide nutrients to the growing tumor, but also to carry away waste products. Once a metastatic tumor has established a proper vascular supply, its growth may be limited by inhibitory growth factors, by a restrictive growth environment, or by a cytotoxic response by the host. There are three principal pathways of metastasis: lymphatic, hematogenous, and direct extension. Lymphatic metastasis Lymphatic invasion occurs primarily at the periphery of the tumor. Lymphatic vessels offer little resistance to penetration by tumor cells because they lack a basement membrane. Clumps or single‐cell tumor emboli may be trapped in the first lymph node encountered, or they may traverse or bypass lymph nodes to form a more distant metastasis, a condition termed skip metastasis. Tumor cells are usually first detected histologically in the subcapsular region of the lymph nodes. Based on extensive studies in humans and limited data from animals, carcinomas have a predilection for metastasis by the lymphatic route compared to sarcomas, although the mechanism is unclear. In dogs with mammary cancer, regional lymph nodes appeared to function as good filters since bypassing the node was found to be uncommon. An enlarged local lymph node does not necessarily mean metastasis has occurred; the node may be enlarged due to hyperplasia and/or metastasis. In most cases, an enlarged lymph node draining a region with malignancy is ­probably no longer immunologically effective, but there is no consensus regarding the value of the removal of such an enlarged node. ­Fine‐ needle aspiration by an experienced cytologist or biopsy for histologic examination is necessary to distinguish lymphoid hyperplasia from metastasis and to allow appropriate clinical staging and treatment planning.

An Overview of Molecular Cancer    15

Hematogenous metastasis Tumor cells can enter the blood directly by invasion of blood ­vessels or indirectly via the lymphatic system that connects with venous tributaries at sites such as the thoracic duct and subsequently enter into the vena cava. Distribution of hematogenous metastases can be initially explained by the hemodynamic theory based on circulatory anatomy. Briefly, metastatic emboli from ­primary tumors spread via the vena cava drainage (mammary, skin, soft tissue, bone, thyroid tumors) unless they arise in the abdominal organs (gastrointestinal, splenic, and pancreatic tumors) drained by the portal vein. The majority of tumor cells are arrested in the first capillary bed they encounter. The first capillary filter of vena caval drainage is the lung, and the liver is the first microvascular field draining the portal vein system. From those sites, tumors can spread to secondary microvascular filters, such as bone marrow. However, in the human, and to a lesser extent also in domestic animals, preferential metastatic sites can  also be explained by organ tropism or the “seed and soil” ­hypothesis described earlier. Direct extension metastasis The coelomic surfaces, covered with a film of fluid, are an ideal site for metastatic seeding. Neoplastic cells shed from a primary tumor can survive when implanted onto the serosal surfaces of body cavities or organs. Implantation of tumor cells in serous cavities is often accompanied by an accumulation of fluid. Peritoneal or pleural ­carcinomatosis is associated either with a primary tumor within a coelomic cavity (ovarian or pulmonary carcinoma) or with metastases from carcinoma elsewhere in the body (e.g., mammary carcinoma). Pleuritic carcinomatosis in dogs and cats with mammary carcinoma was found to be invariably associated with the presence of pulmonary metastasis.88,89 The spread of mesotheliomas is often restricted to the same coelomic cavity as the site of origin. Mechanical transfer can also occur via contaminated surgical tools or fine‐needle aspirates. There are two naturally occurring clonally transmittable malignancies that can be spread by contact. These include transmissible venereal tumor of dogs, in which tumor cells are transferred by coitus but can then spread to other sites in some animals, most of whom are likely immunocompromised.90 A more recently described example is the disease of Tasmanian devils called devil facial tumor disease, in which bite wounds appear capable of transmitting malignant mesenchymal cells that can eventually ­ metastasize.91

Successful metastasis

A tumor cell that has arrived at a new site following dissemination still has a number of challenges before it can expand from a micrometastasis into a metastatic mass. First, the interactions with local stromal cells and inflammatory cells may no longer be present and, in some cases, the new cell cannot expand in the absence of these supporting elements. There may be a considerable time required for sufficient new mutations to develop that enable the micrometastasis to proliferate in the new environment. Inadequate ability to support angiogenesis is a common limitation. Primary tumors may secrete inhibitory factors that suppress growth of the tumor at the new sites. In such circumstances surgical or chemotherapeutic removal of the primary mass can stimulate growth of previously undetected microscopic metastases. These circumstances would explain the sudden appearance of metastases sometimes years after the primary lesion has been removed or treated.

Metastasis site selection

Since extravasation requires adhesion to endothelial cells or underlying basement membrane, tumor cell attachment may be directed to specific sites by receptor and ligand interactions. The release of chemokines can also direct some types of tumor cells to specific organs.87 Clearly, the lung and the liver are common sites of metastasis for many types of neoplasms. Organ tropism seems to play a role in metastasis of melanomas in dog and human with frequent spread to the brain. Prostatic carcinomas in dog and human ­frequently spread to bones. Occult micrometastases are frequently present in these unique sites and in lymph nodes and lung at the time of the primary tumor diagnosis of these tumors. Most osseous metastases have intertrabecular growth. Only at advanced stages do osteolysis or endosteal and periosteal bone formation occur.92 The frequency of osseous metastasis may be underestimated when the bones are not carefully checked radiographically or during the postmortem examination. Bone metastasis in dogs is frequently underestimated, likely from failure to carefully examine the cut surface of long bones. In a detailed postmortem study, examination of transected bones revealed that 17% of dogs with visceral metastasis from a variety of neoplasms also had skeletal metastasis.92 Dogs with epithelial malignancies with visceral metastasis also had bone metastasis in 24% of cases, often affecting more than one bone. Common sites are flat bones, including the ribs, the vertebrae, and the metaphyseal region of the long bones. Frequently, multiple sites in the bones are affected, and metastatic involvement of bone in this study was always accompanied by concurrent soft tissue metastasis. Most primary tumors responsible for bone metastases in the dog are carcinomas, including those of mammary gland,93,94 lungs,92,94,95 and prostate.92,94 Metastasis to bone from mammary96 and pulmonary carcinomas,97 along with various individual case reports have been reported in cats.94 A particularly impressive example of tissue tropism for metastasis is found in the pulmonary carcinomas of cats. These n ­ eoplasms can metastasize widely, but have a predilection for spread to the distal toes. The underlying mechanisms are not known.98

Paraneoplastic syndromes

Paraneoplastic syndromes are defined as systemic complications of neoplasia that are remote from the primary tumor.99 Frequently, the effects of the paraneoplastic syndrome can be more injurious than the associated malignancy and may be the reason the animal was brought to the veterinarian. The common paraneoplastic ­syndromes in veterinary medicine are listed in Table 1.1. Paraneoplastic syndromes may serve as diagnostic aids or as specific tumor markers for treatment response and failure. These effects are generally unrelated to the size of the tumor, the presence of metastasis, or the physiologic activity of the tissue of primary origin. Most of the examples in veterinary medicine are associated with the production of native (true) hormone from cells that normally produce that hormone or from the “ectopic” production of a hormone‐like peptide by tumor cells that are not in an endocrine organ. Excessive insulin production by neoplastic islet cells and production of a parathormone‐like peptide by neoplastic lymphocytes or apocrine cells of the canine anal sac are examples of each category, respectively. In order to definitively establish that a paraneoplastic condition is a result of a specific neoplasm, one or more criteria have to be met. These criteria include the following: (1) concentration of the product (e.g., calcium) decreases after removal or treatment of the neoplasm (e.g., an anal sac carcinoma that was

16    Tumors in Domestic Animals

Table 1.1  Selected veterinary paraneoplastic syndromes and associated neoplasms Paraneoplastic syndromes Endocrine Hypercalcemia of malignancy

Hypoglycemia

Ectopic ACTH Cutaneous Paraneoplastic pemphigus Alopecia Necrolytic migratory erythema Hematologic Hypergammaglobulinemia

Associated neoplasms

Lymphoma Apocrine gland carcinoma of anal sac Mammary carcinoma Thymoma Others Hepatocellular carcinoma Salivary gland carcinoma Leiomyoma/leiomyosarcoma Plasma cell tumor Lymphoma Others Pulmonary carcinoma Lymphoma Pancreatic carcinoma (cat) Others Glucaconoma

Anemia

Multiple myeloma Lymphoma Many neoplasms

Erythrocytosis

Renal carcinoma

Neurologic Myasthenia gravis Peripheral neuropathy Renal Glomerulonephritis

Gastrointestinal

Miscellaneous Hypertrophic osteopathy

Thymoma Others Insulinoma Others Multiple myeloma Polycythemia vera Others Gastroduodenal ulceration Mast cell tumors Gastrinoma Pulmonary carcinoma Other thoracic masses Urinary bladder rhabdomyosarcoma

secreting the trophic hormone PTH‐rp is removed and serum calcium decreases); (2) product concentrations are maintained after removal of the normal gland that controls the concentration of that product (e.g., calcium concentration remains high following removal of a parathyroid gland); (3) a positive arteriovenous concentration gradient of the hormone exists across the tumor; and (4) synthesis and secretion of the product by the tumor in vitro occurs. In veterinary medicine, the first criterion  –  decreased concentration of product after tumor ablation – is most commonly used to diagnose a paraneoplastic syndrome. The pathogenesis of paraneoplastic syndromes is thought to result from several processes. De‐repression of a gene may result in production of a substance with biologic activity. In fact there may be many products from a given tumor, but only the active substances are detectable. One example would be the production of hormone precursors that do not exhibit activity unless metabolized (i.e., prohormone production). Ectopic receptor production by a tumor has also been reported and accounts for displaced activity of a humoral substance (e.g., thymoma and acetylcholine receptor production). The third theory is termed “forbidden contact” and implies that there is exposure to substances that are normally sequestered from the body (i.e., antigens of normal or neoplastic origin) and therefore are recognized by the immune system as foreign. Immune complex

formation from antigenic exposure to these normally sequestered antigens may result in a physiologic or pathologic event leading to clinical signs. Examples include anaphylaxis, coagulopathies, vasculitis, glomerulonephritis, and hemolytic anemia. Endocrine syndromes are a frequent manifestation of paraneoplastic disease. Protein hormones, hormone precursors, or cytokines may be produced or metabolized by tumors. Some types of hormones, such as steroid hormones, thyroid hormone derivatives and catecholamines, are produced exclusively by tumors originating from glands that normally produced these substances. The frequency of biologically active peptide‐producing neoplasms can be explained by the fact that most cells secrete peptide hormones that function in paracrine signaling. These peptide hormones may be expressed in excess when cells become malignant and their ­numbers increase by clonal expansion. Cancer cachexia is one of the more common paraneoplastic syndromes encountered in veterinary and human medicine. ­ Affected animals are anemic, weak, easily fatigued, lose weight, and have diminished immune function. There are characteristic metabolic changes associated with this syndrome that affect carbohydrates, proteins, and lipids.100 Growth of the tumor occurs at the expense of the host. Increased serum lactate levels and insulin levels characterize abnormal carbohydrate metabolism. There is a loss of muscle mass and hypoalbuminemia in affected animals because protein catabolism exceeds protein synthesis. Typically these ­animals will have profound muscle wasting and prominent boney protuberances but surprising amounts of abdominal or subcutaneous fat. Starvation and parasitism drain adipose reserves first, resulting in serous atrophy of fat whereas protein and muscle loss are recognized at later stages of these conditions. Wound healing and immunity are also affected by altered protein metabolism. The loss of protein in cancer patients develops because amino acids are redirected from protein synthesis into gluconeogenesis. Although tumor cells are less capable of using lipids for energy than normal cells, cancer cachexia also promotes fat utilization. Cancer cachexia has been attributed to the effects of tumor necrosis factor, interleukins 1 and 6, and interferons gamma and alpha.100 It affects 50–80% of human patients with malignancies.

References

1. Garraway, L.A. and Lander, E.S. (2013) Lessons from the cancer genome. Cell 153:17–37. 2. Alexandrov, L.B., Nik‐Zainal, S., et al. (2013) Signatures of mutational processes in human cancer. Nature 500:415–421. 3. Negrini, S., Gorgoulis, V.G., et al. (2010) Genomic instability – an evolving hallmark of cancer. Nat Rev Mol Cell Biol 11:220–228. 4. Vogelstein, B., Papadopoulos, N., et al. (2013) Cancer genome landscapes. Science 339:1546–1558. 5. Timp, W. and Feinberg, A.P. (2013) Cancer as a dysregulated epigenome allowing cellular growth advantage at the expense of the host. Nat Rev Cancer 13:497–510. 6. You, J.S. and Jones, P.A. (2012) Cancer genetics and epigenetics: two sides of the same coin? Cancer Cell 22:9–20. 7. Xie, X., Lu, J., et al. (2005) Systematic discovery of regulatory motifs in human promoters and 3′ UTRs by comparison of several mammals. Nature 434:338–345. 8. Jansson, M.D. and Lund, A.H. (2012) MicroRNA and cancer. Mol Oncol 6:590–610. 9. Knudsen, E.S. and Wang, J.Y. (2010) Targeting the Rb‐pathway in cancer therapy. Clin Cancer Res 16:1094–1099. 10. Nambiar, P.R., Jackson, M.L., et  al. (2001) Immunohistochemical detection of tumor suppressor gene p53 protein in feline injection site‐associated sarcomas. Vet Pathol 38:236–238. 11. Nambiar, P.R., Haines, D.M., et al. (2000) Mutational analysis of tumor suppressor gene p53 in feline vaccine site‐associated sarcomas. Am J Vet Res 61:1277–1281.

An Overview of Molecular Cancer    17

12. Burrell, R.A., McGranahan, N., et al. (2013) The causes and consequences of genetic heterogeneity in cancer evolution. Nature 501:338–345. 13. Aird, K.M. and Zhang, R. (2015) Nucleotide metabolism, oncogene‐induced senescence and cancer. Cancer Lett 356:204–210. 14. Hanahan, D. and Weinberg, R.A. (2000) The hallmarks of cancer. Cell 100:57–70. 15. Hanahan, D. and Weinberg, R.A. (2011) Hallmarks of cancer: the next generation. Cell 144:646–674. 16. Croce, C.M. (2008) Oncogenes and cancer. N Engl J Med 358:502–511. 17. Ranieri, G., Pantaleo, M., et al. (2013) Tyrosine kinase inhibitors (TKIs) in human and pet tumours with special reference to breast cancer: a comparative review. Crit Rev Oncol Hematol 88:293–308. 18. Bertoli, C., Skotheim, J.M., et al. (2013) Control of cell cycle transcription during G1 and S phases. Nat Rev Mol Cell Biol 14:518–528. 19. Musgrove, E.A., Caldon, C.E., et al. (2011) Cyclin D as a therapeutic target in cancer. Nat Rev Cancer 11:558–572. 20. Murakami, Y., Tateyama, S., et  al. (2000) Amplification of the cyclin a gene in canine and feline mammary tumors. J Vet Med Sci 62:783–787. 21. Murakami, Y., Tateyama, S., et al. (2000) Immunohistochemical analysis of cyclin A, cyclin D1 and P53 in mammary tumors, squamous cell carcinomas and basal cell tumors of dogs and cats. J Vet Med Sci 62:743–750. 22. Pylayeva‐Gupta, Y., Grabocka, E., et al. (2011) Ras oncogenes: weaving a tumorigenic web. Nat Rev Cancer 11:761–774. 23. Futreal, P.A., Coin, L., et  al. (2004) A census of human cancer genes. Nat Rev Cancer 4:177–183. 24. Edwards, M.D., Pazzi, K.A., et  al. (1993) C‐N‐Ras is activated infrequently in canine malignant lymphoma. Toxicol Pathol 21:288–291. 25. Mayr, B., Holzheu, M., et al. (2003) N‐Ras mutation in a canine lymphoma: short communication. Acta Vet Hung 51:91–94. 26. Mayr, B., Reifinger, M., et al. (1999) Cytogenetic, ras, and p53: studies in cases of canine neoplasms (hemangiopericytoma, mastocytoma, histiocytoma, chloroma). J Hered 90:124–128. 27. Mayr, B., Schaffner, G., et al. (2003) K‐ras mutations in canine pancreatic cancers. Vet Rec 153:87–89. 28. Mayr, B., Winkler, G., et al. (2002) N‐ras mutation in a feline lymphoma. low frequency of n‐ras mutations in a series of feline, canine and bovine lymphomas. Vet J 163:326–328. 29. Murua Escobar, H., Gunther, K., et al. (2004) Absence of ras‐gene hot‐spot mutations in canine fibrosarcomas and melanomas. Anticancer Res 24:3027–3028. 30. Mayr, B., Schaffner, G., et al. (2003) N‐ras mutations in canine malignant melanomas. Vet J 165:169–171. 31. Richter, A., Murua Escobar, H., et al. (2005) Ras gene hot‐spot mutations in canine neoplasias. J Hered 96:764–765. 32. Watzinger, F., Mayr, B., et al. (2001) Comparative analysis of ras proto‐oncogene mutations in selected mammalian tumors. Mol Carcinog 30:190–198. 33. Griffey, S.M., Kraegel, S.A., et al. (1998) Rapid detection of K‐Ras gene mutations in canine lung cancer using single‐strand conformational polymorphism analysis. Carcinogenesis 19:959–963. 34. Griffey, S.M., Kraegel, S.A., et al. (1998) K‐ras mutations in 239pu02 canine lung neoplasms. Cancer Lett 132:1–5. 35. Kraegel, S.A., Gumerlock, P.H., et al. (1992) K‐ras activation in non‐small cell lung cancer in the dog. Cancer Res 52:4724–4727. 36. Usher, S.G., Radford, A.D., et al. (2009) Ras, Flt3, and C‐kit mutations in immunophenotyped canine leukemias. Exp Hematol 37:65–77. 37. Gumerlock, P.H., Meyers, F.J., et  al. (1989) Activated C‐N‐ras in radiation‐induced acute nonlymphocytic leukemia: twelfth codon aspartic acid. Radiat Res 117:198–206. 38. Avery, A. (2009) Molecular diagnostics of hematologic malignancies. Top Companion Anim Med 24:144–150. 39. Letard, S., Yang, Y., et  al. (2008) Gain‐of‐function mutations in the extracellular domain of kit are common in canine mast cell tumors. Mol Cancer Res 6:1137–1145. 40. Webster, J.D., Yuzbasiyan‐Gurkan, V., et al. (2007) Cellular proliferation in canine cutaneous mast cell tumors: associations with c‐kit and its role in prognostication. Vet Pathol 44:298–308. 41. Downing, S., Chien, M.B., et  al. (2002) Prevalence and importance of internal tandem duplications in exons 11 and 12 of c‐kit in mast cell tumors of dogs. Am J Vet Res 63:1718–1723. 42. Sherbenou, D.W. and Druker, B.J. (2007) Applying the discovery of the Philadelphia chromosome. J Clin Invest 117:2067–2074. 43. Figueiredo, J.F., Culver, S., et al. (2012) Acute myeloblastic leukemia with associated bcr‐abl translocation in a dog. Vet Clin Pathol 41:362–368. 44. Breen, M. and Modiano, J.F. (2008) Evolutionarily conserved cytogenetic changes in hematological malignancies of dogs and humans – man and his best friend share more than companionship. Chromosome Res 16:145–154. 45. Cruz Cardona, J.A., Milner, R., et al. (2011) Bcr‐Abl translocation in a dog with chronic monocytic leukemia. Vet Clin Pathol 40:40–47.

46. Culver, S., Borst, L., et al. (2013) Molecular characterization of canine BCR‐ABL‐ positive chronic myelomonocytic leukemia before and after chemotherapy. Vet Clin Pathol 42(3):314–322. 47. McManus, M.T. (2003) MicroRNAs and cancer. Semin Cancer Biol 13:253–258. 48. Pietras, K., Ostman, A. (2010) Hallmarks of cancer: interactions with the tumor stroma. Exp Cell Res 316:1324–1331. 49. Hollstein, M., Sidransky, D., et al. (1991) P53 mutations in human cancers. Science 253:49–53. 50. Bieging, K.T., Mello, S.S., et al. (2014) Unravelling mechanisms of p53‐mediated tumour suppression. Nat Rev Cancer 14:359–370. 51. Bonsdorff, T.B., Jansen, J.H., et al. (2008) Second hits in the FLCN gene in a hereditary renal cancer syndrome in dogs. Mamm Genome 19:121–126. 52. Lingaas, F., Comstock, K.E., et al. (2003) A mutation in the canine bhd gene associated with hereditary multifocal renal cystadenocarcinoma and is ­ nodular d ­ ­ ermatofibrosis in the German shepherd dog. Hum Mol Genet 12:3043–3053. 53. Boerkamp, K.M., Teske, E., et al. (2014) Estimated incidence rate and distribution of tumours in 4,653 cases of archival submissions derived from the Dutch golden retriever population. BMC Vet Res 10:34. 54. Berger, B. and Feldman, E.C. (1987) Primary hyperparathyroidism in dogs: 21 cases (1976–1986). J Am Vet Med Assoc 191:350–356. 55. Raynaud, C., Mallory, A.C., et al. (2014) Chromatin meets the cell cycle. J Exp Bot 65:2677–2689. 56. Jeggo, P.A. and Lobrich, M. (2006) Contribution of DNA repair and cell cycle checkpoint arrest to the maintenance of genomic stability. DNA Repair (Amst) 5:1192–1198. 57. Burkhart, D.L. and Sage, J. (2008) Cellular mechanisms of tumour suppression by the retinoblastoma gene. Nat Rev Cancer 8:671–682. 58. Olivier, M., Hollstein, M., et al. (2010) Tp53 mutations in human cancers: origins, consequences, and clinical use. Cold Spring Harb Perspect Biol 2:a001008. 59. Veldhoen, N. and Milner, J. (1998) Isolation of canine P53 Cdna and detailed characterization of the full length canine P53 protein. Oncogene 16:1077–1084. 60. Zhang, J., Chen, X., et al. (2009) Establishment of a dog model for the p53 family pathway and identification of a novel isoform of p21 cyclin‐dependent kinase inhibitor. Mol Cancer Res 7:67–78. 61. Kirpensteijn, J., Kik, M., et al. (2008) Tp53 gene mutations in canine osteosarcoma. Vet Surg 37:454–460. 62. Koenig, A., Bianco, S.R., et al. (2002) Expression and significance of P53, Rb, P21/ Waf‐1, P16/Ink‐4a, and Pten tumor suppressors in canine melanoma. Vet Pathol 39:458–472. 63. Rungsipipat, A., Tateyama, S., et al. (1999) Immunohistochemical analysis of C‐Yes and C‐Erbb‐2 oncogene products and p53 tumor suppressor protein in canine mammary tumors. J Vet Med Sci 61:27–32. 64. Jasik, A. and Reichert, M. (2011) New P53 mutations in canine skin tumours. Vet Rec 169:684. 65. Ginn, P.E., Fox, L.E., et al. (2000) Immunohistochemical detection of P53 tumor‐ suppressor protein is a poor indicator of prognosis for canine cutaneous mast cell tumors. Vet Pathol 37:33–39. 66. Shearin, A.L., Hedan, B., et al. (2012) The Mtap‐Cdkn2a locus confers susceptibility to a naturally occurring canine cancer. Cancer Epidemiol Biomarkers Prev 21:1019–1027. 67. Hedan, B., Thomas, R., et  al. (2011) Molecular cytogenetic characterization of canine histiocytic sarcoma: a spontaneous model for human histiocytic cancer identifies deletion of tumor suppressor genes and highlights influence of genetic background on tumor behavior. BMC Cancer 11:201. 68. Cotter, T.G. (2009) Apoptosis and cancer: the genesis of a research field. Nat Rev Cancer 9:501–507. 69. Dank, G., Lucroy, M.D., et al. (2002) Bcl‐2 and Mib‐1 labeling indexes in cats with lymphoma. J Vet Intern Med 16:720–725. 70. Novak, K.D. (2003) Telomeres and telomerases in cancer. MedGenMed 5:21. 71. Wright, W.E. and Shay, J.W. (2000) Telomere dynamics in cancer progression and prevention: fundamental differences in human and mouse telomere biology. Nat Med 6:849–851. 72. Renwick, M.G., Argyle, D.J., et  al. (2006) Telomerase activity and telomerase reverse transcriptase catalytic subunit expression in canine lymphoma: correlation with Ki67 immunoreactivity. Vet Comp Oncol 4:141–150. 73. Hanahan, D. and Folkman, J. (1996) Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353–364. 74. Raica, M., Cimpean, A.M., et al. (2009) Angiogenesis in pre‐malignant conditions. Eur J Cancer 45:1924–1934. 75. Nagy, J.A., Chang, S.H., et al. (2010) Heterogeneity of the tumor vasculature. Semin Thromb Hemost 36:321–331. 76. Reymond, N., d’Agua, B.B., et al. (2013) Crossing the endothelial barrier during metastasis. Nat Rev Cancer 13:858–870.

18    Tumors in Domestic Animals

77. Uzzan, B., Nicolas, P., et al. (2004) Microvessel density as a prognostic factor in women with breast cancer: a systematic review of the literature and meta‐analysis. Cancer Res 64:2941–2955. 78. Meert, A.P., Paesmans, M., et al. (2002) The role of microvessel density on the survival of patients with lung cancer: a systematic review of the literature with meta‐analysis. Br J Cancer 87:694–701. 79. Kalluri, R. and Zeisberg, M. (2006) Fibroblasts in cancer. Nat Rev Cancer 6:392–401. 80. Haga, H., Yan, I.K., et  al. (2015) Tumour cell‐derived extracellular vesicles interact with mesenchymal stem cells to modulate the microenvironment and enhance cholangiocarcinoma growth. J Extracell Vesicles 4:24900. 81. Talmadge, J.E., Fidler, I.J. (2010) AACR Centennial Series: The biology of cancer metastasis: historical perspective. Cancer Res 70:5649–5669. 82. Alderton, G.K. (2013) Metastasis: Active lymph nodes. Nat Rev Cancer 13:606–607. 83. Alderton, G.K. (2013) Metastasis: Exit this way. Nat Rev Cancer 13:523. 84. Alderton, G.K. (2013) Metastasis: Polarizing metastasis. Nat Rev Cancer 13:75. 85. Alderton, G.K. (2013) Metastasis: Epithelial to mesenchymal and back again. Nat Rev Cancer 13:3. 86. Qian, B.Z. and Pollard, J.W. (2010) Macrophage diversity enhances tumor ­progression and metastasis. Cell 141:39–51. 87. Balkwill, F. (2004) Cancer and the chemokine network. Nat Rev Cancer 4:540–550. 88. Weijer, K. and Hart, A.A. (1983) Prognostic factors in feline mammary carcinoma. J Natl Cancer Inst 70:709–716. 89. Misdorp, W. and Hart, A.A. (1979) Canine mammary cancer. II. Therapy and causes of death. J Small Anim Pract 20:395–404. 90. Rogers, K.S., Walker, M.A., et al. (1998) Transmissible venereal tumor: a retrospective study of 29 cases. J Am Anim Hosp Assoc 34:463–470. 91. Loh, R., Bergfeld, J., et  al. (2006) The pathology of devil facial tumor disease (DFTD) in Tasmanian devils (Sarcophilus harrisii). Vet Pathol 43:890–895. 92. Goedegebuure, S.A. (1979) Secondary bone tumours in the dog. Vet Pathol 16:520–529. 93. Trost, M.E., Inkelmann, M.A., et al. (2014) Occurrence of tumours metastatic to bones and multicentric tumours with skeletal involvement in dogs. J Comp Pathol 150:8–17. 94. Kas, N.P., van der Heul, R.O., et al. (1970) Metastatic bone neoplasms in dogs, cats and a lion (with some comparative remarks on the situation in man). Zentralbl Veterinarmed A 17:909–919. 95. Brodey, R.S., Reid, C.F., et al. (1966) Metastatic bone neoplasms in the dog. J Am Vet Med Assoc 148:29–43. 96. Waters, D.J., Honeckman, A., et al. (1998) Skeletal metastasis in feline mammary carcinoma: case report and literature review. J Am Anim Hosp Assoc 34:103–108. 97. Goldfinch, N. and Argyle, D.J. (2012) Feline lung‐digit syndrome: unusual metastatic patterns of primary lung tumours in cats. J Feline Med Surg 14:202–208. 98. May, C. and Newsholme, S.J. (1989) Metastasis of feline pulmonary carcinoma presenting as multiple digital swelling. J Small Anim Pract 30:302–310. 99. Finora, K. (2003) Common paraneoplastic syndromes. Clin Tech Small Anim Pract 18:123–126. 100. Argiles, J.M., Lopez‐Soriano, F.J., et  al. (2013) Mechanisms and treatment of ­cancer cachexia. Nutr Metab Cardiovasc Dis 23(suppl 1):S19–24.

Veterinary cancer incidence and molecular approaches to diagnosis and prognosis Genetics and cancer

Inherited tumor risk in dogs

Although it is not possible to accurately determine the number of animal cancers each year, several previous studies have attempted to determine overall incidence rates in various countries, especially for purebred dogs.1–5 Overall, these broad population‐based studies have highlighted those dog breeds that present with higher incidences of certain cancers. Such studies also reinforced the possible bias that may be introduced into regional surveys of pets by factors such as breed popularity, the impact of small population sizes, and diseases that impact breed longevity. There are indeed numerous purebred dogs that have an elevated incidence of certain cancers, compared to the overall purebred dog population, such as histiocytic tumors in Bernese mountain dogs, hemangiosarcoma in golden retrievers, keratoacanthomas in Kerry blue terriers,

c­ utaneous mast cell tumors in pugs, T‐cell lymphoma in boxers, urothelial carcinoma in Scottish terriers, and osteosarcoma in Irish wolfhound. The development of many breeds of modern dog over a relatively short period of time (200–300 years) was based primarily on generations of inbreeding and line breeding, intended to maximize ­conformity to a breed standard. Selective breeding for phenotypes resulted in a very broad range of morphological types (breeds) within which genetic variation has been reduced. Sampling of ­genetic variation across all breeds would provide a level comparable to that present across human populations.6 In individual breeds, however, the level of genetic diversity is variably restricted.7 The process of breed formation over just the past two centuries is estimated to have caused a seven‐fold greater reduction in genetic diversity than did the early domestication process, which lasted thousands of years.8 This is compounded further by the use of popular sires and gene pool decline during the twentieth century. As a result, numerous traits in purebred dogs are a consequence of variants in just a few genes. Since many of these phenotypes are characteristics of the particular breed, their presence in the breed had been positively selected, resulting in high frequency of the genes that cause them. With such intense selection, it is perhaps not surprising that there are now over 350 inherited diseases recognized in dogs and this number will expand annually. While some diseases have simple inheritance patterns, cancers are likely to be more complex. In some breeds the genetic background of the breed may predispose to a higher risk for specific cancers, or cancers in general. It seems likely certain breeds may have suppressor genes missing and they would have multiple different types of cancer, while others may have a promoter gene that enhances the development of a specific cancer. The absence of a suppressor gene would parallel what is seen in Li–Fraumeni syndrome in human families and may explain why certain breeds (e.g., golden retrievers) are at higher risk of developing several cancers compared to the general dog population. For other breeds the risk may be associated with one or more specific genes, where the “insult” is fixed in the genome and inherited, requiring just one or a few additional “insults” to promote cancer initiation. This may explain why we see certain breeds with an extraordinary high incidence of one cancer or closely related cancers (e.g., Bernese mountain dog and flat‐ coated retrievers are less common breeds but they have a high incidence of histiocytic tumors). Rottweilers have an increased incidence of several types of neoplasia and there is a germ line mutation in the proto‐oncogene MET, which encodes a tyrosine kinase in 70% of dogs of this breed but fewer than 5% of dogs of other breeds.9 Susceptibility to tumors has been traced to the family level in life‐ long studies of laboratory beagles in a pattern that is similar to those in some human families.10 There are inherited tendencies to develop melanomas in Sinclair and Hormel miniature pigs and Duroc‐Jersey swine.11 Although the specific genetic damage associated with the increased risk for tumors has been identified for some human ­families, inherited genetic abnormalities responsible for increased tumor susceptibilities in domestic animals are still emerging. From concepts to clinics: molecular genetics usher in an era of opportunity Improvements in our fundamental understanding of cancer biology present opportunities to lower the morbidity and mortality associated with this spectrum of diseases. For half a century, conventional cytogenetic approaches used to investigate human cancers provided

An Overview of Molecular Cancer    19

clues to the genetic basis of cancer. More recently, the introduction of molecular cytogenetics and other molecular genetic tools has revolutionized the way we are able to interrogate cancer cells to identify specific changes in chromosome and gene structure and/or function associated with cancer. Numerical chromosome changes (e.g., whole‐chromosome aneuploidy, insertions, deletions) represent a deviation from the normal gene copy number, potentially leading to increased or decreased expression of genes associated with regulation of growth or survival. Structural changes (e.g., inversions, translocations) result in genome reorganization, which may cause genes that are separated in the normal/healthy genome to be brought into close proximity in the tumor genome with ­consequent effects on gene dysregulation. Abnormal patterns of gene expression can also result from mutations that affect single genes. In fact, every tumor will have a multiplicity of mutations that will create unique patterns of gene expression and contribute to its pathogenesis. The potential impact of faster, less costly techniques for DNA and RNA sequencing in veterinary oncology should not be underestimated. Currently the costs of whole‐genome DNA sequencing and whole‐transcriptome RNA sequencing are still ­prohibitive in a clinical setting for routine evaluation of individual tumor specimens. However, the surge in research activity in these areas will undoubtedly identify genes of particular interest in the genomes of domestic animals for targeted analysis, leading to the emergence of potential new targets for therapy. Cytogenetics in cancer The DNA of all animals is packed into nature’s biological filing ­cabinets, chromosomes, the analysis of which is covered by the field of cytogenetics. In recent years the introduction of new molecular techniques and resources has led to the field of molecular cytogenetics, or cytogenomics, an area particularly suited to providing insights into the level of gross genome reorganization that occurs frequently in cancer cell populations. Using such techniques clonal chromosome aberrations have now been identified in over 65,000 cases of human cancer, representing over 70 different types of cancer (see http://cgap.nci.nih.gov/Chromosomes/Mitelman). Many of these recurrent chromosome aberrations were initially associated with histopathological or immunological subgroups, leading to their use as diagnostic signatures. In addition, the cytogenetic status of tumor cells is of established clinical value for prognosis, guiding therapy and assessing remission for a range of cancers, including ovarian cancer,12 colorectal carcinoma,13,14 gliomas,15 melanoma,16,17 and breast carcinoma.18 For example, in human leukemia patients, significant differences in the duration of treatment‐free interval are reported for different cytogenetic events or deletions of chromosome 17p and 11q.19 Detection of a RUNX1–RUNX1T1 translocation [t(8;21)(q22;q22)] in individuals with acute myeloid leukemia is a cytogenetic marker associated with a more favorable prognosis.20 However, acute ­myeloid leukemia patients with 8;21 translocation who also have a mutation of KIT or FLT3 have decreased survival times.21 In human lymphoma patients, the BCL6 transcriptional repressor (located at 3q27) is regarded as the most frequently involved oncogene in diffuse large B‐cell lymphoma (DLBCL). Chromosome translocations that involve BCL6 have been identified in up to 40% of people with DLBCL, but fewer than 10% of people with follicular lymphomas. The presence of a BCL6 translocation has been reported to have no prognostic significance for DLBCL,22 but in ­follicular lymphoma patients such events have been associated with an increased level of progression to DLBCL.23 When considering

treatment, early studies have suggested that a small molecule inhibitor that binds to the corepressor binding groove of the BCL6 BTB domain may be effective against BCL‐positive DLBCL.24 The ability to identify recurrent cytogenetic aberrations in ­cancers has been useful in helping to localize cancer‐associated genes. This approach has led to the selection of the most appropriate therapeutic approaches for patients and subsequent monitoring for recurrent disease. Cytogenetics has thus proven to be a key approach to improvements in the clinical management of patients, sparing patients with good prognosis from unnecessary treatment and, conversely, allowing patients whose cytogenetic abnormalities indicate poor prognosis to receive more aggressive treatments to improve the probability of positive outcomes. The World Health Organization recognizes that genetic abnormalities are one of the most reliable criteria for the classification of tumors and has stressed the importance of further research into this area. The increasing role of cytogenetics in the development of companion diagnostics and theranostics, which is the development of molecular diagnostic tests and targeted therapeutics in an interdependent, collaborative manner with the goals of individualizing treatment by targeting therapy to an individual’s specific disease subtype and genetic profile, is evident from recent studies. For example, it was discovered that in non‐small cell lung carcinoma (NSCLC), approximately 1 in 25 patients present with a chromosomal rearrangement that results in the fusion of the intracellular region of ALK (anaplastic lymphoma kinase) to the N‐terminal portion of EML4 (echinoderm microtubule‐associated protein‐like 4).25 This discovery led to the investigational use of ALK inhibitors to suppress the constitutive kinase activity in such patients. When treated with the selective MET/ALK inhibitor crizotinib, NSCLC patients with the ALK fusion gene had a 57% overall response rate and a >70% probability of having 6‐month progression‐free survival. The US Food and Drug Administration (FDA) have approved use of crizotinib for use in patients with late‐stage NSCLS where the presence of this fusion gene has been determined. The  FDA have also approved the use of a fluorescence in situ hybridization (FISH) assay to detect different ALK‐associated fusions as a companion diagnostic for crizotinib therapy. Recent progress in molecular cytogenomics of domestic animals (especially the dog) has allowed us to develop species‐specific ­“toolboxes” that will accelerate progress in our understanding of cancer genetics in these species. In addition, the use of comparative genomics is allowing the transfer of key genetic information across multiple species, leading to a greater impact. Although the application of cytogenomics technologies to animal cancers has begun to make an impact for the benefit of veterinary medicine, progress has been limited by a lack of appropriate patient samples associated with standardized therapy and detailed clinical follow‐up. For veterinary medicine to benefit from the full potential of clinical genomics, a greater level of collaboration between clinical and basic sciences is an essential prerequisite. The level of consistency of clinical management of cancer patients through tightly controlled clinical trials would provide the best opportunities to maximize progress towards the development of companion diagnostics/prognostics for veterinary health. Such clinical trials would ideally be conducted across multiple sites to enhance accrual rates. In the absence of sufficient resources to fully fund such trials, the next best alternative is the evaluation of biological specimens from patients with the same diagnosis, made using standardized diagnostic tests, and which are treated with a standard‐of‐care therapy. Sharing of clinical information, with appropriate informed consent of clients, would

20    Tumors in Domestic Animals

then help to expedite the path to improved outcome indicators. More veterinary clinicians need to recognize that opportunities to drive their profession towards the most appropriate treatment/care plans for their cancer patients tomorrow requires that they be open to providing clinical information and appropriate biological specimens from their patients today. This requirement is already evident in the increasing number of multi‐institutional clinical trials, especially through the Comparative Oncology Trials Consortium (COTC), and in the standardized collection of biological specimens from cancer‐bearing dogs by the Canine Comparative Oncology and Genomics Consortium (www.CCOGC.org). At the gross chromosome level, numerous studies have identified cytogenetic aberrations detected in a range of canine n ­ eoplasms,26–37 including hematopoietic malignancies,26–28 intracranial malignancies,29 osteosarcoma,31,32 hemangiosarcoma,33 histiocytic malignancies,34 urothelial carcinoma,35 melanoma,36 leukemia,37 and mast cell tumors (Mochizuki et al., unpublished). In addition, cytogenetic characterization of canine cancer cell lines has been used to compare their status to the primary disease they are reported to represent.36,38–41,79 (Poorman et  al., unpublished). Early work has also reported on c­ytogenetic changes evident in feline sarcomas42 and intestinal ­lymphoma (Thomas et al., unpublished). As the studies above have shown, the increasing use of molecular tools, including arrays comprising thousands of genomic features, either DNA or RNA sequences, are now being used to analyze genome‐wide patterns of changes to DNA content and transcript abundance. In combination, these approaches maximize the efficiency with which we can identify genetic alterations associated with a specific diagnosis and also provide insight into tumor pathogenesis. Assessment of genome‐wide DNA copy number aberrations has been performed in numerous cancer types, in both domestic dogs and cats, as discussed above. In parallel, characterization of the level of transcriptional activity of genes, either via a genome‐wide or a targeted approach, has begun to provide key signatures relating to diagnosis and prognosis in canine cancers.20,43,44 Furthermore, even though about 90% of the genome does not encode proteins, it nevertheless has important functions in maintaining homeostasis. At the turn of the century, noncoding DNA was discarded as mere “junk DNA,” thought to be an anachronism inherited from our evolutionary forebears. We now know that this DNA in fact encodes molecules, such as microRNAs, that have important functions in gene regulation.45,46 Indeed, gain or loss of function of microRNAs may turn out to be just as important in ­cancer causation as gain or loss of function of traditional protein‐ coding genes.46–54 Molecular diagnostics and prognostics Molecular diagnostic testing entered the medical arena with the availability of assays to detect the presence of infectious agents. Although this remains the largest segment of the molecular diagnostics market, molecular testing in oncology is a rapidly growing area. Development of molecular‐based assays to aid in cancer diagnosis and prognosis indicates that they will soon be considered as conventional as morphologic approaches to cancer diagnosis. Access to an increasing portfolio of molecular assays for cancers will have a profound impact on patient care and lead to the need for a more interdisciplinary approach to decision‐making. Although a variety of biological techniques are considered as molecular diagnostics, all are based on the detection and analysis of either specific sequences of nucleic acid (DNA or RNA) or proteins. For

diagnostic purposes the regions analyzed are associated specifically with the presence of a disease or subtype. To provide prognostic value, the regions assessed need to have characteristics that are associated with differing clinical progression and outcomes. In the area of oncology, molecular diagnostics based on DNA include identification of large numerical and/or structural changes to genome organization; ranging from whole‐chromosome copy number changes (aneuploidy), partial‐chromosome copy number change (segmental aneuploidy; deletions, duplications, amplifications), or chromosome rearrangements (translocations), down to changes that affect perhaps just one base of DNA. When considering RNA, a diagnostic parameter may be a specific level of transcription of a gene (or multiple genes) in a cancer subtype or more likely a multi‐gene “signature.” If such a signature is associated with clinical progression and outcome, it may also be considered to be of prognostic value. Molecular approaches to cancer patient management will become a key tool, with assays for cancer prediction, diagnosis, and prognosis becoming intertwined temporally to: (1) identify patients at high risk, (2) provide early detection of a cancer, (3) select the most efficacious therapy options, and (4) spot early signs of relapse. Complementary assays, using cytogenetics, immunohistochemistry, and gene expression platforms, are now in place to determine the status of an individual tumor. Since 2007, the College of American Pathologists and the American Society of Clinical Oncology have provided benchmark guidelines to ensure consistency of reporting across testing laboratories.55 The association of molecular signatures with the patient’s response to therapy is leading to the emergence of companion diagnostics, assays that can inform the clinician of the chances that a specific therapy will be effective for an individual patient (theranostics), based on known efficacy of treating patients/cancers with the same signature. For  example, once it has been determined that a human breast ­carcinoma is overexpressing Her2, ­evidence‐based efficacy data suggest that such c­ ancers will respond favorably to trastuzumab (Herceptin). In general, the development of a companion diagnostic requires close cooperation between the providers of the molecular assay and the therapy. There are numerous stakeholders in the field of cancer patient care (clinician, pathologist, molecular assay developers, molecular assay technologist, pharmaceutical company), each with their own ­ challenges. Communication ­between these is key to ensure that the ultimate stakeholder, the cancer patient, receives the most appropriate care to optimize the quality and duration of their life. Evolutionarily conserved genomic changes in cancers Perhaps the most widely investigated chromosome aberration associated with cancers in people is the Philadelphia chromosome, first described almost half a century ago in patients with CML.56,57 This aberrant human chromosome (HSA) is the result of a translocation event that brings together the c‐abl oncogene (located at HSA 9q34 (ABL locus)) and the breakpoint cluster region (BCR) (located at HSA 22q11) to form a derivative human chromosome 22, technically described as t(9;22)(q34;q11) and referred to as the Philadelphia (Ph) chromosome.57 The juxtaposition of BCR and ABL is considered a hallmark feature of CML, reported in over 95% of CML patients.58 The biological consequence of the generation of this fusion is elevation of tyrosine kinase activity, which results in the uncontrolled proliferation of white (predominantly myeloid)

An Overview of Molecular Cancer    21

blood cells. The identification of STI571 (imatinib mesylate) as a compound that acts as an antagonist to this fusion protein (bcr‐abl tyrosine kinase) and prevents blast crisis,59,60 led to clinical trials and the development of Gleevec®,58 which (with some exceptions) is now generally considered standard of care for patients with the Philadelphia chromosome. Almost 90% of patients treated with Gleevec are free of disease worsening, with an estimated overall survival rate of 91%. A cytogenetic response, defined as a reduction of cells with the characteristic molecular abnormality, is seen in up to 60% of patients61,62 and remains an important surrogate marker of survival in human CML patients.63,64 Cytogenetic testing is used to initially diagnose the CML and then subsequently to monitor remission and identify any elevation in the number of cells ­harboring BCR‐ABL during relapse/recurrence. Although very rare in veterinary species, CML has been reported in dogs and conveys a poor prognosis.27,65–67 A study of canine CML showed that dogs diagnosed with CML also presented with a functionally active BCR‐ABL translocation.27 These data suggest that, cost aside, treatment with Gleevec, or a similar TKI compound could be an option for therapy of canine CML assuming any t­ oxicity issues are overcome. This study resulted in the first molecular ­cytogenetic test for the presence of a clinically significant genomic alteration in a veterinary cancer and has since been used to identify the Raleigh chromosome in numerous additional cases presenting with suspected CML. In addition, cases presenting with the Raleigh chromosome (Figure  1.6) have been followed during treatment with various compounds and the cell population containing the BCR‐ABL event was almost cleared from detection during remission, and then returned at relapse.27,33,68 These data demonstrate that appropriate therapies do have the desired impact on the cancer cell population, opening the door for broader studies of treatment ­efficacy in animal cancers, determined by molecular as well as ­conventional clinical response. The presence of RB1 deletions in canine patients presenting with chronic lymphocytic leukemias and MYC‐IgH translocations in canine patients diagnosed with Burkitt lymphoma have also been reported, supported by functional data.27 These findings reinforce the concept that as mammals, humans and dogs may be ­considered temporally separated, differential organizations of the same collection of ancestrally related genes. Since we have shown that genetic “lesions” associated with human cancers may be similarly associated in cancers of veterinary species, therapies developed for malignancies with specific cytogenetic signatures in human cancers may become applicable to provide improved treatments for cancers in our pet dogs and cats. Cytogenomic screening of cancers in our pets could become common practice in veterinary oncology, used to aid diagnosis, selection of the most appropriate therapy, monitoring of residual disease, and for ­ prognostication. Molecular assays in veterinary oncology The availability of high‐quality genome sequences for the domestic dog and cat laid the foundations for the development of a series of new resources for cancer research for both species. Recent studies using genomics have led to the identification of inherited genetic risk factors associated with canine cancers, cytogenomic changes associated with specific diagnosis and/or prognosis of cancer in dogs and cats, and a series of new molecular tests that will provide new aids to diagnosis and clinical management of pets diagnosed with cancer. Examples of such assays are described below.

PARR (PCR for antigen receptor rearrangement)

PARR is a polymerase chain reaction (PCR)‐based assay developed to detect clonal expansion of lymphocytes in dogs and cats suspected of having lymphoma. The concept of PARR is based on the assumption that the lymphoid neoplasm is the result of clonal expansion of B or T lymphocytes. Ideally, the DNA is isolated from specimens representing the lesion, including aspirates (bone marrow or lymph node), incisional and excisional biopsies, provided that they yield a sufficient amount of cellular DNA. In certain circumstances DNA may be isolated from peripheral blood of the lymphoma patient, but this approach is to be treated with caution. Material to be tested is usually transferred to the laboratory fresh or frozen, but may also be formalin‐fixed, paraffin‐ embedded (FFPE) tumor specimens, or cytologic preparations on glass slides. Each sample is processed to obtain cellular DNA, which serves as the template for PCR analysis. The assay uses PCR primers specific to immunoglobulin antigen receptors in B cells and T‐cell receptors in T cells.69 Rearrangement of these genes is a natural part of lymphocyte differentiation, resulting in polyclonal cell population. With a clonal expansion of cells the usual multitude of rearrangements present in these genes is eradicated, generating a single PCR amplicon. Detection of a monoclonal proliferation strongly favors a diagnosis of lymphoid neoplasia, while a polyclonal proliferation supports hyperplasia. As with all assays it is important to be aware of the limitations of detection. For dogs, PARR is reported to be at least 90% specific, but has only 75% sensitivity for lymphoid neoplasia.70–72 In cats, the specificity is similar, but the sensitivity is reduced to 65%. These data mean that PARR will miss 25% and 35% of actual lymphoma cases in dogs and cats, respectively. Approximately 10% of cases with positive test results (clonality) do not have lymphoma (false positives). Monoclonal proliferation with inflammatory or hyperplastic diseases is uncommon but has been reported with feline infectious peritonitis and canine ehrlichisosis. As with all tests, PARR is not 100% specific or sensitive and results of PARR need to be correlated with clinical signs, cytology, histopathology, and other test results. However, with access to just a fine‐needle aspirate of an enlarged node, PARR is an option for clinicians to determine clonality which would be indicative of lymphoma or to assess recurrence of lymphoma following treatment. Molecular clonality is not a primary or sole diagnostic test and it is not needed when the results of histology or cytology and/or immunophenotyping are definitive. However, determination of clonality is  useful when morphology and immunophenotyping are not definitive.

Cytogenetic assay for prognosis in canine lymphoma

Lymphoma is estimated to affect in excess of 250,000 pet dogs each year in the United States, and is one of the most common canine cancers. In a recent survey of over 150 veterinary oncologists in the United States (Breen, unpublished) the three most common reasons for dog owners hesitating to opt for treatment of the lymphoma are (1) concerns over whether their dog will be even more sick during chemotherapy, (2) the cost of treatment, and (3) uncertainty of outcome. Currently, it is widely accepted that up to 90% of all canine multicentric lymphomas will enter remission if treated with standard‐of‐care chemotherapy, and median survival is 9–12 months. In addition, most B‐cell lymphoma cases are expected to have a longer survival time than most T‐cell cases, but there are exceptions. A review of current indicators for survival has been provided by Valli et al.73

22    Tumors in Domestic Animals

To assist oncologists and owners in making a more informed decision about treatment options and potential outcomes we developed a cytogenetic assay to help predict the duration of remission with two different treatments, either single‐agent doxorubicin or multi‐agent CHOP (cyclophosphamide, hydrodoxorubicin, oncovin/vincristine, prednisone) therapy. Genome‐wide DNA copy number profiling of diagnostic biopsy specimens from dogs diagnosed with multicentric lymphoma revealed numerous recurrent aberrations, including whole‐chromosome aneuploidy and segmental aneuploidy.26,28 In a parallel study to evaluate the prognostic significance of these copy number changes we evaluated patients from two cohorts; the first cohort comprised 160 FFPE diagnostic biopsy specimens from dogs with confirmed lymphoma (any subtype). In all cases the biopsy was taken prior to any treatment and then the dog was subsequently treated with single‐agent doxorubicin. In the second cohort, 100 FFPE diagnostic lymph node biopsy specimens were obtained from dogs diagnosed with multicentric lymphoma and prior to treatment. For this cohort, all dogs were subsequently treated with standard‐of‐care multi‐agent CHOP therapy. Both cohorts comprised dogs with B‐ and T‐cell lymphoma and of various subtypes. For both cohorts, the dogs were clinically evaluated at regular intervals during and after their chemotherapy, and the duration of their first remission recorded. Using cells obtained from the FFPE specimens interphase nuclei were screened using multicolor FISH analysis to determine the mean copy number of selected regions of the canine genome (Figure 1.13). In the doxorubicin‐treated patients, the mean copy number of two regions of the genome (located on dog chromosomes 1 and 6) correlated significantly with the duration of first remission in a positive linear relationship; dogs with low mean copy number of both regions had shorter first remission times than those with higher mean copy number. In the CHOP cohort, the region on chromosome 1 was not associated with the duration of first remission, while the mean copy number of the region on chromosome 6 remained significantly associated. These data were used to develop a molecular cytogenetic assay in which both regions are evaluated simultaneously in cells derived from lymph node samples and the data used to provide a predicted duration of remission if a dog is subsequently

A

treated with either single‐agent doxorubicin or multi‐agent CHOP therapy. This assay should be widely available in 2017.

Cytogenetic assay to separate histiocytic malignancies from lymphoma

Histiocytic neoplasms, benign and malignant, arise primarily from dendritic cells found in the skin and visceral organs. The incidence of all histiocytic malignancies is rare in the general dog population, but remarkably high in several purebred dogs, including the Bernese mountain dog, flat‐coated retriever, rottweiler, and golden retriever. Malignant tumors of histiocytic origin generally have a very poor prognosis (typical survival is just a few weeks post diagnosis) and are considered generally unresponsive to current therapeutic options. In the Bernese mountain dog, 66% of deaths are reported to be due to cancer,74 of which 47% are attributed to histiocytic malignancies, with a further 29% due to lymphoma.74 These data indicate that, strikingly, 75% of all cancers and 50% of deaths in this one breed are due to just these two cancers. Correct diagnosis of a histiocytic neoplasm currently requires specialized immunohistochemistry (IHC). However, IHC is not always readily available, as specific antibodies are required and special tissue ­preparation such as frozen sections are required for some of the antibodies. IHC can be time consuming and requires a particular skill set. The ability to accurately distinguish between canine lymphoma and histiocytic malignancies is an important determinant of treatment and outcome; the prognosis for lymphoma is better than that of disseminated histiocytic sarcoma. Genome‐wide evaluations of DNA copy number in canine lymphoma and histiocytic malignancies across a range of breeds have led to the identification of regions of the canine genome that are aberrant in one of these two cancers but not the other. Specifically, histiocytic neoplasms present with a high frequency of deletion of dog chromosomes 2, 16, and 31; none of these deletions are evident in canine lymphoma. In addition, dog chromosome 31 is frequently increased in copy number in lymph node cells of confirmed lymphoma patients. Using these features, a cytogenetic assay was developed to simultaneously assess the mean copy number status of regions of dog chromosomes 2, 16, and 31 (Figure  1.14). The assay has 97.2%

B

Figure 1.13  Lymphoma cytogenetic prognostic assay. Multicolor FISH of canine interphase nuclei of cells aspirated from lymph nodes of (A) a healthy dog

and (B) a dog with lymphoma. Enumeration of the five differentially labeled single locus probes indicates that in (A) all five have a normal copy number of 2, while in (B) the two probes labeled in red and aqua (arrows) both have an abnormal copy number of 3. Probe enumeration in 100 cells allows derivation of mean copy number value for each probe. With standard‐of‐care doxorubicin‐based chemotherapy for lymphoma, 95% of dogs with low mean copy number (100 cells yields mean copy numbers for each probe of 90% of histiocytic neoplasms, while >90% of canine lymphomas a have a balanced or mean copy number >2.0 for each probe.

specificity and 97.3% sensitivity to distinguish between histiocytic neoplasia and lymphoma. The significance of this assay for veterinarians lies in the ability to readily distinguish these two types of cancer, especially for those breeds that are at high risk of developing both cancers and/or where there is an uncertain diagnosis based on morphology alone. In combination with an assay to predict duration of remission to chemotherapy in canine lymphoma patients treated with standard of care, such an assay has the potential to offer considerable value to patient management, adding new approaches to refine diagnosis, and even prognosis. This assay, available since 2015, is being extended to confirm the presence of a histiocytic malignancy and exclude other round cell neoplasms.

Diagnostic assay for urothelial carcinoma

Urothelial carcinoma, also referred to as transitional cell carcinoma (TCC), is the most common bladder neoplasm in the dog, although compared to other cancers it is uncommon, accounting for 99%. Similar approaches are being used to develop additional cytogenetic assays designed to provide diagnostic and prognostic information for a range of canine cancers, including, for example, canine leukemia subtypes, mast cell tumors, osteosarcoma, oral melanoma, intracranial tumors, and hemangiosarcoma. In addition, studies looking at the cytogenomics of various feline cancers are leading to assays that will provide new tools to aid management of cats diagnosed with injection site sarcoma, gastrointestinal lymphoma/inflammatory bowel disease and mammary carcinoma.42 Conclusions In recent years, remarkable progress has been made in our understanding of the complex pathogenesis of neoplasia. The molecular mechanisms involved in the neoplastic transformation and regulation of cells have been identified for numerous tumor types. This is beginning to be applied to risk assessment, tumor diagnostics, and anticancer therapy. It is hoped that this new understanding will permit more precise identification of the early stages of neoplasia when, it is presumed, therapy can be more effective. Therapies that can be developed to specifically target abnormal properties of cancer cells may spare normal cells and may avoid the side effects of many contemporary treatments. Moreover, as the ­genetic lesions responsible for cancer development and progression are identified, conventional diagnostic techniques and grading algorithms will, it is hoped, be complemented by stronger predictors of outcome. Currently, with few exceptions the gold standard for a diagnosis and determining malignancy remains histologic diagnosis. As molecular studies of animal cancers are pursued, and data become more widely accessible, it is to be expected that new signatures of

malignancy will emerge to aid in the determination of a malignant versus benign phenotype. Provision of an accurate diagnosis remains the key benchmark for the pathologist. However, pathologists are ­seeing only a snapshot in the temporal course of any cancer. Tumors are dynamic yet we may only sample at one point in time and the diagnosis of a benign mass may progress to one of malignancy. Evaluations will need to consider the timeline of transformative events to leverage comprehensive information of patient samples. Detection and quantification of such events will accelerate our understanding of the biological significance of neoplasms, whether assessed through a microscope or molecular means. Whether considering the natural course of the disease, response to primary therapy, or additional response to rescue therapy, the veterinary profession needs to engage in a coordinated way to tackle animal cancers as a team.

References

1. Dobson, J.M., Samuel, S., et al. (2002) Canine neoplasia in the UK: estimates of incidence rates from a population of insured dogs. J Small Anim Pract 43:240–246. 2. Adams, V.J., Evans, K.M., et al. (2010) Methods and mortality results of a health survey of purebred dogs in the UK. J Small Anim Pract 51:512–524. 3. Dorn, C.R., Taylor, D.O., et al. (1968) Survey of animal neoplasms in Alameda and Contra Costa Counties, California. II. Cancer morbidity in dogs and cats from Alameda County. J Natl Cancer Inst 40:307–318. 4. Bronden, L.B., Nielsen, S.S., et al. (2010) Data from the Danish Veterinary Cancer Registry on the occurrence and distribution of neoplasms in dogs in Denmark. Vet Rec 166:586–590. 5. Priester, W.A., McKay, F.W. (1980) The occurrence of tumors in domestic animals. Natl Cancer Inst Monogr 1–210. 6. Lindblad‐Toh, K., Wade, C.M, et al. (2005) Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature 438:803–819. 7. Vonholdt, B.M., Pollinger, J.P., et  al. (2010) Genome‐wide SNP and haplotype analyses reveal a rich history underlying dog domestication. Nature 464:898–902. 8. Gray, M.M., Granka, J.M., et al. (2009) Linkage disequilibrium and demographic history of wild and domestic canids. Genetics 181:1493–1505. 9. Liao, A.T., McMahon, M., et  al. (2006) Identification of a novel germline MET mutation in dogs. Anim Genet 37:248–252. 10. Schafer, K.A., Kelly, G., et al. (1998) A canine model of familial mammary gland neoplasia. Vet Pathol 35:168–177. 11. Oxenhandler, R.W., Adelstein, E.H., et  al. (1979) Malignant melanoma in the Sinclair miniature swine: an autopsy study of 60 cases. Am J Pathol 96:707–720.

An Overview of Molecular Cancer    25

12. Pejovic, T., Himmelmann, A., et  al. (1992) Prognostic impact of chromosome aberrations in ovarian cancer. Br J Cancer 65:282–286. 13. Bardi, G., Fenger, C., et al. (2004) Tumor karyotype predicts clinical outcome in colorectal cancer patients. J Clin Oncol 22:2623–2634. 14. Brosens, R.P., Belt, E.J., et al. (2011) Deletion of chromosome 4q predicts outcome in stage II colon cancer patients. Cell Oncol (Dordr) 34:215–223. 15. Wemmert, S., Ketter, R., et al. (2005) Patients with High‐grade gliomas harboring deletions of chromosomes 9p and 10q benefit from temozolomide treatment. Neoplasia 7:883–893. 16. Ganguly, A., Richards‐Yutz, J., et al. (2014) Molecular karyotyping for detection of prognostic markers in fine needle aspiration biopsy samples of uveal melanoma. Methods Mol Biol 1102:441–458. 17. Hirsch, D., Kemmerling, R., et al. (2013) Chromothripsis and focal copy number alterations determine poor outcome in malignant melanoma. Cancer Res 73:1454–1460. 18. A’Hern, R.P., Jamal‐Hanjani, M., et al. (2013) Taxane benefit in breast cancer – a role for grade and chromosomal stability. Nat Rev Clin Oncol 10:357–364. 19. Giertlova, M., Hajikova, M., et  al. (2011) Cytogenetic abnormalities predict treatment‐free interval and response to therapy in previously untreated chronic lymphocytic leukemia patients. Neoplasma 58:82–88. 20. Li, M.M., Ewton, A.A., et al. (2013) Using cytogenetic rearrangements for cancer prognosis and treatment (pharmacogenetics). Curr Genet Med Rep 1:99–112. 21. Vardiman, J.W., Thiele, J., et  al. (2009) The 2008 Revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood 114:937–951. 22. Iqbal, J., Greiner, T.C., et al. (2007) Distinctive patterns of BCL6 molecular alterations and their functional consequences in different subgroups of diffuse large B‐cell lymphoma. Leukemia 21:2332–2343. 23. Akasaka, T., Lossos, I.S., et al. (2003) BCL6 gene translocation in follicular lymphoma: a harbinger of eventual transformation to diffuse aggressive lymphoma. Blood 102:1443–1448. 24. Cerchietti, L.C., Ghetu, A.F., et al. (2010) A small‐molecule inhibitor of BCL6 kills DLBCL cells in vitro and in vivo. Cancer Cell 17:400–411. 25. Soda, M., Choi, Y.L., et  al. (2007) Identification of the transforming Em14‐Alk fusion gene in non‐small‐cell lung cancer. Nature 448:561–566. 26. Nambiar, P.R., Haines, D.M., et al. (2000) Mutational analysis of tumor suppressor gene p53 in feline vaccine site‐associated sarcomas. Am J Vet Res 61:1277–1281. 27. Breen, M. and Modiano, J.F. (2008) Evolutionarily conserved cytogenetic changes in hematological malignancies of dogs and humans – man and his best friend share more than companionship. Chromosome Res 16:145–154. 28. Thomas, R., Smith, K.C., et al. (2003) Chromosome aberrations in canine multicentric lymphomas detected with comparative genomic hybridisation and a panel of single locus probes. Br J Cancer 89:1530–1537. 29. Thomas, R., Duke, S.E., et  al. (2009) ‘Putting our heads together’: insights into genomic conservation between human and canine intracranial tumors. J Neurooncol 94:333–349. 30. Hershey, A.E., Dubielzig, R.R., et  al. (2005) Aberrant P53 expression in feline vaccine‐associated sarcomas and correlation with prognosis. Vet Pathol 42:805–811. 31. Thomas, R., Wang, H.J., et al. (2009) Influence of genetic background on tumor karyotypes: evidence for breed‐associated cytogenetic aberrations in canine appendicular osteosarcoma. Chromosome Res 17:365–377. 32. Karlsson, E.K., Sigurdsson, S., et al. (2013) Genome‐wide analyses implicate 33 loci in heritable dog osteosarcoma, including regulatory variants near Cdkn2a/B. Genome Biol 14:R132. 33. Thomas, R., Borst, L., et al. (2014) Genomic profiling reveals extensive heterogeneity in somatic DNA copy number aberrations of canine hemangiosarcoma. Chromosome Res 22:305–319. 34. Hedan, B., Thomas, R., et  al. (2011) Molecular cytogenetic characterization of canine histiocytic sarcoma: a spontaneous model for human histiocytic cancer identifies deletion of tumor suppressor genes and highlights influence of genetic background on tumor behavior. BMC Cancer 11:201. 35. Shapiro, S.G., Raghunath, S., et al. (2015) Canine urothelial carcinoma: genomically aberrant and comparatively relevant. Chromosome Res 23:311–331. 36. Poorman, K., Borst, L., et al. (2015) Comparative cytogenetic characterization of primary canine melanocytic lesions using array CGH and fluorescence in situ hybridization. Chromosome Res 23:171–186. 37. Roode, S.C., Rotroff, D., et  al. (2015) Genome‐wide assessment of recurrent genomic imbalances in canine leukemia identifies evolutionarily conserved regions for subtype differentiation. Chromosome Res DOI:10.1007/s10577-015-9475-7 38. Seiser, E.L., Thomas, R., et al. (2013) Reading between the lines: molecular characterization of five widely utilized canine lymphoid tumor cell lines. Vet Comp Oncol 11:30–50. 39. Kisseberth, W.C., Nadella, M.V., et  al. (2007) A novel canine lymphoma cell line: a translational and comparative model for lymphoma research. Leuk Res 31:1709–1720.

40. Thudi, N.K., Shu, S.T., et  al. (2011) Development of a brain metastatic canine prostate cancer cell line. Prostate 71:1251–1263. 41. Simmons, J.K., Dirksen, W.P., et  al. (2014) Canine prostate cancer cell line (Probasco) produces osteoblastic metastases in vivo. Prostate 74:1251–1265. 42. Thomas, R., Valli, V.E., et al. (2009) Microarray‐based cytogenetic profiling reveals recurrent and subtype‐associated genomic copy number aberrations in feline ­sarcomas. Chromosome Res 17:987–1000. 43. Ito, D., Frantz, A.M., et al. (2012) Cd40 ligand is necessary and sufficient to support primary diffuse large B‐cell lymphoma cells in culture: a tool for in vitro preclinical studies with primary B‐cell malignancies. Leuk Lymphoma 53:1390–1398. 44. Scott, M.C., Sarver, A.L., et al. (2011) Molecular subtypes of osteosarcoma identified by reducing tumor heterogeneity through an interspecies comparative approach. Bone 49:356–367. 45. Johnson, C.D., Esquela‐Kerscher, A., et al. (2007) The Let‐7 microRNA represses cell proliferation pathways in human cells. Cancer Res 67:7713–7722. 46. Esquela‐Kerscher, A. and Slack, F.J. (2006) Oncomirs – microRNAs with a role in cancer. Nat Rev Cancer 6:259–269. 47. McManus, M.T. (2003) MicroRNAs and cancer. Semin Cancer Biol 13:253–258. 48. Balkwill, F. (2004) Cancer and the chemokine network. Nat Rev Cancer 4:540–550 49. Calin, G.A. and Croce, C.M. (2006) Genomics of chronic lymphocytic leukemia microRNAs as new players with clinical significance. Semin Oncol 33:167–173. 50. Hammond, S.M. (2006) MicroRNAs as oncogenes. Curr Opin Genet Dev 16:4–9. 51. von Deetzen, M.C., Schmeck, B., et al. (2013) Molecular quantification of canine specific microRNA species. Res Vet Sci 95:562–568. 52. Uhl, E., Krimer, P., et al. (2011) Identification of altered microRNA expression in canine lymphoid cell lines and cases of B‐ and T‐cell lymphomas. Genes Chromosomes Cancer 50:950–967. 53. Noguchi, S., Mori, T., et  al. (2012) Comparative study of anti‐oncogenic microRNA‐145 in canine and human malignant melanoma. J Vet Med Sci 74:1–8. 54. Gioia, G., Mortarino, M., et  al. (2011) Immunophenotype‐related microRNA expression in canine chronic lymphocytic leukemia. Vet Immunol Immunopathol 142:228–235. 55. Wolff, A.C., Hammond, M.E., et al. (2007) American Society of Clinical Oncology/ College of American Pathologists Guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 25:118–145. 56. Nowell, P.C. and Hungerford, D.A. (1960) A minute chromosome in human chronic granulocytic leukemia. Science 132:1497. 57. Rowley, J.D. (1973) Letter: A new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and giemsa ­staining. Nature 243:290–293. 58. Kurzrock, R., Kantarjian, H.M., et  al. (2003) Philadelphia chromosome‐positive leukemias: from basic mechanisms to molecular therapeutics. Ann Intern Med 138:819–830. 59. Mauro, M.J. and Druker, B.J. (2001) Chronic myelogenous leukemia. Curr Opin Oncol 13:3–7. 60. Mauro, M.J., O’Dwyer, M.E., et al. (2001) St1571, a tyrosine kinase inhibitor for the treatment of chronic myelogenous leukemia: validating the promise of molecularly targeted therapy. Cancer Chemother Pharmacol 48(suppl 1):S77–78. 61. Kantarjian, H.M. and Talpaz, M. (2001) Imatinib mesylate: clinical results in Philadelphia chromosome‐positive leukemias. Semin Oncol 28:9–18. 62. Hernandez‐Boluda, J.C. and Cervantes, F. (2002) Imatinib mesylate (Gleevec, Glivec): a new therapy for chronic myeloid leukemia and other malignancies. Drugs Today (Barc) 38:601–613. 63. Deininger, M.W. (2003) Cytogenetic studies in patients on imatinib. Semin Hematol 40:50–55. 64. Rosti, G., Testoni, N., et al. (2003) The cytogenetic response as a surrogate marker of survival. Semin Hematol 40:56–61. 65. Tarrant, J.M., Stokol, T., et al. (2001) Diagnosis of chronic myelogenous leukemia in a dog using morphologic, cytochemical, and flow cytometric techniques. Vet Clin Pathol 30:19–24. 66. Leifer, C.E., Matus, R.E., et al. (1983) Chronic myelogenous leukemia in the dog. J Am Vet Med Assoc 183:686–689. 67. Pollet, L., Van Hove, W., et  al. (1978) Blastic crisis in chronic myelogenous ­leukaemia in a dog. J Small Anim Pract 19:469–475. 68. Perez, M.L., Culver, S., et al. (2013) Partial cytogenetic response with toceranib and prednisone treatment in a young dog with chronic monocytic leukemia. Anticancer Drugs 24:1098–1103. 69. Keller, R.L., Avery, A.C., et  al. (2004) Detection of neoplastic lymphocytes in peripheral blood of dogs with lymphoma by polymerase chain reaction for antigen receptor gene rearrangement. Vet Clin Pathol 33:145–149. 70. Thalheim, L., Williams, L.E., et al. (2013) Lymphoma immunophenotype of dogs determined by immunohistochemistry, flow cytometry, and polymerase chain reaction for antigen receptor rearrangements. J Vet Intern Med 27:1509–1516.

26    Tumors in Domestic Animals

71. Moore, P.F., Woo, J.C., et  al. (2005) Characterization of feline T cell receptor gamma (Tcrg) variable region genes for the molecular diagnosis of feline intestinal T cell lymphoma. Vet Immunol Immunopathol 106:167–178. 72. Burnett, R.C., Vernau, W., et  al. (2003) Diagnosis of canine lymphoid neoplasia using clonal rearrangements of antigen receptor genes. Vet Pathol 40:32–41. 73. Valli, V.E., Kass, P.H., et al. (2013) Canine lymphomas: association of classification type, disease stage, tumor subtype, mitotic rate, and treatment with survival. Vet Pathol 50:738–748. 74. Bernese Mountain Dog Club of American Health Survey (2005) Available from: http://www.bmdca.org/health/pdf/2005_Health_Survey.pdf 75. Mutsaers, A.J., Widmer, W.R., et al. (2003) Canine transitional cell carcinoma. J Vet Intern Med 17:136–144.

76. Norris, A.M., Laing, E.J., et  al. (1992) Canine bladder and urethral tumors: a ­retrospective study of 115 cases (1980–1985). J Vet Intern Med 6:145–153. 77. Henry, C.J., Tyler, J.W., et al. (2003) Evaluation of a bladder tumor antigen test as a screening test for transitional cell carcinoma of the lower urinary tract in dogs. Am J Vet Res 64:1017–1020. 78. Vinall, R.L., Kent, M.S., et al. (2012) Expression of microRNAs in urinary bladder samples obtained from dogs with grossly normal bladders, inflammatory bladder disease, or transitional cell carcinoma. Am J Vet Res 73:1626–1633. 79. Shapiro, S., Knapp, D., and Breen, M. (2015) A cultured report to urothelial carcinoma, molecular characterization of five cell lines. Canine Genet Epidemiol 2:15.

2

Trimming Tumors for Diagnosis and Prognosis Paul C. Stromberg1 and Donald J. Meuten2 1 2

The Ohio State University, USA North Carolina State University, USA

Increases in the complexity of oncology patient management, growing sophistication of veterinary medicine, and elevated client expectations have combined to place more demands on veterinary pathologists than just making a diagnosis. What is a simple matter of tumor diagnosis in autopsies would, in many cases, be inadequate in the surgical pathology arena. Add to this the limited or often complete lack of communication between the clinician collecting the tumor tissue, the technician trimming it into cassettes, and the pathologist reading the cases and formulating the diagnosis and the potential for an unsatisfactory outcome is substantial. Although the most impor­ tant item of information in any assessment of tumors is the diagnosis, in the surgical biopsy arena it is often equally important to note addi­ tional information that may impact the case management, such as tumor grade (if one exists) and completeness of the resection.1 To ensure a diagnosis is possible, representative tissue must be collected, properly preserved and processed. With small masses that fit into a standard 2 cm × 2.5 cm × 5 mm processing cassette, this is not a problem. However, many tumor specimens exceed this size and must be trimmed to fit within those dimensions. It is this critical step which ensures that both diagnostic tissue and the surgical margins are included in the biopsy and properly oriented in the cassette so that the pathologist can complete the evaluation, make a diagnosis and accurately assess the surgical margin so the clinician can formu­ late a prognosis and the appropriate treatment plan. The surgical margin can be defined and reported in several ways. The goal is to provide the clinicians with the information they need to decide what, if any, additional treatments will be recommended.

Specimen sizing

In general, the likelihood of getting a diagnosis from a biopsy is directly proportional to the amount of the tumor tissue ­evaluated. It is highest when 100% of the tumor can be examined. The

optimal specimen is one that fits into a single or several tissue‐ processing cassettes, permitting evaluation of almost all the tissue in the sample. This is common in small skin tumors. But often the entire tumor cannot be sampled by the surgeon or trimmed in by the pathologist. Tumor sampling by the surgeon can be divided into those that completely remove the entire mass (excisional biopsy) (Figure 2.1) and those that remove only a portion of the mass (­incisional biopsy). The choice is the surgeon’s and is often driven by the first rule of medicine (“First do no harm”), client preferences, and cost. Complete removal (total excisional biopsy) is the most desirable option because both diagnosis and treatment can be accomplished in one procedure. A desirable metric of excisional biopsies is a margin evaluation to assess the complete­ ness of the excision. When tumor location and patient concerns dictate only a portion of the mass can be sampled, an incisional biopsy (Figure  2.2) is collected so a diagnosis can be obtained in advance of the treatment decision. Incisional biopsies collect a  variable percentage of the tumor depending on the collection technique. Minimally invasive techniques are employed to collect a small sample and obtain a diagnosis. Endoscopic and Tru‐cut needle biopsies generally collect a very small percentage of the tumor. Laparoscopic wedges and punch biopsies collect more and subtotal wedges often provide a substantial amount of tissue. Minimally invasive procedures are also used when a discrete mass cannot be visualized. The goal of these incisional techniques is diagnosis alone with no concern about margins. The disadvantage is that  sometimes the sample does not contain diagnostic tissue. This  commonly occurs in tumors possessing substantial areas of necrosis, inflammation, reactive stroma, or collection artifacts. The small size of the sample compounds the effect each of these has on the  diagnosis. Samples in which no lesion is visible because  of artifact or inadequate diagnostic tissue are termed

Tumors in Domestic Animals, Fifth Edition. Edited by Donald J. Meuten. © 2017 John Wiley & Sons, Inc. Published 2017 by John Wiley & Sons, Inc.

27

28    Tumors in Domestic Animals

Figure  2.3  Cell Safe™ cassette containing endoscopic intestinal samples.

Samples this small are difficult to orient. Tissues in FFPE block may need to be rotated to obtain correct orientation.

Figure  2.1  Excisional biopsy. The entire mass is removed by the biopsy. Margin evaluation can be performed.

Figure 2.2  Incisional biopsy. Only a portion of the mass is removed by the

biopsy. Margin evaluation is not possible.

“Nondiagnostic biopsies.” When very small percentages of a tumor are collected, the risk of a nondiagnostic biopsy is highest. In all of these cases, diagnosis is the most important outcome of the biopsy. By definition, margin evaluation is not an issue in these samples because complete removal was not attempted. Specimen trimming and orientation in  the cassette are less important because of the limitations associated with tiny specimens. Indeed, most endo­ scopic samples are too small for further manipulation. But these samples must be ­protected from damage or loss during processing. Endoscopic and  needle specimens should be processed in small

screened containers such as Cell Safe™ cassettes that fit right into the standard processing cassettes and protect delicate tissues from damage and loss (Figure 2.3). Large tumors present a different set of problems. The entire tumor or organ containing the tumor (spleen, lung lobe, limb amputation) may be completely removed by the surgeon and sub­ mitted for evaluation and diagnosis. Although the pathologist has 100% of the tumor it is often far too large to evaluate all or even a substantial portion of it. The pathologist must decide what por­ tions to trim and examine. Because the surgeon desires a margin ­evaluation in addition to the diagnosis, the specimens must be trimmed and oriented in the tissue cassettes so as to ensure the most accurate assessment of the margin possible. In many cases, the  diagnosis has  already been made on a previous small inci­ sional or needle specimen so completeness of excision is now the important information desired from the biopsy procedure. This is a very different issue from trimming tumors from autopsies. What and how many portions of the tumor should be trimmed and how they are faced in the tissue‐processing cassette becomes critical to ensure that diagnostic tissue is available on the slides and that the correct margins are identified, included, and ori­ ented so as to permit the most complete evaluation of the surgical margin possible. This is a  critical step. It is subjective and will vary somewhat with the pathologist or technician. If treatment will go beyond the present excision then margin evaluation will help the oncologist decide what treatment plan is optimal for the pet and its owner. In such cases the trimming technician should consult with the pathologist for assistance in proper selection and orientation. The most definitive indication of the surgical margin is made by the surgeon on the specimen at the time of collection.2 Measurement of the margins taken by the surgeon before, during, and immedi­ ately after removal of the mass will be greater than the histologic margin reported by the pathologist. Shrinkage of specimens can be up to 40% and the majority of this occurs immediately after removal and before fixation.3,4 Although suture tags have been a traditional method, inking or dyeing the margin of interest is the superior method because the ink is applied by the surgeon who knows which  margins are to be evaluated. The ink is visible to the

Trimming Tumors for Diagnosis and Prognosis    29

pathologist at trimming and while reading the finished slide. This is a form of direct communication between the surgeon and patholo­ gist. If properly applied by the surgeon, the pathologist can be confident that the inked margin depicts the real margin(s) of ­ interest. The surgeon can color code for regions they are most concerned about. If the surgeon trimmed off fat and or connective tissue around the tumor then this needs to be reported. Applying ink to the regions after adipose or fascial layers are trimmed off may be unnecessary or misleading. Although tissue can be inked by the trimmer during preparation, they must translate the clinician’s notes or other form of designation, which may be imprecise and so create doubt in the pathologist’s mind about the accuracy of the designation. Inking can be done before or after formalin fixation. Surgical ink is ­commercially available in multiple colors. Black, green, and yellow colors are optimal because they provide high con­ trast to the blue and orange colors of H&E‐stained slides. The tis­ sues should be blotted dry and inked with a cotton swab (rather than immersed in the ink) and then allowed to dry for 5–10 min­ utes. Alternately it can be rinsed with a dilute glacial acid solution to set the ink. A specimen can be painted with one or more colors if needed. When properly inked, the surgical margin can be easily seen in the histological section by the pathologist and the relation­ ship of the margin to the tumor assessed and measured. Suture tags can establish a margin but are not observed in the finished slide. Whether the ­surgical margin is inked by the surgeon or the tissue trimmer, the process should be noted on the referral form, clearly indicating which color indicates which margin. Inking establishes a clear line of communication between the surgeon and the labora­ tory about precisely which is the real margin(s) of interest and guides the tissue trimmer about how to orient the specimen for placement in the processing cassette. Specimens that are not inked have a substantial probability of incorrect margin assessment.

A

Orientation of trimmed specimens

Once the true surgical margin has been established and marked, the trimmer orients the tissue in the processing cassette so that the margins can be properly evaluated by the pathologist. The margin principle states that to adequately assess the completeness of the resection you should “examine as much surface area on the margin as possible.” If you can examine 100% of the margin (and the client can afford this many sections), you should do so. However, this is impractical and common sense combined with a few principles will determine how many sections need to be trimmed. In general, two orientations that produce two to four sections are commonly used and their use is dictated by the ­conditions around each case. Perpendicular margins Also called radial sectioning, in this orientation the tissue is cut perpendicular to the surgeon’s plane of section (Figure 2.4A). This is the conventional orientation employed most commonly and familiar to pathologists. The advantage of radial sectioning is that the relationship between the tumor and the surgical margin can be  observed and the histologic distance measured or estimated (Figure  2.4B). The disadvantage of this orientation is that very little of the circumference of the margin is actually examined; just the 2–5 µm thickness of the tissue section in that radius of the tumor and therefore the potential for false‐negative rates is high. Because of this, some feel this is the least favored method for margin evaluation.1 The percentage of margin area can be increased by serially sectioning or “bologna slicing” (“bread loaf­ ing”) the specimen, but even this will only examine 1–5% of the margin (Figure  2.5), which could be prohibitively expensive in large specimens for many ­veterinary clients. A variation of this makes two perpendicular cuts through the specimen at right

B

Figure 2.4  (A) Perpendicular margin. The tissue is cut perpendicular to the surgeon’s plane of cut. (B) When placed on the glass slide, the histologic tumor‐ free margin (HTFM) and mass can both be visualized and the distance between edge of the tumor and non‐neoplastic tissue estimated or measured and reported as M1–M4.

30    Tumors in Domestic Animals

angles (“points of the compass” cut) so that the lateral margin in four directions can be assessed (Figure 2.6A). Although there is only a trivial increase in surface area examined, the pathologist can get some sense of the symmetry of the tumor and its relation­ ship to the lateral margins in four directions. In large specimens these cuts may not include the tumor and the margin in the

Figure  2.5  Perpendicular margin, pictogram. Serial sections or “bread/

bologna slicing” of the entire mass. This technique will assess approximately 1–5% of the circumferential margin depending on the size of the tumor and the number of sections taken at specified intervals, none of which is stan­ dardized for animal tumors.

A

trimmed section so the margin width cannot be observed on the glass slide (Figure 2.6B). Parallel margins (en face) Also called tangential, “orange peel” or “shave” margins, in this ­orientation the tissue is cut parallel to the surgeon’s plane of section. The principal advantage is that the amount of surface area evaluated can be hundreds of times greater than in a perpendicular margin. In some cases, such as digits, ear canal ablations, bowel resections, tail amputations, etc., nearly 100% of the surface area can be examined, fulfilling the criteria for optimal margin evaluation (Figure  2.7A). The principal disadvantage of this is that the relationship between the tumor and the margin is lost from the slide and the margin width cannot be measured (Figure 2.7B). This can be critical, espe­ cially when looking at round cell tumors and soft tissue sarcomas. A  second disadvantage is that in large specimens, such as limb amputations, the cost of trimming and evaluating the entire en face margin could be prohibitive and impractical. It is important for the tissue trimmer to note what type of margin orientation was faced in the cassette because it may not be apparent from just looking at the slide. En face margins should be inked by the trimmer. The presence of any tumor cells on an en face margin indicates an incomplete resection. Large, complicated specimens may present a combination of both types of trim margins. The most effective way to communicate this is to draw a trimming diagram on the submission form indicating the shape of the surgical specimen, the direction and type of trim margins, and in which cas­ sette each was placed. This communication between trimmer and pathologist is critical to ensuring the surgical margins are correctly interpreted. A modification is to cut on a horizontal plane rather than ­perpendicular plane. This will yield tumor and lateral margins but junction of the tumor and deep margin may be lost depending on  the number of sections sampled. This is the principle of the Mohs technique, which can evaluate 100% of the surgical margin.5,6

B

Figure 2.6  (A) Perpendicular margin. Two cuts are made at right angles to evaluate five components of the margin: four lateral (“points of the compass” cut)

and the ventral margin. (B) Perpendicular margin. Modified “points of the compass” cut in an oversized mass to evaluate the margin in five directions. In this modification the specimen is too large to include the margin and mass in the same section so the histologic tumor‐free margin (HTFM) cannot be observed directly and measured or estimated on the glass slide. These type of sections can be reported as M4, >5 mm. Standard histopathology cassettes are 3.0 cm × 2.5 cm.

Trimming Tumors for Diagnosis and Prognosis    31

A

B

Figure 2.7  (A) Parallel or en face margin, pictogram. The tissue is cut parallel to the surgeon’s plane of cut. This technique prioritizes examination of the

outermost tissues supplied by the surgeon. A second or third section must be obtained from the tumor to establish the diagnosis. The cuts (sections) in this pictogram are made in a vertical plane; en face sections in a horizontal plane are seldom done in animal tumors but will yield tumor and peripheral margin until the deep margin is reached. (B) Parallel or en face margin placed on the glass slide. More surface(s) on the margin are examined but the relationship between the mass and the margin is not visible on the slide.

Complete circumferential peripheral and deep margin assessment allows examination of the entire surgical margin and has a lower false‐negative rate than “bread loafing.” In human medicine this is employed on frozen sections of basal cell tumors while the patient and surgeon wait for results.

of the “margins” during routine histological examination may influence treatment plans or the prognosis offered to owners. There is some shrinkage caused by retraction of elastic tissues immediately after removal.3 Studies of oral and cervical carci­ nomas in humans have shown significant shrinkage (30–50%) can be caused by formaldehyde fixation and tissue processing.7–9 Shrinkage in length and width of 20–30% and increase in thick­ ness up to 75% in normal dog skin caused by processing has been Margin evaluation The object of trimming neoplasms is to obtain a diagnosis and to documented.10 The presence of muscle blunted this affect whereas maximize the accuracy of margin evaluation so that completeness fascia did not. There were different patterns seen in different ana­ of excision can be determined. The histologic diagnosis is the most tomic locations of the body. Other reports have indicated the important factor to influence treatment plans and provide a majority of tissue shrinkage occurs before fixation in formalin and prognosis. For many tumors the diagnosis predicts outcome is in the range of 20% for length, 10–15% for width and area.3 ­ (e.g.,  sebaceous adenoma, fibroma, circumanal gland adenoma, Some have proposed pinning the resected specimen to a solid histiocytoma, osteosarcoma, urothelial cell carcinoma, beta cell object to reduce shrinkage. These factors explain the differences in neoplasms, etc.). Margin evaluation, grading, and mitotic count the tumor margin described by the surgeon versus the margins (MC) are not needed for these tumors or objective data that corre­ reported by pathologists. The width of the margin between tumor lates margins, MC, or grades with outcome is not known. In a and non‐neoplastic tissue is greater in vivo than that observed in recent survey, margin evaluation was considered the most impor­ histologic sections. tant component of the biopsy report, presumably after the d ­ iagnosis.1 Studies to compare the in vivo margin to the histologic margin The true margin between tumor and non‐neoplastic tissues is in in animal tumors and tissues are reported.4,10,11 Retraction and the patient. The surgical margin is created by the surgeon and the shrinkage up to 40% occur immediately post removal4,11 and likely histologic margin is created by trimming, paraffin embedding, and will vary between tissues and possibly among species or even histological sectioning. The margin of neoplastic versus non‐­ tumors, adding to the difficulty in arriving at a consensus on what neoplastic tissues in vivo can only be extrapolated via assessment constitutes a “safe margin.” Regardless, the true margin is in the of a specimen that has been removed from the body, fixed in patient and with present techniques we do not know what that is. ­formalin, embedded in paraffin (FFPE), and sectioned for micros­ We can only provide an estimate. It is likely that other parameters copy. Tissues retract or shrink ­significantly following surgical are more predictive of recurrence, disease‐free interval, and ­excision.3,7–9 Additionally, only a small percentage of the histolog­ survival times than the margin width, no matter how carefully it ical margins are assessed. If “bread loaf ”, perpendicular vertical is established or defined. Most important are the inherent biologic cuts are used it is estimated that 1–5% of the margin is examined characteristics of  the tumor and the host. We try to extrapolate depending on the size of the lesion and the number of sections tumor behavior from traditional histological features such as removed.5,6 Despite the inherent inaccuracies, the determination diagnosis, invasion, anaplasia, and MC. Immunohistochemical

32    Tumors in Domestic Animals

determination of proliferation indices (Ki67, PCNA, AgNOR) and gene expressions or mutations may also help predict the behavior of certain tumors. Molecular characterization of tumors is used in human medicine to predict prognosis and select ­treatments. These tools should prove useful if we are asked by an oncologist for information that cannot be obtained from H&E sections. It is likely that these characteristics will vary from tumor to tumor. For instance the MC for some tumors14,15 may have prognostic significance (mast cell tumor) while for others it does not (histiocytoma). Evidence‐based medicine is needed for each type of tumor to identify what the important prognostic charac­ ters are. However, for many cases in veterinary medicine the diag­ nosis and  the information that can be provided from H&E sections or cytology are sufficient. For the diagnostic pathologist communication from the clinical veterinarian is essential; if an owner does not want to consider treatment beyond a primary surgery then applying costly and time‐consuming techniques are not justified. Many techniques can be  applied to archive material should this information be desired at a later time. Applying new techniques to samples of tumors is relatively fast compared with gathering accurate follow‐up data over years. Too often our follow‐up data are based on imprecise clinical assessments of recurrence (­visualization, palpation) and nonstandardized assessment of metastases via imaging, survival and cause‐of‐death data based on phone interviews with owners or primary care veterinarians. We can retrieve archived material if a new technique is devel­ oped but we need follow‐up data (outcome assessment) if we want to determine whether the technique has clinical applica­ bility. This requires partnerships between clinicians, molecular biologists, and pathologists. Methods that help determine the biology of the tumor and the host may prove better ­predictors of prognosis12,13 and theranostics than changing how we enumerate margins and MCs.14

Skin tumors

In veterinary medicine, margin assessment has been predominantly applied to cutaneous neoplasms. The premise of measuring margin widths is simple: by determining if tumor cells are present in histo­ logic sections at the margin or how far they are from the margin we should be able to predict local recurrence. The true distance bet­ ween the tumor and normal tissues is in the patient and that is not known. The distance visualized in histologic sections between the tumor and the inked margin is the histologic tumor‐free margin (HTFM). If desired it can be reported for cranial, caudal, and both lateral and deep margins. In human oncology there is emphasis to determine the HTFM width that will prevent local recurrence. The closest histologic margin to a tumor is referred to as the histologic safety margin (HSM).5,16 The HSM is considered the microscopic distance that will prevent local recurrence. HSM is not known for  animal tumors, and it is more appropriate to report HTFM because HSM implies we know the distance that will prevent local recurrence. Determination of the HSM in certain human tumors is accomplished by the Mohs technique which can assess 100% of the margin.5,6 Some refer to the acronym, CCPDMA – complete circumferential peripheral and deep margin assessment. CCPDMA allows complete examination of the entire surgical margin and therefore has a lower false‐negative rate than “bread loafing” in a perpendicular plane, which only provides a few slides from selected areas. The Mohs

surgical technique is essentially serial frozen sections taken in a horizontal plane while the patient and surgeon wait for results. Sometimes the surgeon also examines the frozen sections. The advantages are immediate knowledge of margin, 100% assessment of surgical margins and maximum preservation of normal tissues. The latter is important for tumors like basal cell carcinomas located on  the head, eyelids, and other sites where the preservation of tissue  is  highly desirable.5 Some disadvantages are the high costs (>US$10,000), frozen sections, large tumors, deeply infiltrative ­neoplasms, and expertise needed. In human medicine, Mohs micrographic surgery is performed by board‐certified individuals. It is used for dermatologic specimens, when no metastases or micrometastases are expected and preserva­ tion of normal tissues is needed. However, its use and advantages have also been reported for tumors in a limited number of animals, most notably when 100% assessment of the margin is desired.6 That degree of certainty has a cost that may not be supported by ­veterinary medicine. The HSM differs between tumors or within subtypes of the same tumor, presumably because of differences in how the tumor grows, whether infiltrative or discrete, aggressive, biologic behavior, etc. There are numerous factors that contribute to local recurrence or metastases that go beyond the HSM. Even the type of anesthesia used for the surgical removal of tumors has been associated with increased risk of recurrence and metastases.17 The biology of the tumor and the host are likely more important deter­ minants of recurrence than how close tumor cells (microscopic assessment) are to the inked margins. Until recently there has been little research in veterinary medicine about this and no standardiza­ tion of how to trim tumors, measure margins or report the results.2,4,6.10,16 The distance of the closest approach of the tumor cells to the his­ tologic edge (ink) should be measured or estimated in the field of vision. This can be assessed along the cranial, caudal, both lateral and deep margins and reported as the HTFM. HTFM should be reported and not the tumor‐free margin. The HTFM varies among the histologic sections examined, is influenced by the pattern of tumor growth (discrete, infiltrative), the type of excision performed, the trimming protocol and total number of sections evaluated. The total number of sections examined is not standardized. HSM has not been determined for animal tumors, and for canine mast cell tumor (MCT) it could not be determined.16 In some human breast cancer studies the standard is “the tumor should not touch the margin ink.”18 In veterinary oncology some have proposed defined distances for canine MCT and soft tissue sarcoma (STS) from 2 mm to 5 cm and/or one fascial plane deep.19–22 Given the variable nature of the periphery, the asymmetry of some tumors, the minimal amount of the total margin actually examined, and shrinkage induced by removal and fixation we recommend pathologists report the HTFM as a range: •  M1 = margin infiltrated; focal or diffuse •  M2 = margin is close, 5 mm. To measure the HTFM more precisely (e.g., “The margin width is 2.6 mm”) implies a greater precision than is real from an FFPE specimen and that the entire margin was examined. The shortest microscopic distance to a margin should be reported, along with any specific regions requested by the surgeon. Gross measurement of margins should be taken by the clinician during surgery, if this is desired, and not performed on a fixed specimen. Immediately at or after removal there is shrinkage of the

Trimming Tumors for Diagnosis and Prognosis    33

specimen. For tumors that are M1 or M2 the report should also approximate the histological amount of tumor present at or near the margin: focal: few foci of tumor cells are seen or diffuse: large number of tumor cells present. Many of us do not use a micrometer, grid, or reticule to determine measurement(s) and estimate a ­distance based on prior determination of how many millimeters are  present in the field of view of our objectives. This can be ­determined with a stage micrometer, a ruler or by this simple calculation: FN (mm)/objective magnification diameter of the field of view in millimeters14

FN 22 mm/40× objective = 0.55 mm diameter FN 22 mm/20× objective = 1.1 mm FN 22 mm/10× objective = 2.2 mm FN 22 mm/4× objective = 5.5 mm

M1 20 mm/40 M2 20 mm/20 M3 20 mm/10 M4 20 mm/4

=0.5 =1 =2 =5

The FN is the field number in millimeters engraved on the side of the ocular and is also referred to as FOV or field of view of the ocular eyepiece. The diameter in the FOV of the pathologist varies with each objective and is calculated by dividing the FN by the objective magnification.14 In newer microscopes the two most common FNs are 20 mm and 22 mm. Newer microscopes have infinity objectives so if a tube lens was also part of the microscope system it does not affect the final FOV and does not have to be part of the calculation. Knowing the diameter in the FOV at the different objective magnifi­ cations makes reporting M1 through M4 relatively easy. We need to standardize how we report histologic margins and MCs (mitotic figures in 2.37 mm2).4,14 As mentioned previously, there is shrinkage of the specimen and the in vivo margin cannot be determined from histologic sections, only estimated.3,4,10,11 Taking these variables into count we believe using histologic descriptors such as M1–M4 provides us with ranges that are appropriate for our  method(s) of measurement (micrometer versus estimate) and although they may seem less precise than a specific number they are likely more accurate, appropriate, and less misleading than reporting an “exact” distance for an imperfect specimen. Additionally we know for canine MCT and STS that grade may be or is more predictive of local recurrence than is the “margin width.”16,25 For financial and other pragmatic reasons we usually only take 1–4 samples to assess the margin of most skin or subcutaneous tumors. Small tumors removed by en bloc resection that fit com­ pletely into the processing cassette (without touching the sides of the cassette) are the easiest specimens to evaluate. These tumors can be trimmed by one or two perpendicular cuts and processed in one or two cassettes (Figure 2.4A). The deep and lateral histologic margins of the specimen can be easily observed in one or two directions and a histologic distance from the tumor to the margin (HTFM) can be  obtained. Even in small tumors only a small percentage of the entire margin is actually seen in these one or two histologic sections. En  bloc resection specimens that exceed the dimensions of the processing cassette need to be subsampled. Depending on clinicians’ preference and the specimen (tumor type, tissues) removed this total tissue mass could occupy an additional 1–5 cassettes, or glass slides. If this is done in four quadrants this will add considerable cost. There is no method to predict where tumor foci may be located, therefore these additional sections are random and add little reassurance that the regions chosen were likely to contain foci of the tumor. They certainly do not represent the entire circumference.

There appears to be more emphasis placed on wide lateral than deep margins. Depending on where the tumor is located the depth of excision will be influenced by anatomical structures that limit deeper excision. Determining whether shrinkage is different at the deep margin versus the lateral margins will be problematic. Establishing the diagnosis is the most important task so a section through the mass should be obtained first, followed by sections that help identify the HTFMs. Perpendicular margins in four directions can usually be obtained and the entire specimen evaluated in sev­ eral cassettes (Figure 2.6). Alternatively, a series of parallel (en face, tangential) margins can be taken around the en bloc excision that increases the surface area examined, although this is still a small percentage of the true circumference of the tumor (Figure  2.7A). The combination of both types of margins is often not practical in en bloc skin resections. The best option may be serial perpendicular sections through the entire specimen (Figure 2.5), but this may be expensive, depending on the size of the specimen. If the clinician knows the client’s tolerance for expenses they could be consulted, but this adds delays and short turnaround times of reports is emphasized. The choice of how the tumor is trimmed is usually made by the pathologist, trainee, or technician. The choice of margin orientation is difficult, with each having advantages and disadvantages. A perpendicular margin is best because it permits visualization of the main body of the tumor and the relationship between isolated tumor cells at the periphery of the tumor with adjacent non‐neoplastic connective tissues. But in an irregular asymmetrical tumor mass, the small percentage of margin actually evaluated may miss cells at the surgical margin but out of the plane of section (false negatives). En face margins could provide more surface area but the definitive identification of individual tumor cells in such sections leaves open the question if these cells are neoplastic or normal components of the tissue. Although epithelial neoplasms generally grow as cohesive cells grouped as a solid mass of connected cords, tubules, or acini (Figure 2.8), round cell neoplasms grow as a spheroid shape with variably dense aggregates of unconnected cells that may produce an irregular, asymmetrical mass with an indistinct edge (Figure 2.9). This pattern of growth is especially true for MCT. Some are well defined, others are infiltrative or some types of tumors tend to be solitary while others are multicentric. The variability in the spatial distribution of cells in MCT makes accurate identification of the true limit of the tumor cell population challenging. The tumor may appear completely excised in one section but incomplete in another section deeper in the block. It can be confounded by the difficulty of separating individual non‐neoplastic round cells (lymphocytes, macrophages, and mast cells) from the neoplastic cells. Likewise, STSs are dense, interconnected mesenchymal cell populations that may form an asymmetrical mass with concentric laminations (pseudo‐capsule) or irregular, elongated finger‐like projections on the edge of the mass. These projections may extend outward from the main body of the tumor like the root ball of a young tree (Figure 2.10). Because of possible irregularities at the periphery of STSs, the HTFM in one slice may be half or twice the distance in another section 1 or 2 mm into the block. Due to the variable nature and asymmetrical shape of the tumor, the width of the histologic margin(s) in some may vary considerably. Clinicians should interpret with caution HTFMs or margin distances obtained by a single measurement in a single perpendic­ ular section of MCT and STS. It is likely that optimal HTFM and HSM widths will need to be established for each tumor type based upon a standardized method that can be applied across

34    Tumors in Domestic Animals

A

B

Figure 2.8  (A) Epithelial tumor, pictogram. Benign types grow as a cohesive sphere of connected cells with even edges. Carcinomas may be infiltrative. The natural surface of both types can be compared to the “tip of an iceberg.” (B) The margin is easily observed in perpendicular section but the HTFM may vary in different parts of an asymmetrical tumor or infiltrative carcinomas. The HTFM in the lower panel is narrower than the upper panel.

A

B

Figure 2.9  (A) Round cell tumor, pictogram. The tumor grows as a sphere of unconnected cells with an irregular edge. (B) The tumor appears completely

excised in the upper panel but tumor cells infiltrate the margin in another section deeper in the block.

laboratories. However, for some tumors margins are not needed (benign neoplasms such as histiocytomas, sebaceous adenoma, circumanal adenoma) or margins are less important compared to the diagnosis of a malignant neoplasm (osteosarcoma, melanoma, urothelial carcinoma, transmissible venereal tumor, prostate car­ cinoma, anal sac carcinoma, islet cell neoplasm). Knowing the biologic behavior of a tumor type is often predictive of recurrence and metastasis.

The different patterns of growth among carcinomas, STSs, and round cell tumors is a basic biologic characteristic of each tumor type and likely contributes to the variability reported in margin evaluation of asymmetrical tumors. This may partially explain the different accuracies in predicting local recurrence among the three different skin tumor types: 94% for carcinomas, 87% for STSs, and only 76% for MCTs.23 The percentages cited are the accuracy of his­ topathology to predict local recurrence, they are not the recurrence

Trimming Tumors for Diagnosis and Prognosis    35

A

B

Figure 2.10  (A) Soft tissue sarcoma, pictogram. The sarcoma grows as an asymmetrical mass with highly irregular edges consisting of concentric lamination or finger‐like extensions similar to a root ball of a tree. (B) The tumor appears completely excised in the upper panel but deeper in the block sections reveal tumor cells infiltrating the margin.

rates for the three types of tumors. Many of these studies should be repeated, the methods standardized, the numbers of cases increased and additional parameters evaluated, including molecular tech­ niques when available. These studies should try to include the methods used in prior studies so the new parameters can be com­ pared to old parameters. This knowledge is needed for pathologists, researchers, and for clinicians providing advice to owners.

Canine soft tissue sarcomas and mast cell tumors

What have we learned from the numerous reports on margins, local recurrence, grading, and biologic behavior of these two common tumors? The theory that residual tumors cells at a margin will predict local recurrence is reasonable but it is only partially predictive when applied to real cases.16,18–25 The majority of STSs (75%) and MCTs do not recur, even when tumor cells are reported to be present at or close to the margin (see Appendix).16,25–27 If tumor cells are seen at the margin then grading STS may add addi­ tional information about local recurrence: approximately 38% of 45 grade 2 or 3 STSs recurred but 60% did not.25 Of the 45 dogs followed with tumor at or close to the margin, 3 of 4 grade 3, 14 of 41 grade 2, and 3 of 41 grade 1 recurred. However, the determina­ tion that surgical margins are free of tumor cells has a high proba­ bility that an STS will not recur: >95% (54/55) of STSs that were not at the margin did not recur.25,27 Interestingly, recurrence did not affect survival for dogs with STS. For cutaneous MCTs the results are similar: local recurrence increases as the grade of the MCT increases16,26 and if the mar­ gins are free of tumor cells then the recurrence of low‐grade or grade 1 or 2 MCT is 95%),1,2 but in dogs only 7 of 61 cases (11%),3 26% of 73 cases,4 or 17 of 435 were of B‐cell type and 1 of 18 cats was B‐cell.6 Approximately 80% (75–90%) of canine CLL cases are T‐cell,3,4 95% (17/18) of cats are T‐cell, and 94% of these were T‐helper CD3+/CD4+/CD8− phenotype.6 Aberrant phenotypes have been reported for dogs.4,5 Horses have been inadequately studied. Cows with CLL are usually of B‐cell type. If tumors are not phenotyped then the diagnosis is CLL. B‐cell CLL is a neoplasm of the bone marrow and is character­ ized by marked lymphocytosis (leukemia) of small mature B lym­ phocytes. Animals often have few or no clinical signs and the disease has a slow but persistent progression. Spleen and liver are usually enlarged due to tumor cells but lymph nodes may be normal size. B‐cell small lymphocytic lymphoma (SLL) is a tumor of small mature B lymphocytes involving lymph nodes, spleen, and solid organs but leukemia is not present or neoplastic cells in blood are few. SLL is an uncommon lymphoma in dogs and is much less common than CLL. It has a slow rate of progression. B‐cell prolymphocytic leukemia (PLL) is related but is an earlier stage of maturation than mature lymphocytes and is more aggres­ sive. In terms of organ distribution it is a mirror image of CLL. Morphologic identification is accomplished better with cytology in which the nuclei appear more “immature,” are larger, the chromatin is in aggregates joined by fine chromatin bands, and the nucleolus is prominent (Figure 7.3). Since the prognosis and therapy differ from those in CLL the distinction is clinically important.

Epidemiology and occurrence

Lymphoma and leukemia of small mature lymphocytes are indo­ lent, B‐cell and T‐cell. They are slowly progressive neoplasms that may involve bone marrow or peripheral tissues or both. They are considered neoplasms of accumulation rather than proliferation. Aggressive variants have been identified, usually associated with atypical phenotypes.4,5 B‐cell CLL is uncommon in domestic ani­ mals, although reported in dogs, cats, and other species.4,7,8–12 In the American College of Veterinary Pathologists (ACVP) study of 1000 cases of canine lymphoma there were 8 cases of B‐cell SLL. In humans there is a strong male predominance but there is no sex or no breed predominance in animals. In animals, as in humans, the tumor is found most often on routine examination of blood. Cells of PLL type are seen in humans,13 mature cows, and dogs, but are relatively rare.9

Clinical presentation

Most cases are asymptomatic and the disease is discovered from a CBC taken at an annual exam that reveals lymphocytosis. Weight loss is reported in approximately 50% of the cases. Cats and dogs with CLL and SLL are usually over 5 years of age, the mean age for 134 dogs with CLL (B and T) was 10 years, the range was 1.5– 19 years3,4 and 17 cats with T‐cell CLL the mean age was 12.5 years.6 Only one cat in this series had B‐cell CLL. Animals in an acceler­ ated phase of the disease occasionally present with epistaxis, likely due to thrombocytopenia; some will have diarrhea and vomiting. In

216    Tumors in Domestic Animals

B

A

C

Figure 7.2  Chronic lymphocytic leukemia (CLL), feline. (A) Nine‐year‐old cat with a lymphocyte count of approximately 11,000/μL. The lymphocytes are

small to intermediate size with irregular shaped or indented nuclei. A few cytoplasmic granules are faintly apparent (see Figures 7.39 and 7.40). Flow cytometric analysis indicates that lymphocytes (light scatter plot, arrow) express CD4 and CD3 (T cells) and do not express CD8 and CD21. In healthy cats there would be a mixture of CD4‐, CD8‐, and CD21‐expressing lymphocytes in blood. Neutrophils (light scatter and control plots, arrowhead) do not express CD3, CD4, CD8, or CD21. Most CLL in cats is of T‐cell type, as is this case, and cells more commonly express CD4 than CD8. This cat tested neg­ ative for feline leukemia virus. Platelets are adequate in this cat, anemia is expected, but neutropenia is not usually seen with CLL. (B) B‐CLL, dog. Marrow core, the marrow is completely infiltrated, few fat cells remain, bone is of normal density and volume. Typically dogs with B‐CLL will have a marked lym­ phocytosis and paraproteinemia. (C) Marrow core, higher magnification. The marrow is heavily infiltrated with mature lymphoid cells of small to intermediate size, nuclei are about 1.5 RBC in diameter and nucleoli are not obvious. There is very little stromal proliferation. IHC is needed to identify phenotype, most CLL are T‐cell type. A young megakaryocyte is in the upper right.

dairy cattle the most common clinical problem is a drop in milk production and feed consumption.

Pathology

Blood, bone marrow, lymph nodes, and spleen The peripheral blood count and cytologic review of a film is diag­ nostic, as the nucleated cell count will be 100,000–400,000/μL and >90% will be lymphocytes (Figure 7.2). Reported mean lymphocyte counts are approximately 100,000/μL but the ranges are wide, from near ­reference interval to >1,000,000/μL.3,4 One study reported a mean of 137,000/μL for B‐CLL and 60,000/μL for T‐CLL LGL type.3 In cats the total median lymphocyte count reported was 34,000/μL (>90% of cases were T‐cell),6 and counts may be as high as 400,000/μL. The higher the lymphocytosis the more likely the diag­ nosis is CLL; however, there are non‐neoplastic causes of lymphocy­ tosis in all species that are more common than CLL. Anemia to some degree is present in approximately 60–75% of dogs and platelets are decreased in 15–25%.3,4 The anemia is

non‐regenerative and can be mild, moderate, or severe; the latter is least common. Neutropenia is not present in dogs with CLL and this is a distinguishing feature from ALL and AML in which neutro­ penia is expected. B‐CLL originates in the bone marrow and numerous bones will be infiltrated with tumor cells in dogs, less so in cats (Figure 7.2B,C). Neoplastic foci are identified by an absence of fat cells and in these regions there are small uniform lympho­ cytes. Some cases will have bone marrow nearly filled by neoplastic cells and others may be as low as 20% lymphocytes. Core samples are more reliable to identify CLL than aspirates.6 T‐CLL is likely of splenic origin. A consistent finding in CLL is hepatosplenomegaly. Animals with SLL are not leukemic or the neoplastic cells in circulation are at a low level because the bone marrow is not involved or only focally so. In SLL all organs and nodes may eventually be involved as there is slow progression and gradual dissemination of the neoplastic cells. The lymph nodes in cats and dogs with CLL can be atrophic or enlarged. With SLL the nodes are neoplastic and the peripheral sinus

Tumors of the Hemolymphatic System    217

like humans, can have larger cells, especially if they enter blast crisis with nuclei 2 RBC in diameter and relatively abundant lightly baso­ philic cytoplasm. Dogs also have a CLL in which the cells are large granular T lymphocytes. The cytologic diagnosis of SLL can be problematic as the cells are mature. Absence of plasma cells, uni­ form small lymphocytes from multiple enlarged lymph nodes, and hepatosplenomegaly all support the diagnosis. Clonality and immu­ nophenotyping may be needed. The cells of PLL are larger than those of CLL or SLL and the chromatin pattern is distinctive (Figure 7.3). PLL cells have round and moderately irregular nuclei that are approximately 2.0 RBC in diameter. The chromatin distribution is characteristic, with large, densely stained chromocenters that are about one‐third the diam­ eter of a red cell apart but joined by narrow chromatin bands. A distinctive feature is the parachromatin clear areas that tend to sur­ round large chromocenters. There is mild anisokaryosis; the larger cells have small but relatively prominent central nucleoli that may be multiple. The cytoplasm is relatively abundant and lightly stained (Figure 7.3). These cells have a low but consistent mitotic count in tissues with 0–2 mitoses/400× field.

Figure 7.3  Prolymphocytic leukemia (PLL) in a cow. The cells of PLL have a

characteristic chromatin pattern that is composed of multiple aggregates of chromatin that are joined by fine chromatin bands. Nucleoli are small and inconspicuous, and there is abundant lightly stained cytoplasm. B‐ versus T‐cell phenotype cannot be determined from cytology.

and medullary cord regions are compressed by the diffuse cortical expansion of neoplastic cells. The tumor forms solid sheets that appear densely stained because the nuclei are small and close together. Neoplastic cells have little cytoplasm. Lymphatic vessels within and close to nodes may be dilated and packed with neoplastic cells. The spleen in B‐cell CLL has atrophy of the thymic‐dependent periarteriolar lymphoid sheaths. There is scattered infiltration of the sinus areas and in later stages of the disease these may become confluent. There can be fading germinal centers with hyalinized centers. Hematopoiesis is restricted to adjacent non‐affected red pulp with megakaryocytes appearing to be the most persistent of the hematopoietic cells. With progression there is colonization of subendothelial areas of the large muscular veins and colonization of hepatic sinusoids. With SLL there is an absence of germinal centers and the infiltrations of the spleen are larger and more solid and coalescing. Animals with B‐cell CLL may have a macroglobulinemia like the  gammopathy of the Waldenstrom type in humans, but it is uncommon. Routine chemistry may reflect this change by an unusually low albumin‐to‐globulin ratio. Serum electrophoresis will verify the presence of a monoclonal gammopathy, usually IgM. When present, macroglobulinemia suggests that the tumor is of B‑cell origin and is more likely of a prolymphocytic type. Although immunophenotyping is needed to definitively separate B‐ and T‑cell CLL, the presence of a monoclonal gammopathy suggests B‑cell and granular lymphocytes favor T‐cell. Neoplastic cells B‐CLL and SLL cells look like mature lymphocytes and are identi­ fied as B‐cell with immunophenotyping. Small lymphocytes have nuclei 1–1.5 RBC in diameter with a narrow rim of cytoplasm and dense nuclear chromatin. It is the marked lymphocytosis that sug­ gests neoplasia (Figure 7.2). Dogs usually have cells of this type or,

Other organs Both CLL and SLL have areas of infiltration in and around hepatic portal tracts. Lymphocytes in the hepatic sinusoids are usually ­present with CLL. Dilated veins containing neoplastic cells may be present in lung, kidney, central nervous system and eyes in CLL.

Cytochemistry and immunohistochemistry

B‐cell CLL and SLL are differentiated from the more common T‑cell phenotype by IHC or flow cytometry (Figure 7.2). The B‐cell CLL and SLL cells are positive in tissues with CD79 alpha, CD20, or CD21 and are negative with CD3. B‐cell CLL cells are negative with CD5 (unlike humans) and should express CD1c (95%) and CD1a (78%).4 Cells from animals and humans with CLL are nega­ tive for CD34, which is expected to be positive in ALL and AML. CD34 is a surface glycoprotein expressed on hematopoietic stem cells (and other cells) and is useful to distinguish acute and chronic leukemia. It should be negative in CLL and most lymphomas and positive in ALL and AML. All of these neoplastic cells can have stages or types in which the morphology of the cells are identical or so similar that phenotyping, especially with flow cytometry and multiple antibodies, may be necessary to correctly identify the neoplasm. Neutropenia is not ­present in CLL and is a feature of acute leukemias. Acute leukemias are symptomatic and have a rapid course with a fatal outcome, while CLL is often asymptomatic and has a clinical course of months to years. Differential diagnoses and how to distinguish them is dis­ cussed in the section on T‐cell chronic lymphocytic leukemia.

Tumor cell transformation and prognosis

These are indolent lymphomas. Animals may live for years without treatments. Since these tumors are slowly progressive and clinically occult they are usually not diagnosed until the tumor has been pre­ sent for an extended period, perhaps 1–2 years. Despite the slow progression a diagnosis of leukemia is considered to be more serious than a lymphoma. Humans with SLL have a 2‐year survival of 60–70% and a 5‐year survival of 50–70%. A monoclonal antibody has been used for treatment in humans.14,15 Dogs are treated with a variety of chemotherapies that may include prednisolone. Some are treated with only prednisolone and others are not treated.

218    Tumors in Domestic Animals

In animals, CLL and SLL can undergo an accelerated phase sim­ ilar to these diseases in humans. When this occurs in the tumor’s life is not known, but the cells gain in mitotic rate and the cells enlarge to prolymphocytic type (Figure 7.3). The next stage is sim­ ilar to Richter’s syndrome, in which the small cells evolve into large cell lymphoma.11,13,16 Further progression in humans is classified as blastic transformation, plasmacytoid transformation, and, finally, immunoblastic transformation.9–11,15 The diagnostic features taken to predict the progression of human CLL are the clinical stage, lym­ phocyte count and doubling time, pattern of marrow involvement, cytogenetic abnormalities, and atypical immunophenotype.1,16 These methods are not available on a diagnostic basis in veterinary medicine. The distinction of blast transformation in CLL or SLL from a large cell lymphoma is not clear unless there was documentation of a prior small cell CLL or SLL. Most canine large cell lymphomas are B‐cell, multicentric, and some may have concurrent leukemia so the distinc­ tion may be difficult when either disease is seen at one point in time. Dogs with CD21‐positive lymphocytosis have different survival times that can be separated by size of the neoplastic cells as deter­ mined from flow cytometry.17 Dogs with large cell B‐lymphocytosis had MSTs of 130 days versus an MST that was not reached for small cell lymphocytosis (>1–2 years) as determined by flow c­ ytometry.17 See T‐cell chronic lymphocytic leukemia section in this chapter and Appendix of this book for more information on survival and prognosis.

References

1. Admirand, J.H., Knoblock, R.J., Coombes, K.R., et al. (2010) Immunohistochemical detection of ZAP70 in chronic lymphocytic leukemia predicts immunoglobulin heavy chain gene mutation status and time to progression. Mod Pathol 23:1518–1523. 2. Chiorazzi, N., Rai, K.R., and Ferrarini, M. (2005) Chronic lymphocytic leukemia. N Engl J Med 352:804–815. 3. Tasca, S., Carlil, E., Caldin, M., et al. (2009) Hematologic abnormalities and flow cytometric immunophenotyping results in dogs with hematopoietic neoplasia: 210 cases (2002–2006). Vet Clin Pathol 38:2–12. 4. Vernau, W. and Moore, P.F. (1999) An immunophenotypic study of canine leuke­ mias and preliminary assessment of clonality by polymerase chain reaction. Vet Immunol Immunopathol 69:145–164. 5. Comazzi, S., Gelain, M.E., Martini, V., et  al. (2011) Immunophenotype predicts survival time in dogs with chronic lymphocytic leukemia. J Vet Intern Med 25:100–106. 6. Campbell, M.W., Hess, P.R., and Williams, L.E. (2012) Chronic lymphocytic leu­ kaemia in the cat: 18 cases (2000–2010). Vet Comp Oncol 11:254–264. 7. Gerou‐Ferriani, M., McBrearty, A.R., Burchmore, R.J., et  al. (2011) Agarose gel serum protein electrophoresis in cats with and without lymphoma and preliminary results of tandem mass fingerprinting analysis. Vet Clin Pathol 40:159–173. 8. Harvey, J.W., Terrell, T.G., Hyde, D.M., and Jackson, R.I. (1996) Well‐differentiated lymphocytic leukemia in a dog: Long‐term survival without therapy. Vet Pathol 18:37–47. 9. Valli, V.E. (2007) Mature (peripheral) B‐cell neoplasms. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 142–161. 10. Valentine, B.A. and McDonough, S.P. (2003) B‐cell leukemia in a sheep. Vet Pathol 40:117–119. 11. Su, Y.C., Wu, W.M., Wu, M.F., and Chiang, B.L. (2001) A model of chronic lymphocytic leukemia with Ritcher’s transformation in severe combined immuno­ deficiency mice. Exp Hematol 29:1218–1225. 12. Vezzali, E., Parodi, A.L., Marcato, P.S., and Bettini, G. (2009) Histopathologic classification of 171 cases of canine and feline non‐Hodgkin lymphoma according to the WHO. Vet Comp Oncol 8:38–49. 13. Reininger, L., Bodor, C., Bognar, A., et  al. (2006) Richter’s and prolymphocytic transformation of chronic lymphocytic leukemia are associated with high mRNA expression of activation‐induced cytidine deaminase and aberrant somatic hyper­ mutation. Leukemia 20:1089–1095. 14. Mavromatis, B. and Cheson, B.D. (2003) Monoclonal antibody therapy of chronic lymphocytic leukemia. J Clin Oncol 21:1874–1881.

15. Zhou, Y., Tang, G., Medeiros, L.J., et al. (2012) Therapy‐related myeloid neoplasms following fludarabine, cyclophosphamide, and rituximab (FCR) treatment in patients with chronic lymphocytic leukemia/small lymphocytic lymphoma. Mod Pathol 25:237–245. 16. Leuenberger, M., Frigerio, S., Wild, P.J., et  al. (2010) AID protein expression in chronic lymphocytic leukemia/small lymphocytic lymphoma is associated with poor prognosis and complex genetic alterations. Mod Pathol 23:177–186. 17. Williams, M.J., Avery, A.C., Lana, S.E., et al. (2008) Canine lymphoproliferative disease characterized by lymphocytosis: Immunophenotypic markers of prog­ nosis. J Vet Intern Med 22:596–601.

Plasmacytoma, plasmablastic lymphoma, lymphoplasmacytic lymphoma, and myeloma‐ related disorders Plasmacytoma, plasmablastic lymphoma, and lymphoplasmacytic lymphoma (LPL) are neoplasms of mature, differentiated B lym­ phocytes, and the neoplastic cells look like or are plasma cells. Plasmacytomas occur outside of lymph nodes, and plasmablastic lymphoma and LPL originate in nodes and may be found in organs. Very rarely does a plasmacytoma arise in a node and equally rare is a primary osseous plasmacytoma. There are other B‐cell neoplasms reported in humans, including lymphoplasmacytoid lymphoma and immunocytoma, however, comparable tumors of similar mor­ phology, immunophenotype, and biologic behavior have not been characterized in animals.1,2 In the ACVP study of 1000 cases of canine lymphoma there were 15 cases of plasmacytoma, 7 cases of plasmablastic lymphoma and one case classified as a B‐cell plasmacytoid lymphoma.3 It is not clear if the plasmacytomas were extramedullary.3 In the 600 dog study there were 6 plasmacytoid B‐cell, 9 lymphoplasmacytic B‐cell, and 19 plasmacytoid T‐cell lymphomas.4 There also is a disease category called myeloma‐related diseases (MRD) that is used for similar tumors and disease entities in cats and humans.5,6 MRD is a group of tumors in cats that are derived from plasma cells or immunoglobulin‐secreting B lymphocytes.5 How tumors are defined influences prevalence and the numbers of tumors in subjective categories. Plasmacytoma is the most common tumor in this group in dogs by a wide margin. Plasmacytomas are a proliferation of differenti­ ated B lymphocytes that originate primarily in soft tissues, oral and subcutaneous locations and rarely in nodes, organs, or bone. Plasmacytomas are common tumors in dogs that were misdiag­ nosed in the older literature and are reported under a wide variety of names, including reticulum cell sarcoma.7 Plasmablastic lymphoma and LPL are uncommon tumors which develop from B lymphocytes in lymph nodes and/or organs. MRD is an umbrella term for a group of about six diseases that arise from B lymphocytes. It has been used to define tumors in cats. Within this group are plasmacytomas, cutaneous and extracutaneous and B‐lymphocyte tumors that secrete immunoglobulins. Some of these latter tumors may be classified as plasmablastic lymphoma or LPL by others, especially if organs were involved and/or leukemia was present. As a generalization the plasma cell–like neoplasms in cats that are not well differentiated are aggressive.

Clinical presentation and occurrence

Plasmacytomas are common in dogs, are seen in cats, and are unusual in horses and cattle.1,2,5,6,8–11 They are described in more detail in Chapter  5. Plasmacytomas are usually solitary, benign tumors in the skin, subcutis, or gastrointestinal tract and most do not recur following excision. Solitary tumors have been identified in the liver of dogs and cats and rarely have been associated with paraneoplastic syndromes of hypoglycemia, hypercalcemia, or

Tumors of the Hemolymphatic System    219

paraproteinemia. Paraproteinemia is a feature of myeloma, MRB, and LPL. Plasmacytomas are also found in spleen, kidney, and intestines, but in these locations consider LPL, plasmablastic lym­ phoma, and MRD. Plasmacytomas are well described in cats. They are divided into cutaneous and extracutaneous and compared to MRD.5,6,8 In cats the data suggest they originate in soft tissues but later may be present in bones, which is a different pathogenesis from that proposed for human MRD.5 Cats with MRD are FeLV and FIV negative. Some of the classifications and categories in MRD are similar to other diagnoses that use different nomenclature. Regardless of names, several important conclusions from these investigations in cats are that cutaneous plasmacytomas respond well to excision, well‐differentiated tumors are associated with survival times of approximately 250 versus 14 days for poorly differ­ entiated tumors, there is good concordance between cytology and histopathology, many tumors will not label with CD79a, mitotic ­figures are not frequent, and giant cells and giant nuclei may be seen in more aggressive tumors, as they are in dogs and humans.5 Most plasmacytomas are benign but infrequently cutaneous plas­ macytomas behave aggressively and shorten the life of the dog or cat. When plasmacytomas are present in the aerodigestive system of dogs they are almost always solitary and benign. However, those that occur in the stomach or rectum may be clonal and aggressive and may cause vomiting and weight loss. When multiple tumors with plasma cell features are present in the intestines, and especially if they are also found in liver, spleen, or lymph nodes, the diagnosis in dogs is more likely plasmablastic lymphoma or LPL. Plasmacytomas in the rectal area usually present with some history of difficulty in defecation and are accompanied by recurrent bloody stool.10 The potential for plasmacytomas in the gastric and rectal areas to be aggressive necessitates wide surgical excision of tumors in these areas.9,10,12 Plasmacytomas in dogs are rarely associated with mono­ clonal gammopathy. If a monoclonal gammopathy and hypercal­ cemia are present the dog is more likely to have multiple myeloma. Cats with MRD are reported to have paraproteinemia and a few have had hypercalcemia.5,6 Degree of differentiation is useful to ­predict survival times in cats. Well‐differentiated tumors had MSTs of 254 days versus 14 days for poorly differentiated tumors.5 Well‐ differentiated tumors are more common. LPLs are rare tumors in animals and humans. In humans this lymphoma as well as other lymphoproliferative diseases are associ­ ated with Waldenstrom‐type macroglobulinemia (IgM) that in about 20% of cases may have cryoglobulinema.1,2 Clinically this protein causes hyperviscosity that results in cold extremities and may require plasma exchange to reverse the syndrome. In animals this syndrome is exceedingly rare but has been associated with mul­ tiple myelomas that produced IgM. A category of MRD includes this syndrome. Plasmablastic lymphomas are uncommon tumors seen in mature dogs and cats. The animals usually present with severe systemic dis­ ease and have profound weight loss and a poor hair coat. On clinical examination multiple nodes are enlarged. Leukemic manifestations are rare but focal lesions in the spleen and perivascular cuffs in the liver occur. The tumor tends to be aggressive. Others may conceiv­ ably classify these as plasmablastic MRD or plasmablastic large cell lymphoma.

Light microscopic features

The features common to plasmacytoma, LPL, plasmablastic lym­ phoma, and MRD are that they are composed of tumors with plasma cell features and they are derived from B lymphocytes.

Histologic and cytologic features of plasmacytomas range from well‐differentiated tumors that are easily recognized to poorly ­differentiated neoplasms that require IHC to confirm the diagnosis (Figure  7.4).5,13 The majority have characteristic plasma cell ­morphology: uniform round cells, moderate amount of basophilic cytoplasm, perinuclear semi‐clear Golgi zone, single or binucleate eccentric uniform nuclei, chromatin that forms aggregates in the centers of nuclei or along the nuclear membrane, and a low mitotic count. The diagnosis on cytology or histopathology of well‐ differentiated­types is straightforward, IHC is not needed and these types of plasmacytomas are benign. Some tumor cells may contain intracytoplasmic inclusions that range from semi‐clear vacuoles to distinct eosinophilic globules or rhomboid‐shaped packets of ­ immunoglobulin (Figure  7.4B,C). These are best appreciated in cytologic preparations and are referred to as Mott cells. The tumors are well delineated and surgical excision is usually curative. Unfortunately, other types have a wider range of cellular and nuclear pleomorphism such that IHC may be required to identify them ­correctly. These types of plamacytomas will have more binucleated cells, multinucleated cells, marked anisokaryosis, no Golgi, no ­cytoplasmic inclusions and may resemble amelanotic melanoma and histiocytic tumors. Plasmacytomas with these features are aggressive and considered malignant. Differentiation of cells was a good means to identify cats with longer survival times.5 Giant multinucleated plasma cells have been described in MRD of cats. Mitotic activity is low in cats with MRD. Only 5 of 26 cases had mitoses seen and when they were found they were 20% plasma cells is a criterion to diagnose myeloma; however, this relative percentage can also be seen with infectious diseases such as ehrlichiosis or leishmaniasis. Each of these usually has lymphocytosis of the marrow and peripheral blood, sometimes marked. Polyclonal gammopathy is common with both but monoclonal gammopathies have rarely been observed, further confounding the correct diagnosis. Persistent high titers to Ehrlichia and other tick‐borne diseases combined with PCR are useful to rule out these differentials. In leishmaniasis there are numerous intracellular organisms in macrophages, so the infec­ tion can be ruled out by cytology. Lymph nodes and spleen The damage to organs other than the bone marrow appears to be due to infiltration with amyloid or M‐protein accumulations rather than actual metastatic tumor in dogs. Metastasis to the spleen does occur. The lymph nodes and spleen are much more likely to be involved with benign reactions of plasma cells, as seen in canine leishmaniasis. In this disease there may be generalized plasmacyto­ sis of bone marrow, lymph nodes, and spleen. Leishmania organ­ isms are very plentiful and easy to see in cytologic or histologic preparations. However, cats will have tumor in abdominal organs, spleen, liver, and lymph nodes and much more frequently than occurs in dogs. The combination of plasma cell proliferation in the marrow, monoclonal gammopathy, and multiple lytic lesions of bones is diagnostic for multiple myeloma in all species.

Cytochemistry and immunohistochemistry

Immune staining is usually not needed for a diagnosis. If performed, the results may vary but most canine cases stain positively with CD79a

in both cytological and histological preparations. CD20 is not reliable in myeloma but the nuclei will stain very well with MUM1.11 CD38 and CD138 are used in humans and myelomas are negative with CD3. Homogeneity of light‐chain restriction can be used in human plasma cell tumors to differentiate neoplasia versus inflammation. This is based on the principle that there is a balanced distribution of lambda and kappa light chains.12 However, in animals (exception is swine) the “normal” ratio is approximately 90:10 lambda to kappa.12 Therefore light‐chain restriction is a part of normal plasma cells and should not be relied on to identify a clonal expansion. If the predominant light chain was kappa, a rare occurrence, then it should represent clonality.

Staging and survival

Treatment of multiple myeloma in animals, as in humans, has not been curative, with most animals succumbing to the tumor. Chemotherapy is used but most cases survive less than 2 years. Factors associated with a poor prognosis are: hypercalcemia, leukemia, azotemia, cytopenias, and Bence‐Jones proteinuria. ­ One parameter that may be useful to predict prognosis is the rela­ tion of the rate of increase in serum M‐globulin to the number of surviving neoplastic cells secreting protein. New treatments in humans have greatly improved the survival of younger patients. A  staging system has been developed for humans based on the levels of β2‐microglobulin and albumin.

References

1. Swerdlow, S.H., Campo, E., Harris, N.L., et  al. (2008) WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. International Agency for Research on Cancer (IARC), Lyon, France. 2. Valli. V.E. (2007) Myeloma. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 189–206. 3. Matus, R.E., Leifer, C.E., MacEwen, G., et al. (1986) Prognostic factors for multiple myeloma in the dog. J Am Vet Med Assoc 188:1288–1292. 4. Tonon, G. (2007) Molecular pathogenesis of multiple myeloma. Hematol Oncol Clin North Am 21:985–1006. 5. Bergsagel, P.L., Kuehl, W.M., Zhan, F., et al. (2005) Cyclin D dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood 106:296–303. 6. Patel, R.T., Caceres, A., French, A.F., and McManus, P.M. (2005) Multiple myeloma in 16 cats: a retrospective study. Vet Clin Pathol 34:341–352. 7. Bienzle, D., Silverstein, D.C., and Chaffin, K. (2000) Multiple myeloma in cats: variable presentation with different immunoglobulin isotypes in two cats. Vet Pathol 37:364–369. 8. Mellor, P.J., Haugland, S., Murphy, S., et al. (2006) Myeloma‐related disorders in cats commonly present as extramedullary neoplasms in contrast to myeloma in human patients: 24 cases with clinical follow up. J Vet Intern Med 20:1376–1383. 9. Edwards, D.F., Parker, J.W., Wilkinson, J.E., and Helman, R.G. (1993) Plasma cell myeloma in the horse. J Vet Intern Med 7:169–176. 10. Thrall, M.A. (2012) Lymphoproliferative disorders and myeloid neoplasms. In Veterinary Hematology and Clinical Chemistry, 2nd edn. (eds. M.A. Thrall, G. Weiser, R.W. Allison, and T.W. Campbell). Wiley‐Blackwell, Ames, Iowa, pp. 179–181. 11. Ramos‐Vara, J.A., Miller, M.A., and Valli, V.E. (2007) Immunohistochemical detection of multiple myeloma 1/interferon regulatory factor 4 (MUM1/IRF‐4) in  canine plasmacytoma: comparison with CD79a and CD20. Vet Pathol 44:875–884. 12. Arun, S.S., Breuer, A.W., and Hermanns, W. (1996) Immunohistochemical examina­ tion of light‐chain expression in canine, feline, equine, bovine and porcine plasma cells. Zentralbl Veterinarmed A 43:573–576.

Marginal zone lymphoma Defining the neoplasm

Marginal zone lymphoma (MZL) originates in lymph nodes or spleen and it is the most common type of lymphoma in the spleen of dogs and humans.1–14 Although it is common in the spleen of

226    Tumors in Domestic Animals

humans, it makes up 10) grade. Most dogs received chemotherapy, however some cases were terminated shortly after the diagnosis. Treatments varied from practice to practice and the stage of the lymphoma was not provided. From the available information the MST of dogs treated for DLBCL was less than 1 year. A few cases survived longer and none survived 2 years or longer. Therefore within the classification of B‐cell lymphoma, most of which will be DLBCL, there are subsets with different prognoses and there are multiple methods that can be used to identify these subsets and predict approximate survival times. In humans, mor­ phologically similar DLBCL can be subdivided by gene expression profiles into GCB and ABC patients that have different survival times.12,13,29 GCB patients have a 5‐year survival of 75% and ABC patients 15% at 5 years. It appears that canine DLBCL or at least large B‐cell lymphomas have germinal center and post‐germinal center subtypes with modestly different survivals; however, the dis­ tinction is not as definitive as in humans and needs clarification.12,13 Molecular classifications of human DLBCL are predictive of bio­ logic behavior and treatment outcomes but are based on hundreds to thousands of analyses. Present data from dogs are based on a low number of cases in which the data are promising.

Tumors of the Hemolymphatic System    245

One 2015 study on high‐grade multicentric lymphomas in dogs did not use phenotyping but used cytology to provide prognostic information.3 Presumably these were a mixture of different lym­ phomas but certainly DLBCL were in the study group. Subjective cytologic features were evaluated in FNA preparations from 20 dogs with high‐grade multicentric lymphoma. In addition, a numerical scoring system was created and morphometry was performed. At diagnosis, multinucleation in the neoplastic cells was associated with decreased survival time and binucleation was associated with shorter remission. None of the other cytologic features evaluated at diagnosis or the scoring system or morphometry were predictive of survival. At relapse, the number of mitoses in lymph node aspirates and the total cytologic scores assigned were greater than those obtained at the time of diagnosis.3 However, the increased total score was due to the increased mitoses and not increased morpho­ logic abnormalities. Mitotic counts as determined in this study did not correlate with survival time or other time points assessed. The overall MST of the dogs followed was 236 days.3 Despite relatively low numbers of cases, no immunophenotyping data, and no staging data, the results suggest that cytologic evaluation for bi‐ and multi­ nucleation at the time of diagnosis has predictive value for survival. Cytology of good‐quality preparations can diagnose 80–90% of lymphomas. Incorporating assessment of bi‐ and multinucleation may provide additional information to help oncologists and owners make decisions. The original paper should be read for the many details and comparisons that were studied.3 Lymphomas can be classified by pathologists as outlined in Box  7.1 and DLBCL in animals can be divided into centroblastic and immunoblastic types. Molecular techniques are being used to subdivide these and other lymphomas to determine their utility as models for similar diseases in humans.12,13,29,30 For future studies we need standardized classification of lymphomas between patholo­ gists and for clinicians to standardize treatment protocols. These data will then need to be correlated with accurate clinical follow‐up, including autopsies on a large number of cases, if we are to deter­ mine reliable survival times, prognoses, and effective therapeutic protocols. It is unlikely there will be a sufficient number of cases in which no treatment is received and the dogs are allowed to live out their lives to compare survival times between treated and non­ treated dogs. A confounder in all types of clinical studies will be the  level of commitment to continue therapy and how different ­clinicians and owners will use the option of euthanasia.

References

1. Valli, V.E., San Myint, M., Barthel, A., et al. (2011) Classification of canine malig­ nant lymphomas according to the World Health Organization criteria. Vet Pathol 48:198–211. 2. Valli, V.E., Kass, P., San Myint, M., and Scott, F. (2013) Canine lymphoma: The effect of age, stage of disease, subtype of tumor, mitotic rate and treatment protocol on overall survival. Vet Pathol 50:738–748. 3. Munasinghe, L.I., Kidney, B.A., MacDonald‐Dickinson, V., et al. (2015) Evaluation of lymph node aspirates at diagnosis and relapse in dogs with high‐grade multicen­ tric lymphoma and comparison with survival time. Vet Clin Pathol 144:310–319. 4. Sato, H., Fujino,Y., Chino, J., et al. (2014) Prognostic analyses on anatomical and morphological classification of feline lymphoma. J Vet Med Sci 76:807–811. 5. Chino, J., Fujino, Y., Kobayashi, T., et al. (2013) Cytomorphological and immuno­ logical classification of feline lymphomas: clinicopathological features of 76 cases. J Vet Med Sci 75:701–706. 6. Valli, V.E., Jacobs, R.M., Norris, A., et al. (2000) The histologic classification of 602 cases of feline lymphoproliferative disease using the National Cancer Institute working formulation. J Vet Diagn Invest 12:295–306. 7. Meyer, J., DeLay, J., and Bienzle, D. (2006) Clinical, laboratory, and histopathologic features of equine lymphoma. Vet Pathol 43:914–924.

8. Durham, A.C., Pillitteri, C.A., San Myint, M., and Valli, V.E. (2012) Two hundred three cases of equine lymphoma classified according to the World Health Organization (WHO) classification criteria. Vet Pathol 50:86–93. 9. McDonough, S.P., Van Winkle, T.J., Valentine, B.A., et al. (2002) Clinicopathological and immunophenotypical features of canine intravascular lymphoma (malignant angioendotheliomatosis). J Comp Pathol 126:277–288. 10. Santagostinol, S.F, Mortellarol, C.M., Boracchi, P., et  al. (2015) Feline upper respiratory tract lymphoma: site, cyto‐histology, phenotype, FeLV expression, and prognosis. Vet Pathol 52:250–259. 11. Mudaliar, M.A., Haggart, R.D., Miele, G., et al. (2013) Comparative gene expres­ sion profiling identifies common molecular signatures of NF‐kB activation in canine and human diffuse large B cell lymphoma (DLBCL). PLOS ONE 8:e72591. 12. Frantz, A.M., Sarver, A.L., Ito, D., et al. (2011) Molecular profiling reveals prognos­ tically significant subtypes of canine lymphoma Vet Pathol 50:693–703. 13. Richards, K.L., Motsinger‐Reif, A.A., Chen, H.W., et al. (2013) Gene profiling of canine B‐cell lymphoma reveals germinal center and post germinal center subtypes with different survival times, modeling human DLBCL. Cancer Res 73:5029–5039. 14. Valli, V.E. (2007) Diffuse large B‐cell lymphoma. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 238–258. 15. Moore, P.F., Moore, P.F., Rodriguez‐Bertos, A., and Kass, P.H. (2012) Feline gastro­ intestinal lymphoma: mucosal architecture, immunopheotype, and molecular clonality. Vet Pathol 49:658–668. 16. Ponce, F., Marchal, T., Magnol, J.P., et al. (2010) A morphological study of 608 cases of canine malignant lymphoma in France with a focus on comparative similarities between canine and human lymphoma morphology. Vet Pathol 47:414–433. 17. Grindem, C.B., Page, R.L., Ammerman, B.E., et  al. (1998) Immunophenotypic comparison of blood and lymph node from dogs with lymphoma. Vet Clin Pathol 27:16–20. 18. Vernau, W., Valli, V.E.O., Dukes, T.W., et al. (1992) Classification of 1,198 cases of bovine lymphoma Vet Pathol 29:183–195. 19. Little, L., Patel, R., and Goldschmidt, M. (2007) Nasal and nasopharyngeal lym­ phoma in cats: 50 cases (1989–2005). Vet Pathol 44:885–892. 20. Flatland, B., Fry, M.M., Newman, S.J., et al. (2008) Large anaplastic spinal B‐cell lymphoma in a cat. Vet Clin Pathol 37:389–396. 21. Higgins, M.A., Rossmeisl, J.H. Jr., Saunders, G.K., et al. (2008) B‐cell lymphoma in the peripheral nerves of a cat. Vet Pathol 45:54–57. 22. Tasca, S., Carlil, E., Caldin, M., et al. (2009) Hematologic abnormalities and flow cytometric immunophenotyping results in dogs with hematopoietic neoplasia: 210 cases (2002–2006). Vet Clin Pathol 38:2–12. 23. Rao, S., Lana, S., Eickhoff, J., et al. (2011) Class II major histocompatibility com­ plex expression and cell size independently predict survival in canine B‐cell lym­ phoma. J Vet Intern Med 25:1097–1105. 24. Williams, M.J., Avery, A.C., Lana, S.E., et  al. (2008) Canine lymphoproliferative disease characterized by lymphocytosis: immunophenotypic markers of prognosis. J Vet Intern Med 22:596–601. 25. Ponce, F., Magnol, J.P., Ledieu, D., et al. (2004) Prognostic significance of morpho­ logical subtypes in canine malignant lymphomas during chemotherapy. Vet J 167:158–166. 26. Rebhun, R.B., Lana, S.E., Ehrhart, E.J., et al. (2008) Comparative analysis of sur­ vivin expression in untreated and relapsed canine lymphoma. J Vet Intern Med 22:989–995. 27. Ruslander, D.A., Gebhard, D.H., Tompkins, M.B., et al. (1997) Immunophenotypic characterization of canine lymphoproliferative disorders. In Vivo 11:169–172. 28. Aresu, L., Martini, V., Rossi, F., et al. (2013) Canine indolent and aggressive lym­ phoma: clinical spectrum with histologic correlation. Vet Comp Oncol DOI: 10.1111/vc0.12048 29. Ito, D., Frantz, A.M., and Modiano, J. (2014) Canine lymphoma as a comparative model for human non‐Hodgkins lymphoma: recent progress and application. Vet Immunol Immunopathol 159:192–201. 30. McCaw, D.L, Chan, A.S, Stegner, A.L., et  al. (2007) Proteomics of canine lym­ phoma identifies potential cancer‐specific protein markers. Clin Cancer Res 13:2496–2503.

T‐cell‐rich large B‐cell lymphoma Definition of the neoplasm

T‐cell‐rich large B‐cell lymphoma (TCRLBCL) is a B‐cell lym­ phoma but the diagnostic characteristic is a mixture of lymphocytes and other cells. Early TCRLBCL may have 80–90% small to intermediate‐sized non‐neoplastic T cells, with the rest of the cells being neoplastic large B cells, histiocytes, and connective tissue

246    Tumors in Domestic Animals

cells. The neoplastic large B cells gradually increase in proportion to the other cells. TCRLBCL is the or one of the most common lym­ phomas in horses.1–4 It is fairly common in cats5 but is uncommon in dogs (1% of all lymphomas).6 In cats this lymphoma has been called “feline Hodgkin’s disease” because of the way the tumor spreads and the presence of large, atypical, binucleated B‐cell lymphocytes that vaguely resemble the Reed–Sternberg cells of human Hodgkin’s lymphoma.5 The characteristic that most resembles the human tumor in cats is that the disease spreads only to contiguous lymph nodes and does not skip to nodes in other anatomic regions.5

Epidemiology, occurrence, and clinical presentation

TCRLBCL constitutes about 10% of all feline lymphomas. Most cats are mature, they are in good body condition, typically without loss of weight or appetite.5 Oddly, the presentation is almost always for a single enlarged node in the neck area, usually in a submandibular node and more often on the right side5 (Figure 7.23). Approximately 25% of cats will have multicentric tumors in almost any organ. In the dog it is seen in a peripheral node but may appear in any area of the body, even in the spinal canal. In horses, TCRLBCL is the most common lymphoma, accounting for approximately 40% of equine lymphomas.1–4 The mean age is 10 years and horses appear in good health with minimal loss of weight or appetite.1 This lymphoma shares the slow progression seen in cats but has very different clinical presentations.1–5,7,8 In horses this lymphoma is unique in that the most common problem is subcuta­ neous nodules (see Figure 7.25). There may be a few or up to 100 or more discoid shaped tumors irregularly distributed over the entire body but most are concentrated at the lower area of the neck.1 These lesions are not in the skin but appear to arise along lymphatics in the subcutis. The subcutaneous tumors are typically 2–3 cm in diameter but some may become large, ulcerate the skin and form a mass weighing up to 30 pounds. The skin lesions have regressed during pregnancy and recurred following parturition attributed to  progesterone receptors on the neoplastic cells.7,8 The other

A

presentation is multicentric lymphoma with tumors in the gastroin­ testinal tract and many other organs. If only a few nodules are found and they are excised fully this has proven an effective treatment in some cases.4 Those cases in which the tumor does not recur post excision have survival times of up to 10 years.4

Pathology

Blood and bone marrow Bone marrow was only noted to be involved in 1 of 200 equine cases.3 Similar results seem to be the case in cats and dogs, with only minor and focal involvement of the marrow. There may be mild to moderate anemia in advanced cases but this may be due to anemia of chronic inflammatory diseases rather than direct infiltration in the marrow. Lymph nodes or skin nodules The diagnosis rests on the recognition of a mixed population of cells with a few characteristic large binucleated cells with large central nucleoli and abundant cytoplasm (Figure 7.23). Unlike most lymphomas, TCRLBCL is not a homogeneous population of uni­ form lymphoid cells. Early lesions have more small cells and are T‑cell‐rich. As the tumor progresses the larger B cells increase ­disproportionately (Figures 7.23–7.27). B cells may predominate in some regions of the tumor. The increase of B cells is usually accom­ panied by increased numbers of macrophages. Histiocytic cells are very obvious in some of the equine subcutaneous tumors and they will be accompanied by multinucleated giant cells (Figure  7.25). Some giant cells may have thin, crystalline‐like cytoplasmic inclusions. An unusual, yet characteristic part of this lymphoma are the large B cells that can have marked anisokaryosis (Figures 7.23, 7.27, and 7.28). The nuclei in these cells will range from 2 RBC in diameter to 3–5. The chromatin of the large B cells is peripheralized, which accentuates a single large central nucleolus. The large cells are often binucleated and a few cells will have 3–4 nuclei. Some large cells are necrotic and have contracted brightly pink cytoplasm and a

B

Figure 7.23  T‐cell‐rich large B‐cell lymphoma (TCRLBCL), lymph node, cat. (A) There is a mixture of large and small cells with the large cells having abun­ dant cytoplasm and a single large nucleus and nucleolus. Single cell necrosis is a characteristic of this lymphoma, as is the fine sclerosis dissecting through the tumor, which is why these tumors are solid and not soft on gross examination. (B) TCRLBCL cytological preparation from a lymph node of a different cat has primarily small and intermediate‐sized lymphocytes with a few larger cells that have large nuclei, prominent nucleoli, and abundant cytoplasm. Cytologic and histologic patterns for TCRLBCL are a heterogeneous mix of large and small lymphocytes and other cells, which is not a pattern typical of lymphomas. Therefore the correct diagnosis may be missed on initial exam.

Tumors of the Hemolymphatic System    247

A

B

Figure 7.24  T‐cell‐rich large B‐cell lymphoma (TCRLBCL), cat. (A) CD20: The large cells are selectively marked and the small T cells are unlabeled. (B)

CD3: Small and intermediate‐sized T cells are numerous and strongly positive. The relatively few cells with large nuclei and the vascular and connective tissue cells are negative. When slides labeled with CD20 or CD3 were examined grossly, both stains appeared to be strongly positive. However, on light microscopic evaluation each antibody labeled different cell populations.

pyknotic nucleus. Necrosis of individual tumor cells should be looked for as this feature is always present (Figure 7.23). In all species there is fine sclerosis which renders the tumor cohe­ sive (Figure 7.23). The nodes are firm to palpation and cut surfaces are solid, the typical soft white appearance of lymphoma is not present. Spleen and other organs There may be focal tumors in the spleen of cats but the spleen is rarely involved in other species. TCRLBCL is a lymphoma pri­ marily of lymph nodes but in late stages liver and kidney may be affected, at least in cats. Subcutaneous tumors are unique to horses.

Cytochemistry and immunohistochemistry

The patterns seen with IHC are unusual because of the heteroge­ neity of cells within a tumor and different tumors from the same animal may vary in the proportion of B‐ versus T lymphocytes. The large B cells mark consistently with CD20 and less frequently with CD79 alpha (Figures 7.24, 7.26, and 7.28). CD20 is preferred for horses. The smaller T cells are positive with CD3 and in some cases, or at some stages, CD3‐positive cells will be predominant (Figure 7.24). IHC for the small T cells accentuates the large nuclei of the unstained B cells. IHC is useful to identify cell types but it does not distinguish neoplastic and non‐neoplastic cells. If histo­ chemical stains are done for reticulin there are diffuse fine fibrillar fibers in all areas of nodal and subcutaneous tumors.

Differential diagnosis

The main differentials for TCRLBCL are hyperplasia or an infectious agent. Some cases, especially in horses will have so many multinucleated giant cells and histiocytic cells that stains for acid‐ fast organisms or fungi are performed. These will be negative. Tumors with foci of ischemic necrosis can look inflammatory. Awareness that TCRLBCL is a mixture of cells combined with the large nuclei and anisokaryosis that is present in the neoplastic large cell population helps focus the diagnosis on lymphoma. A high index of suspicion for this lymphoma in cats and horses is helpful, especially if the lesion is a subcutaneous nodule from a horse. The

Figure  7.25  T‐cell‐rich large B‐cell lymphoma (TCRLBCL), lymph node, horse. The tumor is a mixture of predominantly small cells, T lymphocytes with darkly stained nuclei and fewer large cells, and B lymphocytes with abundant cytoplasm and large, prominent nuclei. Most of the large cells have nuclei with single or multiple nucleoli. Inset: Multinucleated giant cells are a feature of TCRLBCL in horses. This is the most common lymphoma of horses.

large nuclei, binucleation, and prominent nucleoli in large cells are further indications of neoplasia. Clonality detection by PCR should be able to identify clonal antigen receptor genes for B cells if the clonal expansion is greater than 1%. However, the sensitivity and specificity of these assays vary with the expertise of the laboratory. This is a tumor in which expertise with reagents and interpretation is needed.

Evaluation of treatment and survival

Because of the slow rate of progression and the fact that in the cat and dog most of the early cases involve a single node, excision is a consideration. However tumors tend to recur in the same area as the excision. If there is only one or a few subcutaneous tumors then excision may be effective in horses. As stated earlier, some horses respond well to surgical excision and will have long survival times.4 The tumors may be so numerous that excision is not possible.

248    Tumors in Domestic Animals

A

B

Figure 7.26  T‐cell‐rich large B‐cell lymphoma (TCRLBCL), lymph node, horse. (A) CD79a: The large cells are variably marked and the small cells are

universally unlabeled. CD20 is a better B‐cell marker in the horse; consider using multiple antibodies to clarify immunophenotype or flow cytometry. (B) CD3: There is uniform and diffuse staining of the small non‐neoplastic T lymphocytes. In many cases the T lymphocytes may predominate.

Figure  7.27  T‐cell‐rich large B‐cell lymphoma (TCRLBCL), lymph node, dog. The tumor is a mixture of intermediate and small cells with a few very large cells and large nuclei. The pyknotic large cell (left) is a tumor cell undergoing single cell necrosis. Although this is a B‐cell lymphoma, T cells may be the predominant lymphoid cell, at least in different stages of the tumors progression.

Clinical oncology texts and references provide treatment recom­ mendations. In the ACVP study of canine lymphoma there were 9 cases of TCRLBCL.6 These cases had a surprisingly short survival with a range of 89–105 days and a mean of 97 days; however, the follow‐up data in this study need to be confirmed.

Angiocentric B‐cell lymphoma with reactive T cells (lymphomatoid granulomatosis)

Lymphomatoid granulomatosis was the name first used for this unusual lymphoma. This name is still used and it is also classified as angiocentric B‐cell lymphoma. The latter name captures the characteristic histologic distribution of the tumor and part of the oncogenesis of the disease. The neoplasm has angiocentric and angiodestructive histologic patterns.4,9–12 The primary location of this tumor is lung, which is exceedingly unusual for lymphoma. However, it can be found in almost any tissue. The tumor is

heterogeneous morphologically, cellularly, and phenotypically. The large B cells are believed to be the neoplastic population and the numerous T cells a cytotoxic response. Morphologically, it resembles TCRLBCL and the tumor may progress to large cell lymphoma. In humans, the disease is associated with EBV infection. There are human cases in which immunosuppressive treatments caused latent virus to express and induce the tumor, as well as regression of the tumor following removal of the immunosuppressive drugs.4 The disease in dogs has been compared to lymphomatoid granulo­ matosis and the pulmonary form of Hodgkin’s disease in humans.11 Regardless of name(s), it is a lymphoid neoplasm of humans and animals that needs to be characterized further. Grossly and histologically it is well described.9–12 It forms one or more discrete masses in the lung of dogs that can be up to 10 cm or greater in size. It is often only in one lobe, usually a caudal lobe, but there can be multiple masses through the lung and it may be found in other organs. The natural surface is smooth to lobulated. Cut ­surfaces are mottled and blood vessels or airways can be seen trapped in the tumor. In addition to the grossly visible tumor(s), other lesions will be found surrounding blood vessels on histologic sections taken from lung that appeared normal. Histologically, the diagnostic feature is an angiocentric pattern of  heterogenous cells that surround or invade blood vessels (Figures 7.29–7.32). There is an irregular mixture of large and small lymphocytes of mixed phenotype accompanied by binucleate, Reed–Sternberg‐like cells and atypical multinucleated giant cells that have long cytoplasmic tendrils (Figures 7.30 and 7.31). Nuclei in giant cells will be crowded together and in some cells there appears to be hundreds of nuclei. Eosinophils will be numerous in some cases and if histochemical stains are applied many mast cells will be found scattered through the tumor. Histiocytic cells and plasma cells are common. Somewhat like TCRLBCL, the mixed ­cellular infiltration in this disease looks inflammatory on first examination. A distinguishing feature from inflammation is that there are no areas of intensity of the inflammatory cells to form microabscesses or granulomas. Despite the mixture of cell types, the infiltration is uniform and the lesion dissects through tissues the way tumors do. There can be large areas of ischemic necrosis. It is seen primarily in the lung but can occur in the abdominal cavity. In humans this tumor is associated with EBV‐positive large B cells but no type of causation has been defined for dogs. T cells are

Tumors of the Hemolymphatic System    249

A

B

Figure 7.28  T‐cell‐rich large B‐cell lymphoma (TCRLBCL), lymph node, dog. (A) CD20: Tlarge cells are strongly marked as well as cells of intermediate size that are assumed to be of the tumor population. (B) CD3: The smaller neoplastic cells are lightly to moderately marked and the large neoplastic cells (center and left) are unlabeled. This is the typical immunostaining pattern (signature) for this type of lymphoma.

Figure  7.29  Angiocentric B‐cell lymphoma with reactive T cells, Lymphomatoid granulomatosis, lung, dog. The tumor formed multiple large, non‐encapsulated masses that compressed surrounding alveoli. There are large foci of necrosis in center and lower left. Clear foci within the tumor are blood vessels encased by tumor cells: angiocentric pattern.

present in the human lesion, they are not clonal and are believed to be a cytotoxic response. In humans, the large B cells have been shown to have rearranged Ig genes and some may progress into large B‐cell lymphomas. The lesions are graded in humans based on the number of large EBV‐positive B lymphocytes and the higher grades could be a variant of TCRLBCL.12 The immunophenotypic characteristics of the canine disease have been reported and need further clarification.10,11 The cellular infiltrates are positive for B‑cell and T‐cell antigens and the large Reed–Sternberg‐like cells were positive for CD15 and CD30, which is a feature of Hodgkin’s disease.11

References

1. Kelley, L.C. and Mahaffey, E.A. (1998) Equine malignant lymphoma: morphologic and immunohistochemical classification. Vet Pathol 35:241–252.

Figure  7.30  This is a region of the tumor where the giant cells were numerous. They are highly atypical and have very irregular cytoplasmic borders. Nuclei are crowded together, some cells will have what appears to be hundreds of nuclei. The surrounding tissues have fine sclerosis and a mixture of undifferentiated mononuclear cells scattered between reticulin fibers.

2. Meyer, J., DeLay, J., and Bienzle, D. (2006) Clinical, laboratory, and histopathologic features of equine lymphoma. Vet Pathol 43:914–924. 3. Durham, A.C., Pillitteri, C.A., San Myint, M., and Valli, V.E. (2012) Two hundred three cases of equine lymphoma classified according to the World Health Organization (WHO) classification criteria. Vet Pathol 50:86–93. 4. Miller, C.A., Durham,A.C., Schaffer, P.A., et  al. (2015) Classification and clinical features in 88 cases of equine cutaneous lymphoma. J Vet Diagn Invest 27:86–91. 5. Valli, V.E. (2007) T‐cell rich large B‐cell lymphoma. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 260–273. 6. Valli, V.E., San Myint, M., Barthel, A., et al. (2011) Classification of canine malig­ nant lymphomas according to the World Health Organization criteria. Vet Pathol 48:198–211. 7. Henson, K.L., Alleman, R., Kelley, L.C., and Mahaffey, E.A. (2000) Immunohistochemical characterization of estrogen and progesterone receptors in lymphoma of horses. Vet Clin Pathol 29:40–46. 8. Henson, K.L., Alleman, A.R., Cutler, T.J., et al. (1998) Regression of subcutaneous lymphoma following removal of an ovarian granulose‐theca cell tumor in a horse. J Am Vet Med Assoc 9:1419–1422.

250    Tumors in Domestic Animals

9. Jaffe, E.S. and Wilson, W.H. (1997) Lymphomatoid granulomatosis: pathogenesis, pathology and clinical implications. Lymphoma 30:233–248. 10. Smith, K.C., Day, M.J., Shaw, S.C., et al. (1996) Canine lymphomatoid granuloma­ tosis: an immunophenotypic analysis of three cases. J Comp Pathol 115:129–138. 11. Park, H.‐M., Hwang, D.‐N., Kang, B.‐T., et al. (2007) Pulmonary lymphomatoid granulomatosis in a dog: evidence of immunophenotypic diversity and relation­ ship to human pulmonary lymphomatoid granulomatosis and pulmonary Hodgkin’s disease. Vet Pathol 44:921–923. 12. Song, S.Y., Pittaluga, S., Dunleavy, K.,.et al. (2015) Lymphomatoid granulomato­ sis – a single institute experience: pathologic findings and clinical correlations. Am J Surg Pathol 39:141–156.

Figure 7.31  Angiocentric B‐cell lymphoma, dog. Most of the mononuclear

cells have round to oval nuclei and are undifferentiated. Two abnormal mitotic figures are subjacent to the giant cell. Eosinophils were not frequent in this case but can be numerous. Histochemical stains for mast cells will reveal a surprising number scattered through the tumor.

T‐CELL LYMPHOMAS Precursor T‐cell lymphoblastic leukemia/lymphoma Defining the neoplasms

Acute lymphoblastic leukemia (ALL) is a malignancy arising in the bone marrow. The neoplastic cells can be of T‐cell (T‐ALL) or B‑cell (B‐ALL) origin. They look identical cytologically, therefore immunophenotyping is needed to identify the correct cell of origin.1–9 Lymphoblastic lymphoma (LBL) is a lymphoid malig­ nancy that arises in peripheral nodes and it also can be of T‐cell (T‐LBL) or B‐cell (B‐LBL) origin. Canine B‐cell (90%) were of B‑cell origin8 and in other reports the majority were T‐cell.11 These latter authors preferred to designate the neoplasms as lymphoid rather than lymphoblastic but stated the cells “appeared as blast cells.”11 A study of canine leukemia reported approximately 30% B‐ALL and 13% T‐ALL, 35% AML, and 24% acute undifferentiated leukemia.12 Lymphoma LBL originates in lymph nodes and/or lymphoid tissues outside the bone marrow. Most canine LBL are of T‐cell origin and there seems to be agreement on this pattern.9,10 However, if neoplastic cells are in circulation from a dog with LBL, stage V lymphoma (secondary leukemia), the neoplasms were reported to be of B‐cell origin (63%).8 Percentages of distribution of various lymphomas and leukemias will vary between laboratories due to expertise, antibodies used, and how each tumor is defined but hopefully the percentages will be somewhat similar. In the study that reported that approximately 60% of LBL with secondary leukemia were of B‐cell origin it is not clear if the high‐grade lymphoma category seen in 65 of 210 (31%) dogs were a mixture of LBL, DLBCL, or possibly B‐anaplastic large cell or even Burkitt’s‐like lymphomas.8 Criteria to distinguish these diagnoses are published and are in this chapter, but some of the distinguishing features are subjective and therefore the umbrella term of high‐grade lymphoma has practical utility.13 Classifications that divide lymphomas into high, intermediate, or low grade may also add the designation T‐cell or B‐cell, and then under each of these general classifications attempt to list specific lymphomas.7 Therefore, under the general heading of high‐grade T‐cell are: peripheral T‐cell, PTCL‐NOS, and T‐LBL

high‐grade, and under high‐grade B‐cell lymphoma could be: DLBCL, LBL, anaplastic large cell, and Burkitt’s‐like. Specific types of diagnoses are further complicated by the current lack of markers to definitively identify each of these tumors. Molecular character­ ization should help define lymphomas in animals. Each tumor has defining morphologic, cellular, biologic behavior, and immuno­ phenotypic characteristics, but some of the parameters are subjective, and no consensus has been reached. There may be “interpathologist variation,” therefore, as to the type of lymphoma. Nevertheless, some of the most aggressive lymphomas seen in ­animals are ALL or LBL (T‐cell or B‐cell) and have a rapid onset of signs and short survival times of 30% blast cells in bone marrow and tumor burdens are greater in marrow and less in lymph nodes and organs. They can look similar or be identical in blood films and on cytology, especially if AML is poorly differentiated and does not demonstrate maturation. The more differentiated the AML the easier the distinction should be. Granules seen in the cytoplasm of neoplastic cells favor AML. However, granules may be scarce to absent in poorly differentiated AML and granules may be found in LGLs. The granules in LGL may also be ALP positive but they should be granzyme B positive and the immunopheno­ typic profile should be that of lymphoid, T‐cell. Positive reaction for ALP favors AML. The less differentiated the ALL or AML and the wider the organ distribution the more reliant we are on IHC, flow cytometry, and ICC (see section on Immunophenotyping above). The belief is that when ALL or AML has a wide distribution in organs (dissem­ inated) it has progressed and is in a more advanced state. However, AML or ALL confined to bone marrow versus wide organ distri­ bution could be different diseases that have different molecular profiles; one may be programmed for a more aggressive course (wider distribution) than those confined to bone marrow. These are interesting hypotheses, but from a practical view all of these tumors are aggressive and shorten the life of animals with survival times measured in days, and generally less than 4 months for each. CLL is one of the few leukemias with survival times that approach 3 years. The literature provides contrasting data for the prevalence of T‑cell versus B‐cell ALL in dogs.8,11,12 The antibodies selected and the expertise and methodologies used in different laboratories can result in different conclusions. The parameters that researchers use to define B‐ALL, T‐ALL, and AML influence final diagnoses. Until parameters are standardized and definitions agreed upon there will be discrepancies. A battery of antibodies, techniques, and histo­ chemical stains should be used on cases that seem ambiguous. Another approach is to avoid traditional diagnoses and instead define tumors that cause lymphocytosis by an antibody profile, size of the cells in blood, and whether the number of neoplastic cells in blood are greater or lesser than 30,000/μL.15 Briefly, dogs  with CD34‐positive lymphocytosis had survival times of approximately 2 weeks, those CD8 positive (T‐cell) with >30,000 lymphocytes/μL approximately 4 months, those CD8 positive (T‑cell) with 9000/μL, clonal proliferation or homogeneous cells via phenotyping, and either bone marrow >15% mature lympho­ cytes or concurrent cytopenia of at least one cell line. The cytopenia identified most frequently was anemia (50% of 18 cats).10 These authors excluded cats with large cell lymphocytes or if multicentric lymphoma was deemed more likely. Similar criteria have been used to study dogs: lymphocytosis >6000/μL, clonal proliferation or homogeneous cells via phenotyping, negative CD34 and negative testing for ehrlichiosis, leishmania, and other causes of lymphocy­ tosis (Addison’s, epinephrine surge, post vaccine).9 T‐CLL in cats and dogs is an indolent neoplasm composed of small lymphocytes that look cytologically “normal” in blood, but they cause a marked lymphocytosis, often >100,000/μL.4,6,9,10 The lymphocytosis has a wide range, likely dependent when diagnostic techniques intersect the disease, 5000 to >1,000,000/μL. The lym­ phocytes may be small, medium or large, and granulated or not. In cats the lymphocytes are not granular but in dogs granular lympho­ cytes (LGL) are expected in approximately 50–80% of CLL cases.4 CLL composed of LGL have larger cells, pale cytoplasm, larger, round or reniform nuclei, and are more open (chromatin is not as dense). Cytoplasmic granules can be inconspicuous but when found could be confused with myeloid lineage (AML). In general, cytoplasmic granules in LGL cells are easy to see in cats and horses but are smaller and less obvious in dogs. The cytoplasmic granules

A

are easy to see in larger cells and are inconspicuous in small cell types (see Figures 7.40 and 7.41). Visibility is enhanced by Wright– Giemsa stains that have a methanolic base, oil‐immersion objec­ tives, or PTAH. Over 90% of canine CLLs composed of LGL are T‐cell neoplasms and they appear to originate in spleen rather than the marrow. If subtyped, the majority of canine cases are CD8 positive (cytotoxic), a few may be NK cell. The marrow may be involved but usually only late in the progression of the disease. Both dogs and cats can have LGL lymphomas (not CLL) that may or may not have leukemia. In cats, LGL lymphomas usually originate in lymphoid tissue in the gastrointestinal tract.11 These are aggressive but tend to be less so in dogs. CLL composed of LGL is similar to a common and fatal leukemia in Fischer 344 rats.12 Rats with this leukemia typically have white blood cell counts >100,000/μL, severe icterus, massive spleno­ megaly, red blood cell agglutination, erythrophagocytosis, and severe anemia that is immune‐mediated. The leukemia and aggluti­ nation of red blood cells will produce mean corpuscular volumes (MCV) >120 fL. The solid form of this disease is a peripheral lymphoma (SLL). The neoplasm is composed of small lymphocytes, B‐cell or T‐cell, which are indistinguishable until they are phenotyped (Figures 7.36 and 7.37). The majority of SLL are T‐cell but B‐cell types occur. Some cases will have neoplastic cells in blood and the number of lymphocytes can be marked. The line between CLL and SLL becomes blurred when there are neoplastic cells in circulation, bone marrow, spleen, and lymph nodes. Regardless of nomencla­ ture, cases with wider organ distribution and T‐cell lymphocytosis >30,000/μL have a worse prognosis.7 Phenotype needs to be deter­ mined by IHC or flow cytometry, but in general, if a monoclonal gammopathy is identified it suggests the tumor is of B‐cell origin, if the tumor markedly infiltrated the bone marrow it is more likely B‐cell, and if granulated lymphocytes are seen they indicate T‐cell. CLL and SLL should be distinguished from ALL and AML as these latter neoplasms have a much worse prognosis and more aggressive course. CLL and SLL should be distinguished from intestinal LGL lymphoma in cats as many of these have a more aggressive course.

B

Figure 7.36  Small cell lymphocytic lymphoma (SLL), lymph node, dog. (A) Nodal architecture is replaced by a population of uniformly small lymphocytes.

(B) Higher magnification: The neoplastic cells have little cytoplasm, nuclei are approximately the size of red blood cells, chromatin is dense, some nuclei have clefts, and nucleoli are not seen. A mitotic figure is present but overall the mitotic counts in SLL are low.

Tumors of the Hemolymphatic System    257

A

B

Figure 7.37  Small cell lymphocytic lymphoma (SLL), lymph node, dog. (A) CD3: Neoplastic cells are heavily and uniformly positive. There is a fading ger­ minal center at top that is not stained. Inset: CD79a: The lymphoma cells are negative and the non‐neoplastic cells in the fading germinal center are strongly positive. (B) Cytology, lymph node: Cellular and nuclear details are easier to evaluate in cytological preparations than in H&E. Nuclei are only slightly larger than the red blood cells. SLLs generally lack nucleoli and have a narrow envelope of basophilic cytoplasm. Spleen is often neoplastic and splenic or bone marrow involvement may result in a markedly increased, >50,000/μL lymphocytosis. CLL is a tumor of the same cell type but originates in bone marrow and causes moderate to marked lymphocytosis.

Epidemiology, occurrence, and clinical features

Approximately 50% of dogs and cats are asymptomatic and are diagnosed at annual examinations after blood is analyzed and a lymphocytosis is identified.1,6,10 Most affected animals are in good condition but weight loss is reported. Animals presented because of an illness are lethargic with reduced appetite and may have spleno­ megaly, lymphadenopathy, and fever, and very likely the disease has gone undetected for many months. The disease is seen in mature animals and is uncommon in dogs less than 5 years4,6 or cats less than 10 years old.5,6,10 The median age in cats was reported to be 12.5 years, with a range of 5–20 years.10 The median age in 61 dogs was 10, with a range of 5–19 years.4 CLL is identified less often in cattle and horses but likely occurs in most mammals. There do not appear to be any breed‐related tendencies or common causative exposures. Cats with CLL are FeLV negative.

Pathology

Blood and bone marrow The diagnosis of CLL is made from examination of blood with total lymphocyte counts of 50,000–400,000/μL, but incredible ranges of 5,000 to >1,000,000/μL are sometimes seen. If the lymphocytosis is >50,000/μL the diagnosis is fairly easy. In cats, the total median lym­ phocyte count reported was 34,000/μL.10 In dogs median lymphocyte counts range from 36,000 to 166,000/μL. The lymphocytosis is persis­ tent and usually sustained over an extended period of time, >3 months. Lymphocytosis can also be due to non‐neoplastic causes, such as tick‐ borne diseases in dogs, bartonella in cats, hypoadrenocorticism and epinephrine‐induced physiologic response. Physiologic lymphocy­ tosis is more common in cats and young horses but rarely in the range of CLL and it is transient, gone in less than one day if rechecked. The magnitude of lymphocytosis in canine ehrlichiosis, however, can be marked and in the range of CLL.13 The greater the lymphocytosis and the longer it persists the more likely the cause is CLL rather than one of these non‐neoplastic causes. See Differential diagnosis in this section. Approximately 50–75% of dogs and 50% of cats with CLL will have varying degrees of non‐regenerative anemia, but it is not as

severe as with ALL or AML.4,10 Neutropenia was not seen in approx­ imately 180 dogs with CLL4,5,9 and is not reported in cats.4,10 Thrombocytopenia ranged between 10% and 25% in these three studies.4,5,9 Cytopenias are not a feature of CLL but they are expected with ALL and AML. Morphologically, the tumor cells may be granulated or not. Granulated cells are LGL and are T cells. Beyond this, morphology cannot be relied on to identify B‐CLL or CLL or atypical pheno­ types. The cells of CLL are mature and morphologically they look the same as in the non‐neoplastic causes of lymphocytosis. CLL cells have little cytoplasm and small nuclei, 7–9 μm in diameter, which is approximately the size of a canine red blood cell. The chromatin is densely stained without clear areas. Nucleoli are inconspicuous or not visible. The cytoplasm is minimal and lightly stained (Figure 7.37). They look like normal lymphocytes, therefore the key to the diagnosis is the marked lymphocytosis. If CLL is LGL type then the cells are larger, cytoplasm more abundant and pale, and cytoplasmic granules are present (described earlier). Neoplastic cells in ALL and AML can look similar, especially AML without maturation. In CLL the bone marrow is usually neoplastic, >15% lympho­ cytes; however, CLL of LGL cells likely originates in the spleen and bone marrow is not infiltrated. Core biopsy will identify tumor in bone marrow more frequently than aspirational cytology.10 When involved, the marrow is hypercellular, fat is replaced, and 15–90% of the marrow will be CLL cells.1,2 However, the mitotic count is low and dividing cells are not found or are few. The disease is noted to be one of accumulation rather than proliferation, with the neo­ plastic cells having upregulation of the Bcl2 gene that blocks the apoptotic process. There can be focal areas of “reaction centers” in the solid areas of tumor where the dividing cells are of slightly larger type. In the nongranulated form of CLL the marrow is “almost always” involved. Metaphyseal regions of appendicular bones and any axial bone will contain tumor. A bone marrow aspirate or core from the proximal femoral area will have high cellularity and provide a

258    Tumors in Domestic Animals

diagnostic sample. LGL CLL may not be in the bone marrow and samples from the spleen or an enlarged lymph node should contain tumor. LGL leukemia of a larger cell type can arise in the spleen of dogs and LGL lymphoma arises in the intestinal tract of cats. Many of these cases will also have neoplastic cells in blood.11 Lymph nodes The lymph nodes of SLL are diffusely neoplastic. Germinal centers are absent or are reduced to a few fading clusters of mantle cells. The medullary cords are filled and expanded and the medullary sinuses are compressed. Nodes appear dense because the neoplastic cells have little cytoplasm and nuclei are crowded together.3 Nuclei are densely stained and only slightly larger than red cells, with nucleoli not apparent (Figure 7.36). The mitotic count is low, with none in most 400× fields. In general, the nodes have thin capsules and the subcapsular sinus is compressed. Lymph nodes in CLL are often not enlarged and may even be atrophic. Other cases will have enlarged lymph nodes and neo­ plastic cells but the nodes are not as severely involved as with SLL. If CLL is marked, look in the lymphatics of the node as they will be packed with neoplastic cells identical to those in the blood. Megakaryocytes may be present in the collapsed medullary sinuses. The diagnosis of SLL with cytology can be difficult because the lymphocytes are mature. An anatomic pathologist should consult with a clinical pathologist or consider immunophenotyping (flow cytometry or IHC) to assess homogeneity of the cells or PARR for clonality. The cells are uniform, small, mature, and are not the pro­ totypical neoplastic large lymphocytes. There should be no or very few plasma cells and no neutrophils. If granules are seen or sus­ pected, consider staining with a methanolic‐based Wright–Giemsa stain to enhance their prominence. Spleen and liver Splenomegaly is expected in dogs and cats with CLL.2,6,10 The spleen may not be neoplastic in nongranulated CLL as the tumor originates in bone marrow. If the spleen is involved the tumor is usually in the sinus areas, which may be partially or fully occupied by neoplastic lymphocytes. Tumor cells separate the resident smooth muscle trabeculae to varying degrees, depending on the magnitude of the tumor infiltrate. There usually is partial to complete atrophy of the thymic‐dependent periarteriolar lym­ phoid cuffs, and germinal centers are reduced in size. Usually there are no areas of extramedullary hematopoiesis and no sinus histiocytes bearing hemosiderin. The liver is infiltrated in later stages of CLL. Periportal areas and hepatic sinusoids may be filled with CLL cells. In severe cases, ­neoplastic cells can be found in any tissue.2

Cytochemistry and immunohistochemistry

Most CLL and SLL in dogs and cats will have strong cytoplasmic labeling with CD3 and are negative with antibodies for B‐cell markers, CD79 alpha, CD20, and CD21 (Figure 7.37). If subtyped, 75–90% of canine T‐cell CLL are CD8‐positive cytotoxic lympho­ cytes. Of 61 canine cases, 54 (89%) were of T‐cell origin (CD3 positive, CD21 negative), and of these, 49/54 (91%) were LGL (CD3 positive/CD4 negative/CD8 positive), 4 (7%) were CD3 positive/ CD4 negative/CD8 negative, and 1 (2%) was CD3 positive/CD4 positive/CD8 negative.4 Seven of the 61 (12%) canine CLL were of B‐cell origin (CD21 positive).4 A report of 43 CLL indicated 19 were T‐cell and 14 of these were LGL, 17 were B‐cell, and 7 had atypical phenotype, 4 of which were LGL.9 Seven cases had atypical profiles

and these had short survival times.9 A study of 73 dogs reported 73% were T‐cell and 23% B‐cell.5 Forty of the T‐CLL were LGL; all were CD3‐positive and 90% CD8‐positive cytotoxic lymphocytes. These authors also detailed leukointegrin and other CD profiles, including 3 LGL that were double negative. Non‐LGL T‐CLL pro­ files were also characterized.5 No cases were CD34 positive and this is consistent with other reports.5 Almost all of the reported CLL in cats are of T‐helper lympho­ cytes.10 Seventeen of 18 cats with CLL were T‐cell phenotype and 16 of these were CD3‐positive/CD4‐positive/CD8‐negative T‐helper lymphocytes.10 One cat was CD21‐positive B‐cell CLL. Feline CLL cases are CD4‐positive/CD8‐positive T‐helper type which is also consistent with their nongranulated cell type.10 The patterns of immunophenotyping are varied and cannot be predicted from morphology; however, LGL CLL is considered T‑cell and animals with macroglobulinemia B‐cell. Predicting B‑cell, non‐LGL T‐cell, NK, or aberrant phenotypes is not possible from morphology.

Differential diagnosis

AML, ALL, and non‐neoplastic lymphocytosis are the main differ­ entials for CLL. The higher the lymphocytosis and the longer it per­ sists the more likely the diagnosis is CLL. Most antigenic substances, such as a vaccine, stimulate a relatively mild and transient lympho­ cytosis, 30,000 lymphocytes/μL at presentation had significantly shorter median survival (130 days) than dogs with 90% will be T‐cell and a few are considered to be NK.2,3,5,14 They may be further subdivided by size of cells or a ­profile of leukocyte antigen expressions.5,8,14 The biologic behavior of LGL tumors in cats is the same regardless of subtypes. It is an aggressive lymphoma that shortens the cat’s life, and the diagnosis can be determined from cytology or histopa­ thology. In dogs, subtyping LGL tumors has not been correlated with survival or treatments. In dogs, LGL CLL or lymphoma can be indo­ lent, with lymphocytosis lasting years or, in some cases, progressing rapidly to cause death. The latter is associated with larger and more immature cells types. In humans, biologic behavior seems to follow phenotype, CD3‐positive/CD8‐positive types are cytotoxic T‐cell lineage and are indolent. CD3‐negative types are NK lineage and have an acute aggressive course with marked hepatosplenomegaly. There are ample data to conclude that LGL leukemia originates in splenic red pulp.2,5,16 In all species intestinal LGL lymphoma is believed to arise from intraepithelial LGL. This is enteropathy‐ associated intestinal T‐cell lymphoma (EATCL),3,5 which in humans is believed to originate in LGL cells that have undergone a clonal transformation from prolonged antigenic stimulation induced by a variety of inflammatory bowel diseases (IBD). Cats and dogs have IBD and intestinal lymphoma and the association of the two diseases is often suggested but has never been proven. The tumor in cats is aggressive but in some dogs it can have a more indolent course. LGL leukemia or lymphomas have been seen in dogs,2,4 cats,5,6 horses,7–9,17 cows, rodents,10–12 and birds.13 LGL can cause lymphoma or leukemia (acute or chronic) or both and they may increase in a variety of inflammatory or infectious diseases (e.g. canine ehrlichiosis).1 Retroviral particles have been reported in a cell line derived from a canine LGL leukemia,18 but they are likely not causative.

262    Tumors in Domestic Animals

A

B

C

D

Figure 7.39  Large granular lymphocytes (LGL), cat, intestine. (A) The prominent eosinophilic cells are LGL. These cells have been called globular leuko­ cytes. Only in cats and horses are LGL this obvious in H&E. (B) Same section, CD3: The globular eosinophilic intracytoplasmic inclusions (arrows) seen in H&E are strongly positive, as are other intraepithelial lymphocytes and a few lymphocytes in the lamina propria (arrow head). (Image courtesy of Luke Borst, North Carolina State University.) (C,D) LGL, dog blood. Comparison of aqueous (C) and methanolic (D) Romanowsky automated stains. Intracytoplasmic granules stain poorly with the aqueous stain (C) and are not visible but are prominent and easily seen with the methanolic stain (D). Granules in canine LGL tend to be smaller and finer than in cats and horses but there are wide ranges of size and numbers per cell and per species. A few nuclei have characteristic irregular contours with indentations. (C,D Images courtesy of Robin Allison, Oklahoma State University.)

In dogs

Epidemiology and occurrence

LGL diseases in dogs occur in mature and older dogs with a mean age of 10 years, usually of large‐breed types. They present with a variety of signs that may involve the gastrointestinal system and may have a history of weight loss and depression of variable degree. Females are nearly twice as often affected as males.

Pathology

Most dogs with this disease present with an LGL lymphocytosis of 5000 or more/μL. Animals with a clonal LGL tumor likely had lym­ phocytosis for at least 3 months, which makes causes due to inflam­ mation less likely, but duration of the lymphocytosis is usually not known. Neoplasms of LGL may produce a T‐cell CLL or T‐cell lym­ phoma (rarely NK) with or without neoplastic cells detected in the blood. The key to the diagnosis of LGL neoplasms is to recognize

cytoplasmic granules in the neoplastic lymphocytes. These granules are usually near the nucleus, and vary in number from 3–20/cell. In general, they are not as obvious in dogs as they are in cats and horses (Figures 7.40, 7.41, and 7.54). Granules are easier to find in cytologic then histologic slides (see this section on cats). Cytologically, LGL cells are of intermediate to large size with nuclei about 1.5–3 RBC in diameter, round to oval or indented, reniform or infrequently with clefts (Figures  7.40 and 7.41). In histopa­ thology the nuclei appear round and the reniform pattern is only seen well in cytologic preparations. The cytoplasm is fairly abundant and lightly stained in cytology and unapparent in histopathology. Splenomegaly is usually present and may be palpable at initial examination or detected by imaging. Peripheral lymphadenopathy is unusual but may be present. The LGL cells may be present in the marrow but not causing phthisis of normal marrow cells. Unlike the human or feline LGL tumors, the peripheral blood neutrophil

Tumors of the Hemolymphatic System    263

50 um

A

B

Figure 7.40  (A) Large granular lymphocytic (LGL) leukemia in 6‐year‐old cat, lymphocytosis of 64,100/μL. Results of CBC prompted an abdominal

ultrasound. A mass was found in the intestines and mesenteric lymph nodes were enlarged. Treatment was started but the cat declined rapidly and was euthanized. Cats with this combination of intestinal and leukemic LGL have short survival times. Nuclei are 3–4 RBC in diameter, the indentations are prominent, and the granules are juxtanuclear. (Image courtesy of Jessica Bailey, Auburn University.) (B) Large granular lymphocytic (LGL) leukemia, dog. Two‐year‐old English setter presented with intense pruritus. The mucous membranes were pale and there were numerous petechiae. The leukocyte count was 243,000/μL and >90% were LGL, T‐cell. The large neoplastic cells have pink granules (arrows) and basophilic cytoplasm. The cytoplasmic granules are diagnostic for LGL, but they could be confused with myeloid cells, especially since these nuclei have indentations and the granules are fine. In histopathology nuclei look round and the reniform pattern is only seen well in cytologic preparations. Cytoplasmic granules are also difficult to see in H&E. Note mitotic figure (lower left) which is extremely unusual to find in the peripheral blood, but this is a thick area of the film where cells are concentrated and the white blood cell count was incredibly high. Some dogs and rats with LGL leukemia will have concurrent immune hemolytic anemia, but in this image the stacks of red blood cells are rouleaux not agglutination.

numbers in dogs stay within normal limits or there is a mild neutro­ philia. Mild to moderate non‐regenerative anemia (PCV 20–30%) is present in about half of the dogs. The LGL tumor cells or persistent lymphocytosis may be pheno­ typed via IHC, ICC, or flow cytometry. Depending on the expertise, funds, and how detailed a subtype category is desired, a broad panel of proven antibodies should be used. At this time subtyping LGL has not been correlated with clinical outcomes or different treat­ ments in dogs.

Cytochemistry and immunohistochemistry

Almost all LGL tumors in dogs are positive with CD3 (>90%), and those that are negative are assumed to be of NK cell type (3 months) LGL lymphocy­ tosis characterized the immunophenotypic profile of the LGL cells.2 Diagnoses included LGL leukemia, ehrlichiosis, reactive lympho­ cytosis, and persistent lymphocytosis with or without anemia. Clonality was not determined. The original reference should be read for the details provided.2 More than 90% of the LGL were CD3 positive and all were CD21 negative, indicating T‐cell phenotype. The majority (60%) of the LGL cases had an αβ type of T‐cell receptor with approximately 32% having the γδ type of T‐cell receptor; the two negative cases were considered NK (8%).2 In this series all cases had LGL cells that were positive for CD18, CD11a, >90% were positive with CD45Ra, CD11d and CD11b was absent

and CD11c present in about two‐thirds of cases with all of these antigens on the cell surface and not in cytoplasm.8 The pattern of integrin αdβ2 (CD11d) expression was distinctive. It was present in over 90% of cases, suggesting splenic origin. CD11d is the leukoin­ tegrin expressed by macrophages and T cells of the splenic red pulp and by peripheral blood LGLs. To put this level of expression in focus, only about 1% of peripheral blood lymphocytes in normal dogs have the αdβ2 + γδ leukointegrin. However, in the sinus areas of the spleen over a third of the lymphocytes present have this signa­ ture type of leukointegrin. The combination of CD11d expression, splenomegaly, and the marrow only lightly involved suggested the spleen was the source of the LGL cells.2 Recently a gamma/delta (TCRγδ) T‐cell LGL tumor was identi­ fied in a dog with lymphoma in the mediastinum and thoracic lymph nodes.16 An extensive panel of antibodies were used in flow cytometry on neoplastic cells aspirated from the mediastinal mass. The results were used to characterize this LGL lymphoma and ­document the first report of a γδ T‐cell LGL in the dog. The medi­ astinal tumor was CD11d negative, which is the cluster of differentiation antibody that is associated with cells in the splenic red pulp, macrophages, T cells, and LGL. Although the tumor was present in the spleen and the spleen is a common origin of γδ lym­ phomas, the negative CD11d result prompted the authors to suggest that this LGL tumor originated in the mediastinum or possibly liver.16 The dog survived less than a month with treatments. Too few cases of γδ T‐cell lymphomas in dogs have been followed to predict accurate survival data, but it appears γδ T‐cell lymphomas are aggressive with short survival times post diagnosis. LGLs contain perforin‐ and granzyme B‐positive granules. Granzyme B enzyme is a serine protease that induces apoptosis and is found in the granules of LGL of both T‐cell and NK‐cell origin. Granzyme B is a useful marker for LGL.

264    Tumors in Domestic Animals

A

B

C

D

Figure 7.41  Large granular lymphocyte (LGL) intestinal lymphoma, horse. (A) Mesenteric node. (B) Intestine. (C) Touch imprint. LGL lymphoma in the

horse typically affects abdominal organs, such as intestine, liver, and lymph nodes. This is a large cell LGL, the nuclei are large (2–4 RBC in diameter), and cytoplasm is abundant. Many cells contain obvious intracytoplasmic granules typical of LGL in horses and cats. Non‐neoplastic lymphocytes are smaller, the nuclei are darkly stained and no cytoplasm is visible. LGL lymphoma is a T‐cell neoplasm. They will be CD3 positive and granzyme B positive and were misinterpreted as globular leukocyte tumors. (A–C Images courtesy of Allison Boone and Jennifer Neel, NCSU.) (D) Granzyme B‐positive cells in a horse with LGL. Inset: Higher magnification.

Differential diagnosis

If the diagnosis of lymphoma is apparent, then search for gran­ ules in neoplastic cells, preferably in cytologic preparations with alcohol‐based stains.15 The detection of cytoplasmic granules is the key to diagnose LGL neoplasms and distinguish them from other lymphomas. If lymphocytosis is present, it is helpful to monitor or determine how long the LGL cells have been increased in the blood. In humans and in animals it is assumed that if the increase in peripheral blood LGL has been present for 3 months or longer, the condition is a clonal malignancy of LGL cells. PCR for T‐cell gene rearrangement can be performed to determine clonality. The most important differential for LGL leukemia or LGL lym­ phoma with secondary leukemia is an inflammatory or antigenic reactive lymphocytosis.1 Dogs with tick‐borne diseases, especially canine ehrlichiosis, will have clinical signs and blood parameters similar to dogs with neoplasms of LGL. Further complicating the distinction is that the lymphocytes in canine ehrlichiosis can be LGL and rare cases will have a monoclonal gammopathy. Thrombocytopenia can be a feature of tumors and is common in

ehrlichiosis. Hypoadrenocorticism may cause lymphocytosis, but electrolyte patterns and ACTH stimulation will rule in or rule out this differential. Cats with bartonella or an epinephrine surge may have lymphocytosis. These differentials can be ruled out through serology, visualization of organisms, repeat blood work, PCR for specific infectious diseases, response to doxycycline and, if needed, immunophenotyping to determine homogeneity of lym­ phocytes or clonality determination via PCR for gene rearrange­ ment (PARR). Hepatosplenic lymphoma looks similar to LGL lymphoma as they both can have enlarged mesenteric lymph nodes, hepato­ splenomegaly, and the spleen can be the origin of both tumors. The neoplastic cells in both are CD3, CD11d, and granzyme B positive. Gamma/delta T‐cell is the subtype of lymphocyte in hepatosplenic lymphomas, but now there is one report of γδ T‐cell that was an LGL.16 The neoplastic cells in hepatosplenic lymphoma are not reported to be granulated and they exhibit erythrophagocytosis; LGL lymphomas rarely do so although they will in the F344 rat. The liver should be more severely involved in hepatosplenic lymphoma.

Tumors of the Hemolymphatic System    265

Survival

In dogs it is believed that tumors of LGL follow an indolent course, but too few cases have been followed to provide accurate and ­predictive data. Also, the diseases reported may include reactive lymphocytosis, leukemia, or lymphoma. In humans, most cases seem indolent, but phenotyping is used to help predict indolent versus malignant behavior. Long‐term follow‐up correlated with morphology and immunophenotyping has been done in humans but not in animals. There are cases of LGL leukemia and lymphoma in dogs that are rapidly progressive,8,16 but LGL lymphocytosis appears to follow an indolent course in most dogs. In cats LGL neo­ plasms have an aggressive course, especially if there is leukemia and/or transmural intestinal LGL lymphoma.5,19 Concurrent problems in dogs are anemia and splenomegaly. Cytopenias do not appear to be a complicating factor. Some LGL tumors respond to chemotherapy and/or steroids, and the dogs that respond are relatively long survivors. Infectious causes of lympho­ cytosis need to be ruled out. In cats

Epidemiology, occurrence, and clinical signs

Cats with LGL diseases have a wide age range and the mean is approx­ imately 10 years.5,6,19 There is almost a 3:1 female predominance. They present because they are sick; this is not an indolent tumor in cats. Common problems are weight loss, anorexia, abdominal masses, enlarged mesenteric lymph nodes, hepatosplenomegaly, and reno­ megaly. Less common problems are icterus, diarrhea, vomiting, body cavity effusions, and peripheral lymphadenopathy.5 Lymphoma in the abdomen is a common differential diagnosis for cats with these types of abnormalities. Some cats will have a history of IBD. Cats with LGL diseases are almost always FeLV and FIV negative. A diagnosis of LGL lymphoma/leukemia can often be made from examination of a blood film. Lymphocytosis ranges from mild (100,000/μL) and some cases will be >300,000/μL in blood. In one study, 18 of 21 cats had neoplastic LGL cells in the blood, although sample bias is likely as the results of blood cell counts were used to find cases of LGL.5 Anemia is ­present in about one‐third and cytopenia of leukocytes is not expected. In fact, feline LGL diseases will have neutrophilia and a paraneoplastic mechanism has been speculated to produce this neutrophilia.5 Serum hepatic enzymes and bilirubin are increased in about half of the cases but there are no characteristic patterns. The neoplasm is not associated with FeLV or FIV; both are negative in almost all cases.5,6 FeLV‐associated lymphomas are usually mul­ ticentric or mediastinal and in young cats. FIV‐associated tumors in cats are usually high‐grade B‐cell lymphomas.

Pathology

LGL can produce lymphoma and some of these cases will have neoplastic cells in the circulation. LGL may also be a leukemia and some of these cases will have neoplastic cells in solid tissues. LGL lymphoma appears to arise most frequently in the intestinal tract in cats.19 Cats with LGL lymphoma will have thickened intestines, usually jejunum and ileum, but other regions can be involved. The size of the intestines, white discoloration imparted by the tumor and transmural involvement all vary with the severity of the neo­ plasm. Invariably, the lamina propria is filled with neoplastic LGL cells. In most cases, the tumor extends into the submucosa, in many it extends into muscle layers, and some will be transmural. The tumor has an epitheliotropic pattern that may be obvious or may need to be searched for and highlighted by immunostaining

for CD3.19 Numerous neoplastic LGL cells will be seen between intestinal epithelial cells, but this pattern is never as obvious as the infiltration in the lamina propria. Mesenteric lymph nodes are enlarged and neoplastic in all cases. The tumors will efface nodes and fill cortex and medulla in about half of the cases. Liver was neo­ plastic in 12 of 13 cases and the spleen in 8 of 13.5 The kidneys will be involved bilaterally in severe cases, and if enough parenchyma is compromised the cats will be azotemic. Other sites that may be involved include the skin, liver, spleen, and bone marrow. Less than 25% will have enlarged peripheral lymph nodes and a few will have effusions in thoracic or abdominal cavities. The effu­ sions contain neoplastic LGL cells and cytofuge preparations of the fluid provides excellent visualization of the cytoplasmic granules. Bone marrow is only lightly infiltrated by the LGL cells in cases of lymphoma and myelophthisis by the neoplasm does not appear to be a clinical problem. Gross and histologic sections will establish the diagnosis of lym­ phoma easily. However, the diagnostic cytoplasmic granules are not easily seen with H&E‐stained slides. Cytologic preparations of blood, effusions, or touch impressions of mesenteric nodes or intestinal lesions will reveal the granules and therefore establish the diagnosis of LGL lymphoma easier than histopathology. The identification of LGL is important clinically as nongranulated forms of gastrointestinal lym­ phoma, especially small cell types that are mucosal, have a more protracted course than the rapidly progressive LGL lymphoma.19 If a diagnosis of lymphoma is made from biopsy material of mesenteric lymph node or gastrointestinal specimens from a cat, consider exami­ nation of cytologic preparations from the same specimens and/or blood films to look for granules. Slides stained with methanolic‐based stains are superior to stains with an aqueous base (see Figure 7.39).15 Flow cytometry to search for and characterize neoplastic cells in circulation is helpful as well. If none of these are available then search for eosinophilic granules at 100× magnification in H&E‐stained sec­ tions or purple granules in PTAH‐stained sections. Two cell types are recognized by morphology. Phenotypically there may be several cell types, but the biologic behavior of all is similar, comprising an aggressive course with survival measured in days to weeks post diagnosis.5,19 The smaller cell type of LGL has a cell diameter of 8–15 μm, and the cytoplasmic granules are small and inconspicuous. They have round nuclei, often with an indenta­ tion, and the chromatin is coarse and dense. Nucleoli are not usu­ ally visible. The cytoplasm and nuclei impart a mature cell appearance. The other cell type is larger, 15–35 μm in diameter, the cytoplasm is lightly stained, and the cytoplasmic granules range from small to large and are distinctive. These cells look immature, they have more cytoplasm, and the nuclei are larger; the chromatin is also less dense and nucleoli are usually visible. The granules are much larger than those of the small cell type. Some will be 3.0 μm and are obvious in cytologic preparations, especially if examined with oil immersion (Figure  7.41). The granules stain dark purple with Wright–Giemsa, not as well with Diff‐Quik, and in H&E‐ stained slides the granules are much less obvious. They are eosino­ philic in H&E and purple in PTAH but not all tumors will react with PTAH. They are seen most easily with oil immersion in cyto­ logic preparations of blood, body cavity effusions, or touch imprints of tumors. All LGL tumors should be granzyme B positive.

Cytochemistry and immunohistochemistry

The neoplastic cells are positive with the leukocyte marker CD18. Most cells in an LGL lymphoma stain positively with CD3 (approximately 90%) and are negative with multiple B‐cell markers.

266    Tumors in Domestic Animals

The minority that are negative for both are assumed to be of NK‐cell type.5 The majority of the T‐cell lymphocytes are cytotoxic lympho­ cytes (CD8 alpha positive) and a few are helper T lymphocytes (CD4 positive). CD11d is associated with splenic cells and was only expressed by approximately 25% of feline LGL lymphomas, suggest­ ing they are not of splenic origin in cats as they are believed to be in dogs. Intestinal origin is more likely in cats, especially given the gross and histologic distribution of the tumors and the biology of LGL.5,19 Original manuscripts contain details of leukointegrin pro­ files and their interpretation.5,16,19

References

1. Weiser, M., Thrall, M., Fulton, R., et al. (1991) Granular lymphocytosis and hyper­ proteinemia in dogs with chronic ehrlichiosis. J Am Anim Hosp Assoc 27:84–88. 2. McDonough, S.P. and Moore, P.F. (2000) Clinical, hematological, and immuno­ phenotypic characterization of canine large granular lymphocytosis. Vet Pathol 37:637–646. 3. Chan, W.C., Foucar, K., Morice, W.G., and Catovksy, D. (2008) T‐cell large gran­ ular lymphocytic leukemia. In WHO Classification of Tumours of Haematopoietic and Lymphoid Tissue (eds. S.H. Swerdlow, E. Campo, N.L. Harris, et  al.). IARC Press, Lyon, France, pp. 272–273. 4. Wellman, M.L., Couto, C.G., Starkey, R.J., and Rojko, J.L. (1989) Lymphocytosis of large granular lymphocytes in three dogs. Vet Pathol 26:158–163. 5. Roccabianca, P., Vernau, W., Caniatti, M., and Moore, P.F. (2006) Feline large gran­ ular lymphocyte (LGL) lymphoma with secondary leukemia: primary intestinal origin with predominance of a CD3/CD8αα phenotype. Vet Pathol 43:15–28. 6. Wellman, M.L., Hammer, A.S., DiBartola, S.P., et al. (1992) Lymphoma involving large granular lymphocytes in cats: 11 cases (1982–1991). J Am Vet Med Assoc 201:1265–1269. 7. Kramer, J., Tornquist, S., Erfle, J., and Sloeojan, G. (1993) Large granular lympho­ cyte leukemia in a horse. Vet Clin Pathol 22:126–128. 8. Grindem, C.B., Roberts, M.C., McEntee, M.F., and Dillman, R.C. (1989) Large granular lymphocyte tumor in a horse. Vet Pathol 26:86–88. 9. Herraez, P., Berridge, B., Marsh, P., et al. (2001) Small intestine large granular lym­ phoma in a horse. Vet Pathol 38:223–226. 10. Losco, P.E. and Ward, J.M. (1984) The early stage of large granular lymphocyte leukemia in the F344 rat. Vet Pathol 27:186–291. 11. Stromberg, P.C., Rojko, J.L., Vogtsberger, L.M., et  al. (1983) Immunologic, biochemical, and ultrastructural characterization of the leukemia cell in F344 rats. J Natl Cancer Inst 71:173–181. 12. Miyajima, R., Hosoi, M., Yamamoto, S., et al.(1999) Eosinophilic granulated cells comprising a tumor in a fischer rat. Toxicol Pathol 27:233–236. 13. Patnaik, A.K. (1993) Histologic and immunohistochemical studies of granular cell tumors in seven dogs, three cats, one horse, and one bird. Vet Pathol 30:176–185. 14. Williams, M.J., Avery A.C., Lana, S.E., et  al. (2008) Canine lymphoproliferative disease characterized by lymphocytosis: immunophenotypic markers of prognosis. J Vet Intern Med 22:596–601. 15. Allison, R.W. and Velguth, K.E. (2010) Appearance of granulated cells in blood films stained by automated aqueous versus methanolic Romanowsky methods. Vet Clin Pathol 39:99–104. 16. Ortiz, A.L., Carvalho, S., Leo, C., et al. (2015) Gamma delta T‐cell large cell gran­ ular lymphocyte lymphoma in a dog. Vet Clin Pathol 44:442–447. 17. Mastorilli, C., Cesar, F., Joiner, K., et al. (2015) Disseminated lymphoma with large granular lymphocyte morphology diagnosed in a horse via abdominal fluid and transtracheal wash cytology. Vet Clin Pathol 44:437–441. 18. Ghernati, I., Corbin, A., Chabanne, L., et al. (2000) Canine large granular lympho­ cyte leulemia and its derived cell line produce infectious retroviral particles. Vet Pathol 37:310–317. 19. Moore, P.F., Moore, P.F., Rodriguez‐Bertos, A., and Kass, P.H. (2012) Feline gastro­ intestinal lymphoma: mucosal architecture, immunopheotype, and molecular clonality. Vet Pathol 49:658–668.

Peripheral T‐cell lymphoma not otherwise specified

Peripheral T‐cell lymphoma not otherwise specified (PTCL‐NOS) is a classification used for T‐cell lymphomas that are not yet charac­ terized or that cannot be fully classified.1 In human oncology the tumors under this umbrella have decreased as subtypes within PTCL‐NOS are defined as disease entities, so the designation NOS

is often not used and the specific diagnosis provided. This is also happening in veterinary pathology as we recognize the different T‑cell lymphomas. Classifications in humans use IHC and molecular markers as well as known disease patterns, biologic behaviors, and histopa­ thology to characterize the diseases. In veterinary pathology we do not have all the tools or economic resources needed to characterize these lymphomas and we lack accurate follow‐up data, which is a critical component to characterize the lymphoma as a disease entity. Despite these limitations many of the tumors in this classification have now been recognized and characterized, at least to the limit of our antibodies and resources, although the number of animals in some groups are small. Subtypes of PTCL include: angioimmunoblastic T‐cell lym­ phoma (AILT), angiocentric T‐cell lymphoma, hepatosplenic, intestinal T‑cell, subcutaneous panniculitis type, mycosis fungoi­ des, mature T‐cell leukemia/lymphoma, anaplastic large cell lym­ phoma, and T‐zone lymphoma (TZL).1 TZL is an uncommon lymphoma in humans and is listed under PTCL‐NOS, however, it is described as a distinct diagnosis in this chapter and is listed indi­ vidually in Box 7.1 because it is a common and well‐characterized lymphoma in animals, especially dogs. In addition, we know the biologic behavior of this lymphoma is indolent. All the other tumors listed are aggressive lymphomas. Hepatosplenic T‐cell lym­ phoma is also listed separately in Box 7.1 as it is a defined entity in humans and is characterized in dogs, although the total number of cases is low. Hepatocytotropic T‐cell lymphoma is not a recognized disease but it has been reported in dogs and cats and is partially characterized. It is described with hepatosplenic T‐cell lymphoma in this section because they share similar features and it allows comparisons between the two tumors. If a tumor can be classified that diagnosis should be provided; the names PTCL or PTCL‐NOS should be reserved for those T‐cell tumors that cannot be classified, perhaps because the tumor lacked differentiating features, complete antibody profiles or molecular tests were not available, the organ distribution of the tumor was not known, or the owners or oncologists did not want or could not afford complete characterization. Some generic descriptors for these tumors are the following. Most importantly, they are extrathy­ mic T‐cell tumors and most are high grade. Many will look like DLBCL and are only differentiated by phenotyping. Most are in lymph nodes and cause a paracortical expansion; compressed sinus and tumor cells will extend beyond the capsule. A few are extrano­ dal, usually occurring in the skin or subcutis. They may be associ­ ated with vascular proliferation, vascular invasion, and necrosis. The cells can be calssified as pleomorphic small, pleomorphic mixed (small, medium and large cells), or pleomorphic large. One study reported the majority to be large cell2 and another indicated the majority were mixed.3 Eosinophils and macrophages may be noted. The mitotic rate will be variable and can be used to assign high grade or low grade. Nuclei are round to oval and nucleoli quite variable. Cytologically, PTCL look like immature aggressive lymphomas and immunophenotyping is needed to distinguish B‐ versus T‐cell type. Large cell types and those that efface nodes with extranodal tumor will look like DLBCL and are only differentiated by pheno­ typing. The mixed cell types look like TCRLBCL and are also distin­ guished via phenotyping. Inflamed cutaneous non‐epitheliotropic T‐cell lymphoma resembles cutaneous histiocytosis or even inflammation. Combining data from two studies of canine lymphoma produces almost 1600 cases, of which approximately 15% were classified as

Tumors of the Hemolymphatic System    267

PTCL‐NOS and 3–11% were T‐zone lymphoma.2,3 Therefore these two T‐cell classifications accounted for about 20% of the canine lymphomas. Identification of TZL is crucial as its indolent behavior is in contrast to the other lymphomas in this group. The unique his­ topathology of TZL is diagnostic. Flow cytometry with broad panels of antibodies have also been used to help subdivide T‐cell tumors.4–6 By flow cytometry, CD21 positivity, CD25 positivity, CD45 nega­ tivity, and high expression of class II MHC was used to identify TZL and predict long survival times.7,8 CD21 negativity, CD25 positivity, CD45 positivity, and low expression of class II MHC identified other T‐cell lymphomas that have short survival times of approxi­ mately 160 days.9 Only 15 cases had histopathology, 10 of which were PTCL‐NOS and 5 LBL. The following sections include specific diagnoses that are in the broad group of PTCL.

A

Mycosis fungoides and Sézary syndrome

Defining the neoplasm

Cutaneous lymphoma can be B- or T-cell, epitheliotropic, or nonepitheliotropic. The most common is T-cell epitheliotropic; that is, mycosis fungoides. Mycosis fungoides affects the skin and/or mucous membranes of humans and many species of animals, including rodents.1–3,7–15 The diagnosis requires histology and the disease is characterized by linear infiltrations of the epidermis by small to intermediate‐size lymphocytes that typically have sharp shallow nuclear indentations. In the fully developed lesion there are clefts in the epidermis known as Pautrier’s microabscess that are filled with neoplastic cells (Figures  7.42–7.45). The disease has stages beginning with an interface dermatitis that progresses to a patch or plaque. In the untreated case the skin lesion will enlarge, and if tumor cells appear in the blood the condition is called Sézary

B

Figure  7.42  Mycosis fungoides, oral cavity, dog. (A) The epithelium is thickened and markedly infiltrated by lymphoid cells that extend through the

basement membrane and are in the submucosa. There are small cystic areas in the epithelium filled with edema and lymphocytes (Pautrier’s microabscess). (B) CD3: Neoplastic cells in the epithelium are strongly positive, indicating T‐cell origin and those in submucosa are moderately positive. See Figure 7.43.

A

B

Figure 7.43  Mycosis fungoides, oral cavity, dog. (A) Neoplastic lymphocytes have considerable cytoplasm, nuclei are 1.5 times the size of RBC, chromatin is dense and several nuclei are folded or convoluted. (B) Cytology: Tumor cells exfoliated in large numbers. The neoplastic cells are 2–3 times as large as the dense small non‐neoplastic lymphoid cells. Cytoplasm of neoplastic cells is abundant and deeply basophilic; mitoses (arrow) were present. Most nuclei are round or oval, some have indented nuclear profiles, and a few horseshoe‐shaped nuclei are present.

268    Tumors in Domestic Animals

A

B

Figure 7.44  Mycosis fungoides, skin, cat. (A) A 12‐year‐old domestic short hair cat presented with hair loss and swollen skin over the tail. The skin is inter­

mittently thick and thin with large cystic areas in the epidermis filled with edema and lymphocytes. Neoplastic lymphoid cells extended into the deeper dermis and infiltrated adnexa. (B) CD3: The infiltrating lymphoid cells are strongly and uniformly immunostained, including those in the epithelial cysts.

A

B

Figure 7.45  Pagetoid reticulosis, skin, dog. (A) A 6‐year‐old Schnauzer dog presented for foci of hair loss and depigmentation that were unresponsive to

antibiotics. A biopsy of affected skin has a greatly thickened epidermis with a marked infiltration of lymphocytes in the epidermis and in the hair follicles (arrows). Some regions of the epidermis are devoid of melanocytes (left) and in other areas they are still present (right); there also are melanophages in the dermis. (B) CD3: The infiltrating cells within the thickened epidermis are uniformly and heavily marked. The contrast in colors makes the intraepidermal pattern easier to see. Note intraepithelial infiltration in one of the hair follicles (arrow). Some of the cells stained “brown‐black” in the dermis are melano­ phages (see A).

syndrome (Figure 7.46). In the early stages of the infiltration or at the edges of an established area of tumor the neoplastic cells form a single row just inside of the basement membrane known as the “string of pearls.” As the disease progresses the cellular infiltration separates dermal collagen fibers and infiltrates the mural aspect of the hair follicles and glandular adnexa. In dogs the infiltration may completely fill the apocrine glands. Tumor extends to the level of the deeper dermal vessels.

Epidemiology and occurrence

This is not a common lymphoma in animals but is well described in dermatology and pathology texts. Depending on the type of acces­ sions, the prevalence of the disease can be exaggerated.2 There were 13 cases of mycosis fungoides in a collection of 502 canine

lymphomas11 and 53 (9%) in another series of 600.2 In cats, 15 cases of mycosis fungoides were present in a collection of 751 lym­ phomas.11 Only one feline case involved the gingiva, all the others were in haired skin. The mean age was 11 years and there was no gender predominance. In cattle, mycosis fungoides typically occurs at 2 years of age and is not associated with infection with BLV.11 The lesions in cattle are usually along the back and sides. They depilate and ulcerate and then heal spontaneously before appearing else­ where. In the early stages of the disease cattle remain in apparent good health, but over a period of about a year the disease progresses to become systemic. In the advanced cases the tumor fills the abdomen and virtually all of the organs may be involved with tumor. When this severe, mycosis fungoides looks like the enzootic type of bovine lymphoma.

Tumors of the Hemolymphatic System    269

A

B

C

D

Figure 7.46  Sézary syndrome, horse, buffy coat. (A) A thoroughbred 10 years old presented in late pregnancy with ventral edema and muffled heart sounds.

The animal had brick red oral mucosa and atypical cells in circulation. A buffy coat preparation of blood was prepared, which had a high population of intermediate and large cells and occasional “blast‐type” cells (center). The white blood cell count in this horse ranged between 11,000 and 15,000/μL. The atypical cells were originally classified as monocytic but after review were identified as lymphocytic. (B) The cell in the center is typical of the large cells seen in the marrow, blood, and pericardial effusion. The nucleus has a fine chromatin pattern and two very large nucleoli. (C) Bone marrow and lymph nodes contained mononuclear cells similar to those found in the blood. The reniform nuclei are monocyte‐like but the tumor cells will be CD3 positive. (D) There is a dense infiltration of neoplastic lymphoid cells in the oral submucosa. The mucous membranes were markedly hyperemic, resulting in fiery red oral mucosa which is a clinical feature of this disease. Inset: Neoplastic cells are dissecting through the myocardium and pushing muscle fibers apart in a pattern typical of neoplasia. Neoplastic cells were present in the pericardial effusion.

Mycosis fungoides is rare in horses but also progresses to involve viscera.8,15 There may be a prolonged period before the cells appear in the blood of horses (Figure 7.46A–E).11 A human neoplasm called lymphomatoid papulosis is a chronic infiltration of the skin with clonal T cells that has several forms or stages that resemble mycosis fungoides.1,10,11 It is not reported in animals.

Clinical presentation

Dogs present with thickened gingiva or skin.11–14 In the dog the pre­ senting lesions may be singular within the mouth or lip, or may appear in multiple sites, usually on the head or ventral area of the body. The mean age was 10 years and there appeared to be a male predominance.8 In the cow, the lesion is very distinctive and should initiate biopsy to confirm the diagnosis and salvage the animal before there is internal progression. The disease in horses may appear first in the mouth as plaque‐like areas that appear reddened but do not blanche with pressure. There may be mild peripheral

lymphadenopathy that is not apparent unless looked for. Similarly, atypical cells may not be noted in peripheral blood but can be found if they are searched for.11

Pathology

The lesions of mycosis fungoides are primarily in the skin and/or mucous membranes. The lesions progress from an interface derma­ titis into plaques that become confluent and form distinct gross lesions. There is epidermal or epithelial hyperplasia and intraepi­ dermal aggregates of neoplastic T lymphocytes (Pautrier’s microab­ scess) (Figures 7.42–44). A variant is pagetoid reticulosis, in which the neoplastic cells are confined above the basement membrane of the epidermis, epithelium, or adnexa (Figure 7.45). The intraepidermal tumor cells react positively to CD3 and the lymphocytic inflammation subjacent to the basement membrane will have scattered T cells but is predominantly filled with B lym­ phocytes and other inflammatory cells. The skin is markedly thick­ ened by the intraepidermal neoplasm and the dermal inflammation.

270    Tumors in Domestic Animals

In the dog, mycosis fungoides cells in the skin are CD8 positive, unlike mycosis fungoides lesions in humans where the skin infil­ trates are CD4 type. An epitheliotropic lymphoma in a cat had CD3‐positive neoplastic cells that expressed perforin and there were LGL in the blood.7 The neoplasm can be widespread in multiple tissues (Figure 7.46). In cattle, the lesions are raised, circular tumors or thickened areas of skin that form plaques 5–50 cm in diameter. The affected areas are hairless and may be ulcerated and bleeding. In the cow the lesions have been misdiagnosed clinically as ringworm (Trichophyton). Cytologically the infiltrating cells are small to intermediate with nuclei 90% lymphoma cells in the abdominal fluid. These are usually large cells, with fairly abundant basophilic cytoplasm and immature large nuclei with prominent nucleoli. These events occur in cattle and horses as well as pets. If the lymphoma starts in the small intestine then involvement of nodes, liver, and spleen may not be severe. If the lymphoma is multicentric and an immature cell type then liver and spleen will be enlarged and neoplastic. Some of the peripheral nodes as well as mesenteric nodes will be neoplastic. Enteric lymphoma in cats This is a common disease in cats.1–4 It is well described and elo­ quently illustrated in a series of 120 cats with over 125 tumors (some tumors were multifocal).2 The tissue samples from this latter study came from surgical biopsies (n = 47), endoscopic biopsies (n = 35) or autopsy (n = 38). Lesions ranged morphologically from mild to marked and they were categorized by anatomic distribution in the gastrointestinal tract by histologic patterns of transmural versus mucosal (lamina propria and intraepithelial at the surface of intestines and in crypts), cell size (small, large), and by phenotype. Approximately 80% were T‐cell and 15% B‐cell.2 Another study reported that approximately 50% of feline enteric lymphomas were B‐cell.4 Both studies agreed that T‐cell lymphomas were more common than B‐cell in the small intestine and that some tumors were not immunoreactive with B‐ or T‐cell antibodies.2,4 The majority of T‐cell lymphomas were mucosal (80%) and 95% of B‑cell were transmural.2 The jejunum was the most common site for T‐cell lymphomas. B‐cell lymphomas (n = 19) were distributed fairly evenly from stomach to colon but none were found in the duodenum.2 All the B‐cell tumors were large cell type2 which is sim­ ilar to another report.4 The T‐cell tumors were subdivided: if they were mucosal then 80 of 84 were small to intermediate size and if transmural 11 of 19 were large cell type. Nine of these 11 large cell type were LGL and all 9 reacted positively for granzyme B.2 Only two of 84 small cell type were LGLs. All B‐cell lymphomas in cats were large cell‐type and they were further divided by nuclear morphology into centroblastic (17/19) or immunoblastic (2/19).2 Immunoblastic types have a single large central nucleolus, whereas centroblastic types have multiple nucleoli, often adjacent to the nuclear membrane. In another study all gastric lymphomas in cats were of B‐cell type and large cell immunoblastic was most common.4 B‐cell lymphomas in the small or large intestine were situated close to lymphoid follicles, but this pattern was not

apparent in the stomach.2 Eighteen of 19 were transmural and the tissues were markedly infiltrated, which made the diagnosis straight­ forward but the association with lymphoid follicles difficult to deter­ mine. Intraepithelial location was seen in one case. As the tumors were divided into groups by phenotype, size of cells, and whether mucosal or transmural, the number of cats in each group was reduced, and this was accentuated by the small number of cats that had follow‐up data. Cats with follow‐up data that had mucosal T‐cell lymphoma (n = 54) had an MST of 29 months and cats with transmural T‐cell lymphoma (n = 13) had an MST of 6 weeks.2 Dividing groups by size of cells revealed similar survival predictions. Cats with small cell T‐cell lymphoma (n = 54) had an MST of 28 months versus 6 weeks for large T‐cell lymphoma (n = 13).2 Six of 19 cats with follow‐up data that had B‐cell lym­ phoma survived about 3.5 months. Tumors and survival data were not correlated with treatment regimens. Although the numbers of cats in the different groups were small and treatments could not be factored in, trends were apparent. Cats with small T‐cell lymphoma in the mucosa have an indolent course (>2‐year survival), whereas cats with large T‐cell lymphoma, B‐cell lymphoma, transmural lymphoma, or LGL cell lymphoma have much shorter survival times. The worst prognosis may be feline LGL lymphomas that have concurrent lymphocytosis (leukemia), and transmural pattern since they had an MST of 19 days.6 T‐cell lymphomas (mucosal and transmural) were almost exclu­ sively localized to the small intestine, especially the jejunum, and were uncommon in the duodenum or stomach.2 Practical consider­ ations of this anatomic distribution is that endoscopic biopsies taken from the duodenum would miss the most common location of T‐cell lymphomas in cats (jejunum) and B‐cell tumors are rare in the duodenum. There were only 7 instances (approximately 7%) of T‐cell lymphoma involving the stomach or large intestine. Lesions in the lamina propria of cats with enteric lymphoma will range from mild infiltration to severe effacement.2–4 Histologically, mild lesions will have shortened or fused villi with variable infiltra­ tion in the lamina propria. Villi should be 2–4 times the height of crypts but it is preferential to examine villi in which the superficial epithelial cells align with subjacent crypts. Pathologists should search sections until this alignment can be assessed as these regions are perpendicular sections and are not cut in a tangential plane that will produce blunting or fusion. Avoid villi over Peyer’s patches as they are normally short and interspaced with domes of the lym­ phoid cells from the follicles beneath. The villi are increased in diameter to 2–3 times normal size due to neoplastic infiltration in the lamina propria. The neoplastic cells may fill the lamina propria or form aggregates (patches) at various levels of the lamina propria (Figure  7.50). Adjacent villi may be unaffected or markedly affected. This characteristic of “skipping” villi is a feature of lym­ phoma and if seen is an aid to differentiate lymphoma and IBD. The mucosal infiltration in IBD does not skip adjacent villi. Neoplastic cells are always in a greater concentration in the lamina propria than in an intraepithelial location. An epitheliotropic pattern is more common in surface epithelium than crypt and it is present in approximately 50–60% of mucosal or transmural T‐cell lymphomas.2,4 Intraepithelial lymphocytes can be singular or clus­ tered. If they occur in aggregates of 4–6 cells then clonal prolifera­ tion is likely. The colonization of the epithelium by neoplastic lymphocytes may be so great that there are more nuclei of lympho­ cytes than those of the epithelial cells. The more intraepithelial lymphocytes present in a lesion favors lymphoma. Immunostaining with CD3 will enhance the intraepithelial pattern and is useful

Tumors of the Hemolymphatic System    277

A

B

C

D

Figure  7.50  Lymphoma, duodenum, dog. (A) Villi are about twice normal width and there is lymphoid infiltration. However, there is distortion of the

sample and sections taken over a Peyer’s patch usually have shorter villi. At this magnification inflammatory bowel disease (IBD) is a likely differential. This sample only contained mucosa, therefore infiltration into or past muscularis mucosa could not be evaluated to help differentiate IBD and lymphoma. (B) Higher magnification: The surface epithelium is thin and there are numerous lymphocytes in intraepithelial areas and in lamina propria; however, there are also neutrophils and eosinophils, which favor IBD. In cases like this, PCR for clonality and IHC may help differentiate IBD and lymphoma. Of equal value is to examine as many other sections as possible to determine the uniformity or heterogeneity of the infiltration. (C) CD3: Lymphocytes in the villus tips, in intraepithelial areas, and in the solid area of lymphoid proliferations in the deeper mucosa are all T‐cell type, supporting a clonal proliferation. Inset: Higher magnification of villus tips. Clusters of intraepithelial lymphocytes as in this image also support lymphoma. (D) CD79a: Relatively few cells are positive, in contrast to the numerous T cells seen in (C). The absence of a mixture of B and T cells in the mucosa favors the diagnosis of lymphoma. When the cellular infiltration extends transmurally, as in Figure 7.49D then this strongly favors lymphoma and PCR or IHC should not be needed. When intracy­ toplasmic granules are seen, the tumor is T‐cell type, but B‐ versus T‐cell requires phenotyping in lymphomas without cytoplasmic granules.

when there does not appear to be many lymphocytes in the epithe­ lium.2 There is often a clear space around or close to the nuclei of the intraepithelial lymphocytes. Search these cells for eosinophilic granules which indicates they are LGL. Of 84 mucosal T‐cell lymphomas in one study,2 80 were small to intermediate with nuclei about 1.5 RBC in diameter with a nuclear chromatin‐dense, mature appearance. Small intraepithelial lym­ phocytes were the most common T‐cell lymphomas in an earlier study.4 Mitoses may be present but are not frequently encountered, even where there is heavy lymphoid proliferation in the deeper mucosa and lymphoma is apparent.1,15 The crypts may be separated from the muscularis mucosa by a laminar band of small lympho­ cytes. A few eosinophils may be present in lymphomas but eosino­ phils will not be predominant. These changes may be accompanied

by a secondary and nonspecific moderate or mild dilation of the villus lymphatics. When the diagnosis is small cell type, mucosal intraepitheliotropic lymphoma there is little to be gained by IHC for B‐ versus T‐cell since >95% will be T‐cell and IHC only iden­ tifies phenotype, it cannot identify neoplastic cells. However, visu­ alization of the intraepithelial distribution is enhanced by staining with CD3 and the homogeneity of the proliferation is apparent. Therefore IHC is helpful to recognize these characteristic features. Mucosal T‐cell lymphoma of small cell type is analogous to EATCL type II. Transmural types have the most severe lesions, with effacement of the lamina propria, moderate to marked infiltration of the sub­ mucosa, and tumor cells dissecting to the serosa or beyond. The diagnosis of lymphoma is easy in these cases and the majority will

278    Tumors in Domestic Animals

be large cell type. However, they can be either T‐cell or B‐cell phe­ notype and this is determined by IHC. The B‐cell types are large cells and are comparable to DLBCL. They have a median survival of approximately 3.5 months. Eleven of 19 transmural T‐cell lym­ phomas were large cell type and 9 of these were LGL.2 Large T‐cell types have nuclei greater than 2 RBC in diameter and nuclei appear more immature and pleomorphic; they may also have variable con­ tours. Cytoplasmic granularity should be assessed, which is easiest to do in cytologic preparations using methanolic‐based stains rather than aqueous‐based ones16 (see Figures  7.39 and 7.54). If cytology slides are not available then consider oil‐immersion objec­ tives and look for eosinophilic granules close to the nuclei or in sections stained with PTAH look for dark purple granules. LGL will be granzyme B positive and, if available, this is an easy method to identify LGL in histopathology preparations. Granzyme B labeling is cytoplasmic, usually juxtanuclear. See the section in this chapter on LGL tumors. Transmural T‐cell lymphoma of large cell type is analogous to EATCL type I, and many of these will be LGL. Clonality for T‐cell receptor gene (TCRG) rearrangement will support the diagnosis in about 90% of mucosal and transmural T‑cell lymphomas.2 Approximately 80% will be monoclonal and 10% oligoclonal. The other approximate 10% are polyclonal and clonality was undetected, which can happen if there are gene segments used in the rearrangement that are not detected by the primers used in the assay. Clonality was detected in 50% of the B‐cell lymphomas and pseudoclonality in about 40%; 10% were polyclonal. The authors discussed possible explanations.2 No test is 100% sensitive and specific; however, if morphology and IHC are not definitive than clonality testing may help these ambiguous cases.

Small intestine and differential diagnoses

Dogs, cats, and horses have IBD that mimic lymphoma clinically, grossly, and microscopically (Figures  7.49 and 7.50).11–15 In many cases the differentiation is clear from histology or gross examination, but in some the differentiation of these two diseases is difficult.11 Problematic cases are endoscopic biopsies that have mucosa only (depth of invasion cannot be assessed) or if there are limited number of samples submitted or the quality of the sections is not good. The distinction of these two diseases is often difficult and, although this is not proven, it is thought that lymphoma may arise in areas of IBD and that the two lesions may be present in the same patient. Difficult cases require additional parameters and integration of results from lymph node biopsy, possibly including samples of liver, cytology, PARR for clonality, IHC, flow cytometry, FeLV, and BLV. Clonality can be help­ ful but there are false‐negative and false‐positive results just like any test. IHC cannot identify neoplastic cells but it can help identify homogeneous versus heterogeneous populations and an intraepithe­ lial pattern. We tend to emphasize or remember those cases that are exceptions or that were difficult to diagnose. The primary histologic lesion in both diseases is in the mucosa of the small intestine, and lymphocytes are the most numerous cell type. Both diseases will have enlarged mesenteric lymph nodes. Endoscopic mucosal biopsies are taken from the duodenum, which is not the most common location of enteric lymphoma in cats and is a very uncommon location for B‐cell lymphoma. The following are features to help differentiate IBD from lym­ phoma. In almost all cases of IBD the inflammation is confined to the mucosa, whereas lesions that enter the submucosa or extend further favor lymphoma. In general, all or the great majority of each villus is affected in IBD, but in lymphoma some villi may be skipped. If one villus has numerous cells in the lamina propria and the next

villus is unaffected, this favors lymphoma. The inflammatory cell population is mixed and it is not monomorphic in IBD. If the inflam­ mation is granulomatous then the distinction is easy and the task is to determine the cause (e.g., granulomatous enteritis of horses, M. avium, ulcerative colitis, etc.). The more mixed the cellular infiltra­ tion and the greater the plasma cell population the more likely it is IBD. IBD has areas in which the inflammation is intensified, if sec­ tions are searched carefully. Intensification may appear as foci of plasma cells, pockets of eosinophils or neutrophils, and the normal mucosal structures are destroyed rather than pushed aside. Neoplasia should be uniform, monomorphic, it should infiltrate and push normal structures apart but it does not cause focal inten­ sification with inflammatory cells or discrete areas of necrosis sur­ rounded by inflammatory cells. Neoplastic lymphoid cells will dissect between muscle fibers rather than produce areas of necrosis or myositis (see Figure 11.24C,D). Examine the superficial epithe­ lium for the type of cells and clustering of cells. The presence of more neutrophils here favors IBD, whereas more lymphocytes favors lymphoma, especially if they form intraepithelial aggregates of 4–8 cells. LGLs with large cytoplasmic granules that are scattered in the epithelium favors IBD. Numerous LGLs in the lamina propria and intraepithelial favors lymphoma. The greater the expansion of the lamina propria by lymphocytes, the more monomorphic this population and the more likely it is lymphoma. Mucosal lymphoma is accompanied by a loss or absence of plasma cells. If lesions are present in liver, spleen, and kidney then lymphoma is the working diagnosis. If white tumors are present in organs then lymphoma, another tumor, or one of the granulomatous diseases is present, not IBD. Peripheral lymphadenopathy or hypercalcemia are features of lymphoma. TCRLBCL is an uncommon form of enteric lymphoma but its heterogeneous cell population can make distinction of inflamma­ tion versus neoplasia difficult.4 Polyclonal patterns in PARR and heterogeneous lymphoid populations in IHC are characteristic of IBD and monoclonal or oligoclonal patterns with homogeneous lymphoid cells are features of lymphoma.

References

1. Cesari, A., Bettini, G., and Vezzali, E. (2009) Feline intestinal T‐cell lymphoma: assessment of morphologic and kinetic features in 30 cases. J Vet Diagn Invest 21:277–279. 2. Moore, P.F., Rodriguez‐Bertos, A., and Kass, P.H. (2012) Feline gastrointestinal lymphoma: mucosal architecture, immunopheotype, and molecular clonality. Vet Pathol 49:658–668. 3. Mahony, O.M., Moore, A.S., Cotter, S.M., et al. (1995) Alimentary lymphoma in cats: 28 cases (1988–1993). J Am Vet Med Assoc 207:1593–1598. 4. Pohlman, L.M., Higginbotham, M.L., Welles, E.G., et al. (2009) Immunophenotypic and histological classification of 50 cases of feline gastrointestinal lymphoma. Vet Pathol 46:259–268. 5. Coyle, K.A. and Steinberg, H. (2004) Characterization of lymphocytes in canine gastrointestinal lymphoma. Vet Pathol 41:141–146. 6. Roccabianca, P., French, R.A., Seitz, S.E., and Valli, V.E. (1996) Primary epithelio­ tropic alimentary T‐cell lymphoma with hepatic involvement in a dog. Vet Pathol 33:349–352. 7. Woo, J.C., Roccabianca, P., van Stijn, A., and Moore, P.F. (2002) Characterization of a feline homologue of the αE integrin subunit (CD103) reveals high specificity for intra‐epithelial lymphocytes. Vet Immunol Immunopathol 85:9–22. 8. Bacon, C.M., Du, M.‐Q., and Dogan, A. (2007) Mucosa‐associated lymphoid tissue (MALT) lymphoma: a practical guide for pathologists. J Clin Pathol 60:361–372. 9. Burton, A.J., Nydam, D.V., Long, E.D., and Divers, T.J. (2010) Signalment and clinical complaints initiating hospital admission, methods of diagnosis, and path­ ological findings associated with bovine lymphosarcoma (112 cases). J Vet Intern Med 24:960–964. 10. Dukes, T.W., Bundza, A., and Corner, A.H. (1982) Bovine neoplasms encountered in Canadian slaughterhouses: a summary. Can Vet J 23:28–30.

Tumors of the Hemolymphatic System    279

11. Carrasco, V., Rodriquez‐Bertos, A., Wise, A.G., et  al. (2015) Distinguishing intestinal lymphoma from inflammatory bowel disease in canine duodenal endo­ scopic biopsy samples. Vet Pathol 52:668–675. 12. Steinberg, H., Dubielzig, R.R., Thompson, J., and Dzata, G. (1995) Primary gastro­ intestinal lymphosarcoma with epitheliotropism in three shar‐pei and one boxer dog. Vet Pathol 32:423–426. 13. La Perle, K.M., Piercy, R.J., Long, J.F., and Blomme, E.A.G. (1998) Multisystemic, eosinophilic, epitheliotropic disease with intestinal lymphosarcoma in a horse. Vet Pathol 35:144–146. 14. Pinkerton, M.E., Bailey, K.L., Thomas, K.K., et al. (2001) Primary epitheliotropic intestinal T‐cell lymphoma in a horse. J Vet Diagn Invest 202:150–152. 15. Valli, V.E. (2007) Enteric T‐cell lymphoma. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 318–327. 16. Allison, R.W. and Velguth, K.E. (2010) Appearance of granulated cells in blood films stained by automated aqueous versus methanolic Romanowsky methods. Vet Clin Pathol 39:99–104.

Hepatosplenic and hepatocytotropic lymphoma (γδ T‐cell lymphoma) These are two different lymphomas that share certain morphologic and phenotypic characteristics. Both tumors are rare; they are gamma/delta (γδ) cytotoxic T‐cell lymphomas that progress rapidly and dogs die or are euthanized because of the neoplasm in a matter of days or weeks after diagnosis. Hepatosplenic lymphoma (HS‐TCL) is a recognized classification and has been reported in dogs,1–3 cats,4,5 and a horse.6 HS‐TCL is a rare lymphoma in humans associated with systemic signs of illness, anemia, thrombocytopenia, and a poor prog­ nosis.7 There is infiltration of the spleen, liver, and bone marrow without lymphadenopathy and the spleen is suspected to be the origin of the lymphoma.1,7 Neoplastic cells in circulation are not a feature of HS‐TCL or hepatocytotropic T‐cell lymphoma (HC‐ TCL). The cells in human and canine HS‐TCL have a γδ type T‑cell receptor.1,7 Only a few cases of HS‐TCL in dogs have been reported and most were from retrospective material.1,3 Patterns were noted and they are summarized below but the original publications should be read for the details they provide. Dogs with HS‐TCL have T‐cell lym­ phoma in the liver and spleen. Some cases have tumor in mesen­ teric lymph nodes and other abdominal organs but not peripheral lymph nodes; however, autopsy data is incomplete therefore it is difficult to know how extensive these lymphomas are. The diagnosis of HS‐TCL is made from histology of the liver which contains a lymphoma that is primarily located sinusoidal rather than periportal or perivascular. The tumor cells form a  linear pattern along hepatic cords. They may compress and cause  atrophy of hepatocytes. There will be neoplastic cells in portal tracts and around central veins, but this pattern is less pro­ nounced than the sinusoidal distribution (Figure  7.51, see also Figure  14.19A,B). The infiltration is usually marked and easily seen at lower magnifications. Involvement of bone marrow is patchy, with some areas heavily infiltrated and other areas so mildly involved it may require IHC with CD3 to be certain the tumor is present (Figure  7.52). Mesenteric nodes are irregularly infiltrated. The colonization of nodes spares the outer cortex and germinal centers and involves the inner paracortex and medullary areas. Blood vessels in the lungs will contain tumor cells and this lesion ranges from mild to marked.1 Cytologically or histologically, a key to the diagnosis is erythro­ phagocytosis by neoplastic lymphoid cells and non‐neoplastic his­ tiocytes. Erythrophagocytosis is more prominent in macrophages and it must be searched for in the neoplastic lymphocytes. The spleen contains the tumor in sinudoidal areas and red pulp. The

white pulp may be atrophic. Erythrophagocytosis by macrophages is prominent in the spleen. The neoplastic lymphocytes are intermediate to large, nuclei are 1.5–3 RBC in diameter.1 Cytoplasm ranges from a thin rim to abundant and sometimes is clear. Mitotic figures range from 0 to 2/10 HPF, which is fairly low for aggressive lymphomas. Six out of 7 dogs with HS‐TCL were dead within 1 month of their diag­ nosis.1 The immunophenotypic profile was CD3 positive, CD11d positive, γδ type T‐cell antigen receptor, and granzyme B positive.1 Taken together, these data support splenic cytotoxic T cells as a likely origin of this lymphoma. None of the tumor cells were identified as LGL, although CD11d is expressed on LGL (as well as macrophages and T lymphocytes in the splenic red pulp), the cytoplasm was clear, and LGL can be of γδ T lymphocytes. Clinical parameters indicated that dogs with HS‐TCL had a regenerative anemia, thrombocytopenia, hypoalbuminemia, var­ iable hepatic enzyme increases, and an absence of leukemia and bilirubinemia.1 Dogs with HC‐TCL had sufficiently different clinicopathologic results, histologic lesions, and immunophenotype that they were considered a separate entity and not a subtype of HS‐TCL.1 There were only two dogs found with this tumor so data were limited. The principal lesions in each dog were centered on the liver and con­ sisted of neoplastic lymphocytes that were not just sinusoidal but followed and invaded hepatic cords and which appeared to be intra­ hepatocytic (see Figures  7.53 and 14.20). This appearance was ­distinctive. Transmission electron microscopy revealed the lym­ phocytes were not truly in the cytoplasm of hepatocytes and the appearance was due to an invagination of the hepatocyte cell mem­ brane as lymphocytes encroached.1 Neoplastic lymphoid cells looked the same as in HS‐TCL. The phenotypic profile was the same, with the exception that the lym­ phocytes in HC‐TCL were negative for CD11d.1 Other differences were the absence of erythrophagocytosis, no anemia but pro­ nounced icterus, hyperbilirubinemia (>20 mg/dL), and massive increases of serum ALP (>20,000 IU/L) and gamma glutamyltrans­ ferase.1 Both dogs died within days of diagnosis. The cell of origin for HC‐TCL is believed to be cytotoxic γδ T cells but perhaps the liver is the source of these cells rather than the more expected splenic source. Sinusoidal distribution of lymphoma cells in the liver is a diagnostic aid for HS‐TCL and the unique pattern follow­ ing hepatic cords for HC‐TCL. Sinusoidal distribution of tumor cells in the liver is a characteristic of myeloid neoplasms or leu­ kemic forms of lymphoma. Clinical features and response to treatments of hepatic lym­ phoma were described in 18 dogs.2 These dogs had heterogeneous types of lymphomas that included B‐cell, multicentric, and medias­ tinal.2 Presumably a few were HS‐ or HC‐TCL but insufficient char­ acterizations were performed. The MST in this group of dogs was approximately 2 months with a range of 2–402 days.2 The histologic signature of HC‐TCL has also been observed in the liver of cats with T‐cell lymphoma.4,5 The authors chose the term emperipolesis‐like to emphasize the distinctive histologic pattern of lymphoid cells that appeared to be in hepatocytes. The light microscopic appearance of an intracellular location was due to invaginations in the cell membrane of hepatocytes, and the tumor cells were not actually in the cytoplasm of hepatocytes.5 Four of 12 cats were of LGL cell type (Figure 7.54) and five had monoclonal proliferations of neoplastic cells as determined by PCR for the TCRG gene.5 Survival and prognostic data were not reported but they are probably similar to those in humans and dogs.

280    Tumors in Domestic Animals

A

B

C

D

Figure 7.51  Hepatosplenic T‐cell lymphoma (HS‐TCL), dog. (A) The pattern of infiltration in this liver is periportal and sinusoidal. Although most lym­

phomas in the liver are primarily periportal or perivascular, hepatosplenic types of lymphomas are sinusoidal. Myeloproliferative diseases are also primarily sinusoidal. These patterns help distinguish the neoplastic processes but are not definitive. (B) Portal area, bile duct (arrow): The sinusoids are dilated by neoplastic lymphocytes that are intermediate to large cell size. (C) Spleen: There is cuffing around arterioles by the infiltrating lymphocytes. Inset: Lymphoma cells are much larger than the red blood cells. (D) CD3, spleen: About half of the infiltrating tumor cells stain strongly and the rest appear unlabeled. Hepatosplenic lymphomas are of gamma/delta type and therefore not all neoplastic cells will stain with CD3. Inset CD79a: The tumor cells are negative. A few residual lymphocytes near an arteriole are non‐neoplastic.

References

1. Keller, S.M., Vernau, W., Hodges,J., et al. (2012) Hepatosplenic and hepatocyto­ tropic T‐cell lymphoma: two distinct types of T‐cell lymphoma in dogs. Vet Pathol 50:281–290. 2. Dank, G., Rassnick, K.M., Kristal, O., et al. (2011) Clinical characteristics, treatment and outcome of dogs with presumed primary hepatic lymphoma: 18 cases (1992– 2008). J Am Vet Med Assoc 239:966–971. 3. Fry, M.M., Vernau, W., Pesavento, P.A., et al. (2003) Hepatosplenic lymphoma in a dog. Vet Pathol 40:556–562. 4. Ossent, P., Stockli, R.M., and Posposchil, A. (1989) Emperipolesis of lymphoid neoplastic cells in feline hepatocytes. Vet Pathol 26:279–280. 5. Suzuki, M., Kanae, Y., Kagawa, Y., et  al. (2011) Emperipolesis‐like invasion of neoplastic lymphocytes into hepatocytes in feline T‐cell lymphoma. J Comp Pathol 144:312–316. 6. Roccabianca, P., Paltrinieri, S., Gallo, E., and Giuliani, A. (2002) Hepatosplenic T‑cell lymphoma in a mare. Vet Pathol 39:508–511. 7. Belhadj, K., Reyes, F., Farcet, J.P., et  al. (2003) Hepatosplenic gammadelta T‐cell lymphoma is a rare clinicopathologic entity with poor outcome: report on a series of 21 patients. Blood 102: 4261–4269.

Intravascular large T‐cell lymphoma, subcutaneous panniculitis‐like T‐cell lymphoma, angioimmunoblastic T‐cell lymphoma, aggressive NK‐cell leukemia/lymphoma

Intravascular large T‐cell lymphoma In this rare disease, neoplastic lymphocytes proliferate inside blood vessels rather than in hemolymphatic organs.1–7 Despite intravas­ cular proliferation and numerous neoplastic cells in blood vessels on histopathology, leukemia is not present, at least not when assessed with routine means of detection. In approximately 5% of human cases neoplastic cells can be detected in circulation. One canine case documented a LGL intravascular large T‐cell lym­ phoma (IVL) that had neoplastic cells in circulation but the number of tumor cells were low (approximately 500/μL, 8%).6 The authors concluded that it was not a histiocytic tumor, there was erythropha­ gocytosis by tumor cells, and it was CD18/CD45 positive but

Tumors of the Hemolymphatic System    281

A

B

C

D

Figure 7.52  Hepatosplenic T‐cell lymphoma (HS‐TCL), dog. (A) Spleen, cytological preparation: Erythrophagocytosis by neoplastic cells is prominent

(arrows); mitotic figure is present; note erythropoiesis with numerous metarubricytes. Tumor cells and macrophages are erythrophagocytic in HS‐TCL. (B) CD3, cytological preparation: The neoplastic cells in the spleen are strongly marked and the mature red blood cells and developing erythroid cells are not. (C) Liver: Neoplastic lymphoid cells fill and dilate the hepatic sinusoids. This distribution, along with tumor cells that form a linear pattern along hepatic cords, are characteristic of HS‐TCL. Many of them are of LGL origin. (D) CD3: Lymphoma cells in sinusoids are strongly marked. Inset: Bone marrow stained with CD11d illustrates positive staining of the tumor. CD11d is an integrin expressed by cells in the splenic red pulp, predominantly ­histiocytes and granular lymphocytes that traffic to the red pulp. The latter gives rise to some types of hepatosplenic lymphoma and splenic T‐CLL. CD11d is considered a marker of splenic red pulp origin. The combination of neoplastic lymphocytes in sinusoids of liver, lymphoma in the spleen, erythropha­ gocytosis by tumor cells, and positive staining pattern of tumor with CD3 and CD11d provides the diagnosis of HS‐TCL. CD11d will also mark granular lymphocyte T‐CLL and hemophagocytic histiocytic sarcomas (of splenic red pulp macrophage origin), but the more common histiocytic sarcomas of dendritic cell origin are usually CD11d negative. (Images courtesy of W. Vernau, UC Davis.)

E‐cadherin negative. Therefore it could be histiocytic. The cells cer­ tainly look histiocytic in published images. IVL occurs in humans and is very rare in animals but is reported in dogs,2 cats,3 and a horse.7 The majority in humans are B‐cell (90%) but in dogs T‐cell is more common. B‐cell is reported in dogs, as is non‐T non‐B.2 The most common clinical signs in dogs are related to the brain or spinal cord, which are among the most common tissues to contain IVL.2 Veins and arteries from multiple organs have been found to contain IVL and the numbers of neo­ plastic cells per vessel vary widely from only a few to numerous. Some vessels are packed with neoplastic cells which distend and/or occlude the vessel (Figures 7.55, 7.56, and 19.38) and adjacent tis­ sues are infarcted. There are a variety of secondary lesions in the walls of blood vessels. Vessels in any tissue may be affected, and

there are examples in which vessels of the nasal turbinates were involved and the tumor bridged the cribriform plate and invaded the meninges of the forebrain (Figure 7.56). Diagnosis of IVL in animals is made on histopathology based on seeing neoplastic lymphocytes in the lumen of blood vessels but without neoplasia in nodes or other tissues. The endothelium is still visible and muscular walls are unaffected. Intravascular neo­ plasia needs to be differentiated from tumor cells that form a cuff around a vessel and tumors that invade the wall of a vessel. The latter is seen in some histiocytic sarcomas. The lumen of the vessel is devoid of tumor cells but there are intramural tumor cells, and these are not IVL.3 Vaccine reactions can create a pattern that looks similar to IVL.3 Occasionally, a lymphoma may form a cuff or mantle between the vessel wall and the pericyte sheath that enlarges

A

B

Figure  7.53  Large granular lymphocyte‐type lymphoma, liver, cat. (A) The tumor cells appear to be within the cytoplasm of hepatocytes providing an

emperipolesis‐like pattern characteristic of hepatocytotropic T‐cell (HC‐TCL) lymphoma. (B) Higher magnification demonstrates the pseudo‐intracellular location of the lymphoma, but cytoplasmic granules in the lymphoid cells are difficult to see in H&E sections. Touch imprints or methanolic‐based stains would demonstrate the granules better. PTAH staining to visualize granules is not reliable but sometimes is helpful. Intracytoplasmic granules are difficult to see in histopathology.

A

B

C Figure 7.54  Large granular lymphocyte (LGL) leukemia, cat (A,B) and dog (C). (A) Six‐year‐old male mixed‐breed cat with large mid‐abdominal mass. Fine‐needle aspirate of mass indicated LGL lymphoma, likely in intestine. Neoplastic cells were also found in blood and there was peripheral lymph node enlargement. Cats with intestinal LGL lymphoma and leukemia have very short survival times (2 RBC in diameter, with irregular chromatin and multiple large central nucleoli. One cell is undergoing necrosis (arrow) and below this is a mitotic figure. It can be difficult to distinguish apoptosis and mitosis: eosinophilic cytoplasm and round nucleus (arrow) favor apoptosis, whereas hairy extensions from the metaphase aggregate indicate mitotic figure. Eosinophils are numerous in this example. Infrequently lymphomas, as well as other neoplasms (fibrosarcoma), can cause paraneoplastic eosinophilia.

A

B

Figure 7.58  Large T‐cell lymphoma, subcutis, same dog as in Figure 7.57. (A) CD3: About a third of the cells are lightly positive and a few are strongly

positive. Large oval nuclei were stromal cells. (B) CD79a: Tumor cells are completely negative. Interpretation: Most lymphomas are more uniform and have a higher percentage of positive cells than seen in this example. When CD79a is negative and CD3 is not conclusive, consider using other B‐cell antibodies (CD20). Loss of CD3 expression is a feature of some T‐cell lymphomas. When CD3 is positive in only a fraction of the neoplastic cells the T cells could be of gamma/delta type. Additional batteries of antibodies are needed for this diagnosis (CD11d). If results are still inconclusive, consider PCR for clonality, additional sections of subcutis, and examination of a lymph node. This is a good example of combining morphology, which provided the diagnosis of ­lymphoma with ancillary tests to help identify cell of origin.

This forms a grid‐like pattern that is appreciated at low or medium magnifications (Figure  7.59). The vessels have deeply stained cyto­ plasm and vesicular endothelial nuclei. The enclosed lymphocytes are of variable type with round to oval nuclei of 1.5–2.0 RBC in diameter with branched chromatin and a single prominent central nucleolus. The cytoplasm is relatively abundant, deeply stained, and ampho­ philic. The amphophilic cytoplasm and eccentric nuclei impart a ­plasmacytoid appearance. The mitotic count is low. About one‐third to half of the cells are strongly positive with CD3 but many cells are unmarked, perhaps due to the γδ receptor configuration.3 The only cells positive with B‐cell markers are an occasional plasma cell.

Aggressive NK‐cell leukemia and blastic lymphoma

The NK‐cell leukemias and lymphomas are rarely encountered in humans or animals and few labs in veterinary medicine have the capability to diagnose these tumors, so they are largely unknown. In the WHO classification these are considered separate diseases. There are no markers to recognize NK tumors in domestic ani­ mals, so this designation is extrapolated from all the data. It is important to note that they do not mark with surface receptors for B or T lymphocytes.1,4 There is no phenotypic specificity to NK tumors. They are germ‐line cells without rearrangement of the

Tumors of the Hemolymphatic System    285

A

B

C Figure 7.59  Angioimmunoblastic T‐cell lymphoma (AILT), lymph node, dog. (A) At low magnification there are clear spaces and the tumor cells are in

clusters or are individualized, creating a grid‐like pattern. (B) There are small vessels within the tumor. The neoplastic cells have eosinophilic cytoplasm and are surrounded by clearly recognizable halos (inset). Neoplastic cells vary from intermediate to large size; chromatin is densely stained and the larger cells have irregular dispersion of chromatin with prominent central nucleoli. The cytoplasm is densely stained and the clear areas are not artifact. These tumors could be misinterpreted as mast cell tumors via H&E. (C) CD3 to right, CD79a to left: CD3 strongly marks about a third of the tumor cells and CD79a is negative (the nuclear staining is an artifact). AILT is a recognized entity in human medicine but it is poorly defined in veterinary medicine. This case resem­ bles the disease seen in humans.

T‐cell receptor gene and therefore express neither the more common αβ nor the less common γδ type of T‐cell antigen receptor.1,4 These cells usually express the ε‐chain part of the CD3 complex and as a result may have cytoplasmic labeling but not membrane reactivity. The absence of T‐cell or B‐cell receptors can also be due to reasons other than NK lineage (e.g., methodologies, errors, prolonged storage). In the human forms of this tumor the cells are consistently of intermediate size and of LGL cytoplasmic type, with moderately abundant clear cytoplasm that contains a few small eosinophilic granules.1,4 The cell type is the same for both the leukemic and solid forms of the neoplasm and in humans both are associated with EBV infection. The tumor cells are identified by positive immunoreac­ tivity with CD56, CD2, T1A‐1, and granzyme B, with the T‐cell receptor remaining germ line (negative).1,4 The tumor cells exhibit their deleterious effect through major histocompatibility unre­ stricted cytotoxicity.

Aggressive NK‐cell leukemia This tumor occurs in people of Asian origin with a mean age of approximately 40 years.1,4 The clinical presentation includes fever, hepatosplenomegaly, lymphadenopathy, and leukemia. Neoplastic cells in the blood vary from normal‐appearing LGL cells to larger and more atypical cells with lobulated nuclei and lightly basophilic and granulated cytoplasm. The bone marrow may be lightly to heavily invaded by the tumor. The condition is rapidly fatal with a median survival of approximately 2 months. The major differential diagnosis is the more common LGL leukemia that follows a more indolent progression in humans. Blastic NK‐cell lymphoma The extranodal NK‐cell blastic lymphoma has a higher prevalence in Asian, Mexican, and South American Indian populations.1,4 The presentation is almost always in adults, with a 2:1 male predomi­ nance. The syndrome includes the lymphoma previously known as

286    Tumors in Domestic Animals

lethal midline granuloma or nasal lymphoma. Other sites include the skin, gastrointestinal tract, lung, eye, and soft tissues. Skin lesions ulcerate and present with areas of necrosis. The disease in dogs that may be similar is hepatosplenic lym­ phoma (see earlier section). The neoplastic cells have moderate cytoplasmic staining with CD3 but not membrane and are found on clonal examination to have germ‐line T‐cell receptor.4 The neo­ plastic cells are LGL or nongranulated lymphocytes that may be small or large. Some nuclei will be indented, reniform, or folded. Giemsa‐stained preparations with a methanolic base are recom­ mended for visualizing the cytoplasmic granules. Involvement of the gastrointestinal tract is usually transmural and there may be coagulative necrosis. In deeper skin lesions the NK tumor resem­ bles the panniculitis‐like T‐cell lymphoma.5,8

References

1. Cheuk, W. and Chan, J.C. (2011) NK‐cell neoplasms. In Hematopathology (eds. E.S. Jaffe, N.L. Harris, J.W. Vardiman, et  al.). Saunders/Elsevier, Philadelphia, PA, pp. 473–491. 2. McDonough, S.P., Van Winkle, T.J., Valentine, B.A., et al. (2002) Clinicopathological and immunophenotypical features of canine intravascular lymphoma (malignant angioendotheliomatosis). J Comp Pathol 126:277–288. 3. Valli, V.E. (2007) Angioimmunoangioblastic lymphoma. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 312–316. 4. Valli, V.E. (2007) Aggressive NK‐cell leukemia. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 309–312. 5. Valli, V.E. (2007) Subcutaneous panniculitis‐like T‐cell lymphoma. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 327–330. 6. Lane, L.V., Allison, R.W. (2012) Canine intravascular lymphoma with overt leu­ kemia. Vet Clin Pathol 41:84–91. 7. Raidal, S.L., Clark, P., and Raidal, S.R. (2006) Angiotrophic T‐cell lymphoma as a cause of regenerative anemia in a horse. J Vet Intern Med 20:1009–1013. 8. Sen, F., Rassidakis, G.Z., Jones, D., and Medeiros, J. (2002) Apoptosis and prolifera­ tion in subcutaneous panniculitis‐like T‐cell lymphoma. Mod Pathol 15:625–631.

Adult T‐cell leukemia/lymphoma Defining the neoplasm

The current WHO classification gave this disease its name. The human neoplasms are caused by an endemic retrovirus (HTLV‐1 virus) on the south Island of Japan and in the Caribbean basin.1–8 The virus is highly cell associated and can be spread by cells in breast‐fed infants. This was also shown years ago as the mechanism by which BLV is spread from dam to calf.7 There is no direct animal counterpart of the HTLV‐1 virus of humans, but the mode of transmission is highly comparable to that of BLV.6,7 The conversion from a viral infec­ tion to a neoplastic disease in humans is low: approximately 1.5 males/5000 infected men occurred over a period of about 40 years. In cattle the rate of neoplastic conversion is estimated to be about 1% of infected animals and occurs over a 7‐ to 8‐year time line.7

Epidemiology and occurrence

The cells containing virus can be spread by sexual contact in humans and cattle but the disease in cattle is not spread by frozen semen. The presence of the virus and the lymphoma is much more common in dairy herds than in beef herds. This is partly due to the proximity of the animals in dairy herds and procedures such as dehorning or vaccinations if the equipment used becomes contam­ inated with blood containing BLV‐infected cells. In cats FeLV is in the blood and is not cell associated so the mode of spread is more by aerosol and fighting.8 Testing and vaccinating for FeLV has switched the pattern of lymphomas from multicentric and mediastinal in young cats to extranodal in adult cats. An aggressive lymphoma developed in genetically SCID mice given intradermal and intraperitoneal injections of splenic cells from transgenic mice

given the Tax gene of the human HTLV‐1 virus.9 These mice all died within 28 days with extensive lymphoma involving the spleen, nodes, bone marrow, liver, kidney, and lung. The tumor cells had the characteristic floral cells of the human lymphoma.

Pathology

Once the retroviral lymphoma has developed, there are many sim­ ilarities in the tumors of humans, cats, and cattle. In the human lymphoma induced by HTLV‐1 the involved tissues include the nodes, marrow, spleen, liver, gastrointestinal tract and central ner­ vous system. All of which can be involved in lymphomas caused by FeLV and BLV.6–8 The cells in the blood of humans with adult ­leukemia/lymphoma are intermediate to large with multilobated nuclei of a large atypical cell type. All of the cells in the human tumor are CD4 positive and strongly express IL‐2 and the IL receptor IL‐2R and CD25.1.3–5 The diagnosis in cattle may be made in an asymptomatic animal that has a high level of circulating lymphocytes (100 × 103/μL).6,7 In bovine lymphomas the cells are usually of large cell type (66%), about 20% are of intermediate size, and 10% are small cell type as determined on a review of over 1000 cases.6 The multilobated nuclei seen in the human adult T‐cell type of lymphoma are also present in the blood of cows with BLV‐positive lymphomas. In contrast to the retroviral‐induced human tumor, most of the large cell lymphomas in cattle are of B‐cell type.6.7 In 602 cats with lymphoma there were large cells in 54%, with 18% of intermediate cell type and 17% of cases with small cells.8 In cats the large cell lymphomas can be of either B‐ or T‐cell type.

References

1. Hasegawa, H., Sawa, H., Lewis, M.J., et  al. (2006) Thymus‐derived leukemia‐ lymphoma in mice transgenic for the Tax gene of human T‐lymphotrophic virus type I. Nature 12:466–471. 2. Jaffe, E.S. (2011) Adult T‐cell leukemia/lymphoma. In Hematopathology (eds. E.S. Jaffe, N.L. Harris, J.W. Vardiman, et  al.). Saunders/Elsevier, Philadelphia, PA, pp. 521–531. 3. Semmes, J.O. (2006) Adult T cell leukemia: a tale of two T cells. J Clin Invest 116:858–860. 4. Taylor, G.P. and Matsuoka, M. (2005) Natural history of adult T‐cell leukemia/ lymphoma and approaches to therapy. Oncogene 24:6047–6057. 5. Tobinai, K., Watanabe, T., and Jaffe, E.S. (2010) Human T‐cell leukemia virus type I‐associated adult T‐cell leukemia lymphoma. In Non‐Hodgkin Lymphomas, 2nd edn. Wolters Kluwer/Lippincott Williams & Wilkins, Philadelphia, pp. 404–414. 6. Vernau, W., Valli, V.E.O., Dukes, T.W., et al. (1992) Classification of 1,198 cases of bovine lymphoma. Vet Pathol 29:183–195. 7. Valli, V.E. (2007) Adult T‐cell lymphoma/leukemia. In Veterinary Comparative Hematopathology. Blackwell, Ames, IA, pp. 346–355. 8. Valli, V.E., Jacobs, R.M., Norris, A., et al. (2000) The histologic classification of 602 cases of feline lymphoproliferative disease using the National Cancer Institute working formulation. J Vet Diagn Invest 12:295–306. 9. Gessain, A., Mahieux, R., and de Thé, G. (1996) Genetic variability and molecular epidemiology of human and simian T cell leukemia/lymphoma virus type I. J Acquir Immune Defic Syndr 13:132–145.

Anaplastic large cell lymphoma

The lymphomagenesis of anaplastic large cell lymphoma (ALCL) is partially understood.1–7 The tumor is rare, and it may be T‐cell, B‑cell, or NK. In humans it is most commonly T‐cell or NK. Some tumor cells will also express CD30 cytokine receptor for the tumor necrosis factor family. Additionally most cases also express the cytotoxic granule–associated proteins. A morphologic signature is horseshoe‐shaped nuclei in some of the large anaplastic cells, so‐ called “hallmark cells” of ALCL. ALCL also has Hodgkin’s‐like cells admixed with numerous large undifferentiated cells.2 There will be  variable mixtures of histiocytes, neutrophils, eosinophils,

Tumors of the Hemolymphatic System    287

lymphocytes, and plasma cells. Inflammation in the tumor can make the diagnosis challenging. A genetic signature is t(2;5) chro­ mosomal translocation that results in the anaplastic lymphoma kinase gene fusing with the nucleoplasmin gene.6 This fusion results in the production of the anaplastic lymphoma kinase (ALK) pro­ tein, which can be detected by IHC and is used for the diagnosis of ALCL in humans. In humans an ALK‐negative form of ALCL is seen in older patients.7 There are cutaneous and systemic forms of this disease. There is an animal model in mice1 and dogs have tumors that fit morphologic descriptions of ALCL.7 Cutaneous anaplastic large cell lymphoma T cell: Non-epitheliotropic cutaneous lymphoma In dogs the cutaneous form of ALCL has a diffuse infiltration of large lymphoid cells that are primarily in the superficial dermis, nonepitheliotropic (NE).7 The neoplastic cells extend from the basement ­membrane of the epidermis into the dermis and may have irregular deep extensions into the panniculus but infiltration of epidermis or hair follicles is absent. When this pattern is homogenous in HE or CD3 the diagnosis is straightforward (Figures 7.60 and 7.61). However, cases with pleocellular infiltration by histiocytes, lymphocytes, eosin­ ophils, and neutrophils can mimic cutaneous histiocytosis or even

Figure 7.60  Anaplastic large cell lymphoma (ALCL), cutaneous nonepithe­ liotropic T‐cell type, skin, dog. The dermis is filled by neoplastic cells. The epidermis, hair follicles, and adnexa are not invaded. The epidermis does not bulge and is not ulcerated. Some cases will have thrombosis and ischemic necrosis of the dermis and subcutis.

inflammation such that IHC or PARR are required for diagnosis.8 Inflammatory NE cutaneous T-cell lymphoma is a diagnostic challenge because of its heterogeneous cellular infiltration, variable CD3 expression, and variable T-cell antigen receptor gene rearrange­ ment results (Figure 8.17). Neoplastic T cells often express CD18, which further confounds the interpretation of IHC patterns in tumors heavily infiltrated by histiocytes. Neoplastic lymphoid cells are of varying sizes, including large anaplastic cells with large nuclei 3–4 RBC in diameter. Binucleated cells are common. Characteristic horseshoe‐shaped nuclei can be found, along with nuclei that have complex foldings of nuclear membranes. Large nuclei of convoluted or reniform shape have prominent nucleoli. Neoplastic cells have relatively abundant cytoplasm that is lightly stained with indistinct cell boundaries. There are 1–5 mitoses/400× field. Part of the definition of this disease in humans is that extracuta­ neous disease is not recognized for at least 6 months following initial diagnosis. Immunoreactivity to CD30 is used in humans to assist diagnosis. Most NE-cutaneous lymphomas in dogs are of T-cell type, but intensity of CD3 expression is highly variable. When CD3 expression is partially “lost” it makes interpretation of the staining patterns difficult.8 Despite the aggressive nature suggested by the name of the tumor, the 5‐year survival in humans is 80–90%. Median survival time of 9 months has been reported in dogs with inflammatory form of NE cutaneous lymphoma.8 Systemic anaplastic large cell lymphoma of T‐cell type The systemic form of ALCL is rare in dogs or is rarely recognized as a disease entity.7 It is seen in young, large‐breed dogs that p ­ resent with severe systemic illness, generalized skin disease, enlarged nodes, and dependent edema.7 Neoplastic cells are as described for the cutaneous form. They are large cells with anaplastic nuclei that have reniform, horseshoe shapes, or multinucleation. That they are CD3‐positive helps differentiate them from histiocytic diseases. Mitotic figures are common, occurring at 10–20/400× field. Neoplastic involvement of nodes ranges from patchy to diffuse. The capsule is thinned and the subcapsular sinus compressed and irreg­ ularly obliterated. There may be vasculitis and thrombosis with hemorrhage and necrosis in surrounding tissue (Figure  7.62). Systemic ALCL in dogs has a rapid onset and progression.

Figure  7.62  A lymph node from a different dog with ALCL is no longer Figure 7.61  Anaplastic large cell lymphoma, dog. CD3: The tumor is posi­

tively stained, diffusely and strongly. Inset: Higher magnification. Inflamed types of NE cutaneous lymphoma will not be stained this homogenously. Epidermis and hair follicles are infiltrated in MF.

r­ ecognizable due to marked ischemic necrosis, hemorrhage, and congestion due to arterial thrombosis. Thrombosis and hemorrhagic necrosis of systemic tissues, as depicted here in skin and lymph nodes, are characteris­ tics of this very aggressive disease. ALCL is a high‐grade T‐cell lymphoma that typically affects young dogs.

288    Tumors in Domestic Animals

References

1. Bittner, C., Feller, A.C., Renauld, J.C., et al. (2000) An animal model for anaplastic large cell lymphoma in the immunocompetent syngeneic C57BI/6 mouse. Lab Invest 80:1523–1531. 2. Falini, B. and Gisselbrecht, C. (2010) Anaplastic large cell lymphoma. In Non‐ Hodgkin Lymphomas, 2nd edn. (eds. J.O. Armitage, B. Coiffier, P.M. Mauch, et al.). Wolters Kluwer Philadelphia, PA: pp. 415–432. 3. Kinney, M.C. and Kadin, M.E. (1999) The pathologic and clinical spectrum of ana­ plastic large cell lymphoma and correlation with ALK gene dysregulation. Am J Clin Pathol 111:56–67. 4. Krenacs, L., Wellmann, A., Sorbara, L., et al. (1997) Cytotoxic cell antigen expression in anaplastic large cell lymphomas of T‐and null‐cell type and Hodgkin’s disease: evidence for distinct cellular origin. Blood. 89:980–989. 5. Li, C., Takino, H., Eimoto, T., et  al. (2007) Prognostic significance of NPM‐ALK fusion transcript overexpression in ALK positive anaplastic large‐cell lymphoma. Mod Pathol 20:648–655. 6. Delsol, G., Lamant‐Rochaix, L., and Brousset, P. (2010) Anaplastic large cell lymphoma, ALK positive and ALK negative. In Non‐Hodgkin Lymphomas, 2nd edn. (eds. J.O. Armitage, B. Coiffier, P.M. Mauch, et al.). Wolters Kluwer, Philadelphia, PA, pp. 564–579. 7. Valli, V.E. (2007) Anaplastic large cell lymphoma. In Veterinary Comparative Hematopathology. Blackwell, Ames. IA, pp. 339–345. 8. Moore, P.F., Affolter, V.K., and Keller, S.M. (2013) Canine inflamed nonepitheliotropic cutaneous T‐cell lymphoma: a diagnostic conundrum. Vet Dermatol 24:204–211.

MYELOID NEOPLASMS

Myeloid neoplasms are clonal cancers that originate in hematopoi­ etic tissue from granulocytic, monocytic, erythrocytic, and mega­ karyocytic or mast cell precursors. Myeloid leukemia in most instances can be distinguished from lymphoid leukemia by tissue distribution, cell appearance, and cell markers. This distinction is important since the different leukemias have unique prognoses and responses to chemotherapy. Myeloid neoplasms are extremely het­ erogeneous, and may present with severe acute or very indolent ill­ ness, with hypo‐, normo‐, or hypercellular bone marrow, with marked leukocytosis or leukopenia, with marked cell dysplasia or with relatively normal cell morphology. Myelophthisis is a form of bone marrow failure with replacement or infiltration of normal hematopoietic tissue by malignant cells that release suppressive or destructive cytokines or fibroblast growth factors resulting in reduced hematopoiesis. Myelophthisis is a factor contributing to cytopenia in myeloid neoplasia. Cases with overt leukemia characterized by a large number of abnormal cells in blood or bone marrow are straightforward to diagnose, but others require data derived from sequential blood

counts, review of blood and bone marrow films, histopathology of bone marrow, and/or other hemolymphatic tissue, and immunohis­ tochemistry and/or flow cytometry to establish the diagnosis. Consultation with cytopathologists is essential to assess hematopoi­ etic cells in blood, when cell populations are heterogeneous or dif­ ficult to identify, or when dysplastic cells are present. Most leukemia cases are more complicated than looking at a single histopathology section of a solid tumor. Careful assessment of clinical history and morphological features of blood and bone marrow samples should enable a diagnosis of myeloid neoplasia to be made and then placement into one of the three main categories: acute myeloid ­leukemia (AML), myeloproliferative neoplasm (MPN, formerly called chronic leukemia), or myelodysplastic syndrome (MDS) (Table 7.1). The terms acute and chronic refer to the clinical course of disease entities and not the morphological descriptors. The pathogenesis of myeloid neoplasms consists of proliferation of hematopoietic cells with release of a variable number of neo­ plastic cells into blood, myelophthisis, suppression of normal hema­ topoiesis, and variable infiltration of spleen, liver, lymph nodes, and other tissues. Most myeloid neoplasms in animals are similar to types recognized in people; therefore the classification of myeloid neoplasms in people is reasonably easy to adapt to animals.1 This classification utilizes nomenclature based on hematological and morphological features of the neoplasm. Immunophenotyping of neoplastic cells with antibodies to specific antigens, and analysis of neoplastic cells for mutations and cytogenetic changes, are utilized to subcategorize myeloid neo­ plasms in people. Immunophenotyping of undifferentiated blast cells with a more limited range of antibodies is available at several North American and European academic veterinary laboratories, but genetic and cytogenetic analysis of cancer in animals is in its infancy. Furthermore, knowledge of the natural progression and response to therapy for different subtypes of myeloid neoplasms in domestic animals is largely lacking. Therefore, the clinician, clinical pathologist, and anatomic pathologist have to carefully integrate a thorough history, the hematology results, and bone marrow findings to derive a possible diagnosis of myeloid neoplasia. In most cases information from these sources is sufficient to make the diagnosis of leukemia and to categorize the neoplasm as AML, MPN or MDS. Immunophenotyping by flow cytometry or immunohistochemistry

Table 7.1  Categories of myeloid neoplasms Category

Definition

Subcategory

Markersa

Acute myeloid leukemia (AML)

Cytopenia and ≥20% blast cells in blood or bone marrow

According to cell morphology and/or i­mmunophenotypic features of cells: •  Acute undifferentiated leukemia (AUL) •  AML with neutrophilic differentiation •  AML with myelomonocytic differentiation •  AML with megakaryoblastic differentiation

CD34 CD4, CD11b, CD18, CD34 CD4, CD11b, CD14, CD18, CD34, MHCII CD9, CD34, CD41, CD61 NAc NA NA NA NA

Myeloproliferative neoplasms (MPN, old term = chronic leukemia)

Cytosis of mature appearing cells in blood, hypercellular bone marrow, 5:1 • Undifferentiated blast cells comprise 106/μL) with none or only mild cytopenia in other cell lines (i.e., anemia). If bone marrow is examined it  shows hypercellularity relative to the age of the animal and marked skewing toward one type of fairly differentiated cells and their precursors. Splenomegaly is usually prominent and hepato­ megaly may be present. There usually are leukemic cells in hepatic sinusoids, while in lymphoma or extramedullary hemato­ poiesis (EMH) neoplastic cells are at periportal or perivascular sites. Organ‐infiltrating cells in MPN are of one cell type (i.e., segmented neutrophils), whereas in lymphoma infiltrates lack differentiated granulocytes, rubricytes, and megakaryocytes, and  EMH comprises more than one cell lineage and different stages of each. Immunophenotyping is not required to identify the neoplastic cells in MPN. MPN are clonal hematopoietic cell disorders with common genetic lesions. The actual phenotype of the predomi­ nant cells is determined by mutations that usually involve a specific growth factor receptor (Table 7.1). Most MPN in people and animals can be diagnosed from history and CBC since they consist of accumulations of relatively normal‐appearing seg­ mented neutrophils (chronic neutrophilic leukemia, CNL), monocytes (chronic monocytic leukemia, CMoL), platelets (essential thrombocythemia), erythrocytes (polycythemia vera), or other leukocytes (Figure 7.67). Blast cells in bone marrow are 95%). The cell composition closely recapitulates the cytologic picture.

morphologically or functionally dysplastic cells. The proportion of bone  marrow blasts is 4/10 HPF at 400×), pseudo‐palisading necrosis, and microvascular proliferation.

Immunohistochemistry

Well‐differentiated ependymomas have a uniform and consistent cytoplasmic GFAP immunoreactivity. Vimentin and S100 are also expressed but to a much lesser intensity and extent, while pancyto­ keratin expression (Lu5, AE1AE3)is variably immunonegative and almost always Olig2 immunonegative. The PI is very low at  95%). Cytology is an excellent means to diagnose MCT that is rapid, accurate, less invasive, and less expensive than histopathology and is performed at the patient’s side. There is a resurgent interest in evaluating FNA cytology to grade MCT.2–4 One study used 37 low‐grade (LG) and 13 high‐grade (HG) MCT to evaluate the applicability of the two‐tier system on the grading of cytology samples that contained approximately 1000 cells. Through application of the same histopathology cut‐off values the authors were able to correctly predict the histologic grade in 94% of the cytology cases stained with May Grünwald–Giemsa (sensitivity 85%, specificity 97%).2 Another study characterized the utility of cytology to predict MCT grade using both features of two‐tier histologic grading system and other cellular and extracellular features on 100 or 500 cell analyses.3 The study evaluated 48 cases of MCTs (39 LG and 9 HG) and identified that: moderate to marked anisokaryosis, karyomegaly, ≥5 binucleate cells, and low tumor cell granularity were pre­ dictive of HG MCT using 500 cell counts. Additionally, a finding of low tumor cell granularity or ≥5 binucleate cells was 100% sensitive and 97% specific for the detection of HG MCT using 500 cell counts.3 This algorithm should be confirmed with samples evaluated without the restriction of a predetermined cell count. A recent publication used similar cellular and nuclear criteria of the two‐tier histologic system to evaluate 152 cutaneous MCTs cytologically.4 There were approximately 90% LG, 10% HG MCTs with a 77% agreement (three anatomic pathologists) for assigned histologic grades and 76% agreement for assigned cytology grades (three clinical pathologists). Using the cytology grading scheme outlined below, 128 of 135 histologic LG MCTs were LG via cytology (95% specificity), 15 of 17 histologic HG MCTs were HG via cytology (88% sensitivity). Two histologic HG MCTs were considered LG via cytology (12%) and 7/135 (5%) LG MCTs were considered HG via cytology. A cytological diag­ nosis of LG MCTs has 98% likelihood of being LG via histology and a cytological diagnosis of HG MCTs has a 68% likelihood of being HG via histology. Therefore, approximately 32% of the cytological HG tumors may be LG via histopathology. Histopathology to confirm HG as well as FNA of regional lymph node(s) and staging would likely be recommended by most oncologists. A cytological grade of LG should be sufficient to elect treatment options for most oncologists and owners. Follow‐up data were obtained for a 2‐year period from 13 HG and 126 LG MCTs.4 Dogs with cytologic or histologic HG MCTs were 25 times and 39 times (respectively) more likely to die within the 2‐year follow‐up period, respectively, than dogs with LG MCTs. The MST for dogs with HG MCT was approximately 365 days versus 560 days for LG MCTs; results were similar for histologic and cytologic groups. Probability of developing additional tumors was approximately 35% for HG and 8% for LG; histologic or cytologic grade. Recurrence was more likely for HG. The cytological features with the highest association with survival were mitotic figures, multinucleation, and granularity. Nuclear pleomorphism had the lowest association with survival and correlated poorly with histologic grade.2,4 Future studies should deter­ mine if this criterion should be eliminated from cytologic and histologic schemes. Granularity of tumor cells, mitoses, and mulitnucleation appear to be the best cytologic features on which to determine grade. Many other details can be found in the original publications.1–4 Cytological diagnosis of MCT is accurate. Grading via the cytologic criteria below correlates very well with histologic grades and provides prognostic information that is useful to plan treatment options. Knowledge of HG versus LG canine MCTs prior to excision could influence surgical and medical treatments.

Cytological criterion

Methanolic Romanowsky type* stain: >100 intact mast cells. HG: 1. Poorly granulated or 2. At least 2 of 4 findings: mitotic figures, multinucleated (or binucleated) cells, nuclear pleomorphism, or >50% anisocytosis. A tumor that is well granulated but has 2 of the 4 features above would be HG. *Note: Methanolic Romanowsky type. These are alcohol‐based stains such as Wright–Giemsa or May Grünwald. Do not use aqueous‐ based stains such as Diff‐Quik as cytoplasmic granules do not always stain adequately even if fixed in methanol (see Figure 7.39C,D). There are aqueous Romanowsky stains that are hand dipped or used in auto‐stainers.

References

1. Kiupel, M., et al. (2011) Vet Pathol 48:147–155. 2. Scarpa, F., et al. (2014) Vet Comp Oncol. DOI: 10.1111/vc0.12090. PubMed PMID: 24717019 3. Dekker, S., et al., eds. ACVP/ASVCP/STP Combined Annual Meeting 2015, Minneapolis, MN. 4. Camus, M.S., et al. (2016) Vet Pathol. DOI: 10.1177/0300985816638721

Appendix   953

Evaluation of regional lymph node metastasis in canine cutaneous mast cell tumors Erika L. Krick and Donald J. Meuten Although tumor grade remains the most consistent prognostic indicator for cutaneous mast cell tumors (MCTs) in dogs, other clinical information should be integrated with the tumor grade in order to develop an appropriate treatment plan and estimate prognosis. One key aspect of this information is lymph node status, which will often change further staging and treatment recom­ mendations and may affect prognosis. A study that presented cytological criteria for lymph node evaluation in dogs with MCT reported ST of approximately 8 months in dogs with cytologically confirmed lymph node metastasis (stage 2 disease) as compared to 6 years for dogs with normal or reactive lymph nodes. 1 However, this may be a self‐fulfilling prophecy. Once owners are informed that the prognosis is not as good due to regional node involvement, this information may influence their decisions to treat, which in turn affects survival data. Since the publication of that study, several additional studies examining treatment outcome in dogs with stage 2 MCT have reported prolonged ST in those patients.2–4 Specifically, the reported MST for 21 dogs with low‐grade stage 2 MCT which were treated with adequate loco‐regional and systemic therapy (surgery ± radiation therapy to the primary tumor and lymph node as well as chemo­ therapy) was 1359 days.2 An earlier study describing the outcome of 19 dogs with stage 2 MCT (1 grade 1, 16 grade 2, and 2 grade 3) treated with surgery and radiation therapy to the primary tumor, radiation therapy to the lymph nodes, and prednisone reported an MST of 1240 days.3 A third study found no significant difference in survival between dogs with grade 2 MCT with or without metastasis to lymph nodes; treatments varied widely.4 A survival benefit was noted if the metastatic lymph nodes were removed versus not (MST not reached versus 43 months).4 For dogs with lower grades of MCT, lymph node metastasis will change the treatment recommendation but not necessarily the prognosis, provided that definitive treatment is administered.2–4 Long ST (years) and disease‐free intervals can be seen in dogs with low‐grade cutaneous or subcutaneous MCTs that have metastasized to a regional lymph node. Multiple treat­ ments can be offered. For dogs with high‐grade MCT, lymph node metastasis will affect prognosis and change treatment recommendations. The reported MST of dogs with (high) grade 3 MCT with metastatic versus non‐metastatic regional lymph nodes was 176 versus 503 days.5 Treatment of the metastatic lymph nodes also affects prognosis, with a reported MST of 240 days for dogs who had surgical removal or radiation therapy of their metastatic lymph nodes compared to 89 days for dogs who did not undergo treatment specific to the metastatic lymph node.5 Although lymph node metastasis does affect prognosis for dogs with high‐grade MCT, treatment of those metastatic lymph nodes confers a survival advantage and is therefore still appropriate to recommend in many cases. Grade 3 and high‐grade MCT have higher risk of lymph node metastasis on initial clinical staging compared to lower grade tumors.5–7 In addition, a recent study found that the risk of lymph node metastasis at diagnosis was significantly higher in tumors larger than 3 cm diameter, ulcerated tumors, tumors located on the digit, shar‐pei dogs, and dogs with substage b disease (dogs are clinically ill).6 If a patient with an MCT has metastasis to a regional lymph node (regardless of grade), then further staging, such as additional lymph node evaluation and abdominal ultrasound, are often recommended. The likelihood of additional lymph node or visceral metastasis in a patient with regional lymph node MCT metastasis ranges from 10 to 20%.6,7 In a series of 220 dogs with MCT who underwent complete staging (lymph node evaluation, chest radiographs, and abdominal ultra­ sound), none of the dogs with non‐metastatic regional lymph nodes had clinically detectable metastasis in other areas.7 All of the dogs with distant metastasis had metastasis in the regional lymph node; pulmonary metastases were not seen and are not expected in dogs with MCT. Approximately 30% had lymph node metastasis and 7% of dogs had distant metastasis. In most cases the determination of metastasis to a loco‐regional lymph node will be accomplished by FNA cytology. Mast cells can be found in lymph nodes, bone marrow, buffy coats,9 and organs of dogs that do not have MCT. There are five interpretations for cyto­ logic assessment of lymph node status, three of which are straightforward. There are always gray zones, however, in which the status may be reported as: probable or possible.1 We do not have 100% certainty for any diagnostic technique and until a marker is developed to distinguish neoplastic from non‐neoplastic mast cells the distinction will be subjective and is based on numbers of mast cells seen and if the mast cells are aggregated. The diagnoses, cytological criteria, and representative images (Wright–Giemsa stain) for dogs with a MCT are shown in Figure 1. Below is a summary of criteria for histopathological node (HN) evaluation of H&E‐stained sections of lymph nodes in dogs with MCT.8 Forty‐one dogs with nodal metastasis of MCT, 26 had follow‐up data, variety of treatments, MCT not graded, HN classification based on number and distribution of mast cells and disruption of nodal architecture; nodes may not have been sentinel. MST approximately 1800 days, 77% alive at 1 year, 74% and 60% alive at 2 and 3 years; 71% were disease‐free at 3 years and median DFI not reached. Many additional details are in this publication.8 •  HN0 (No metastasis): 0–3 individualized mast cells in sinuses and/or parenchyma per 400× field. •  HN1 (Pre‐metastatic): >3 individualized mast cells in sinuses and/or parenchyma in >4 400× fields. •  HN2 (Early metastasis): Aggregates of >3 mast cells in sinuses or parenchyma, or sheets of mast cells in sinuses. •  HN3 (Overt metastasis): Disruption or effacement of nodal architecture by mast cells.

954   Appendix

50 μm

A

50 μm

B

50 μm

50 μm

C

D

50 μm

E Figure 1  (A) Normal (no significant abnormalities): no evidence of hyperplasia, no or rare individualized mast cells seen and mature lymphocytes are predomi­ nant. (B) Reactive lymphoid hyperplasia: >50% small lymphocytes with mixed population of other inflammatory cells (plasma cells arrows, eosinophil arrowhead) and/or few individualized mast cells. (C) Possible metastasis: 2–3 mast cell aggregates of 2–3 cells; arrows eosinophils. (D) Probable metastasis: >3 aggregates of 2–3 mast cells and/or 2–5 aggregates of >3 mast cells; arrows eosinophils. (E) Certain metastasis: effacement of lymphoid tissue by mast cells and/or >5 aggregates of >3 mast cells, and/or aggregated, poorly differentiated mast cells with cytological criteria of malignancy and/or decreased or variable granulation.

References

1. Krick, E.L., et al. (2009) Vet Comp Oncol 7:130–138. 2. Lejeune, A., et al. (2013) Vet Comp Oncol 31. DOI: 10.1111/vc0.12042 3. Chaffin, K. and Thrall, D.E. (2002) Vet Radiol Ultrasound 43:392–395. 4. Baginski, H., et al. (2014) J Am Anim Hosp Assoc 50:89–95. 5. Hume, C.T., et al. (2011) J Am Anim Hosp Assoc 47:37–44. 6. Stefanello, D., et al. (2015) J Am Vet Med Assoc: 246:765–769. 7. Warland, J., et al. (2014) Vet Comp Oncol 12:287–298. 8. Weishaar, K.M., et al. (2014) J Comp Pathol 151:329–338. 9. Bookbinder, P.F., et al. (1992) J Am Vet Med Assoc 200: 1648–1650.

Appendix   955

Canine oral perioral mast cell tumors Mast cell tumors were located orally, at mucocutaneous junction, or haired muzzle (perioral).1–3 Survival time was similar for all three locations.1 Mitotic count (area not defined) and cytological examination of regional lymph node are the two best parameters to predict ST. Dogs (n = 21 total) with MC >5 had shorter MST and less time to disease progression than did dogs with tumors with MC 19 figures

1: ≤ 50% necrosis 2: > 50% necrosis

3: Undifferentiated sarcoma*

Add scores. Sum

Grade

Metastases3

Local recurrence2 if tumor at excision margin

≤3 4–5 ≥6

Grade 1 Grade 2 Grade 3

 0% 27% 22%

  7% (3/41) 35% (14/41) 75% (3/4)

•  n = 60: 27 were grade 1, 23 grade 2, and 10 grade 33 •  n = 86: 41 were grade 1, 41 grade 2, and 4 grade 3; used a different MC scoring system.2 Note: both studies modified how they followed the above means to assign a grade. If margins are complete, then 90% were CD4‐positive T‐cells; dogs are usually CD8 positive. Lymphocytosis23,24 Factors associated with survival: total lymphocytosis, CD34 positive,23,25 B versus T,19 and size of lymphocytes. MST are approximations. Canine total T‐cell lymphocyte count at presentation, MST: •  >30,000 lymphocytes/μL, 130 days23 •