Hematopoietic Tumors

Genetic and Molecular Factors

Recent advances in molecular cytogenetics (see Chapter 1, Section A), including gene microarray techniques, have been and are currently being applied to investigations of chromosomal aberrations in dogs with lymphoma.^10-16 The publication of the canine genome and the commercial availability of canine gene microarrays (GeneChip Canine Genome 2.0 Array, Affymetrix, Inc.) have led to advances in our understanding of genetic events occurring in lymphoma.¹⁷ Breen’s group has documented gain of dog chromosomes 13 and 31 and loss of chromosome 14 as the most common aberrations in a group of 25 cases analyzed.¹¹ Chromosomal aberrations have also been associated with prognosis in dogs with lymphoma. A study of 61 dogs with lymphoma demonstrated a prognostic advantage in dogs with trisomy of chromosome 13 (25% of the dogs studied) as evidenced by increase in duration of first remission and overall survival time.¹⁸ Germline and somatic genetic mutations and altered oncogene/tumor suppressor gene expression, epigenetic changes (e.g., DNA hypomethylation), signal transduction, and death-pathway alterations (e.g., Bcl-2 family) are common in human lymphomas and have been reported in the dog as well (see Chapter 1, Section A, and Chapter 14, Section B). These include N-ras, p53, Rb, and p16 cyclin-dependent kinase aberrations.^19-24 Additionally, differences in the prevalence of immunophenotypic subtypes of lymphoma among different breeds indicate heritable risks.²⁵ Additionally, telomerase activity (see Chapter 2) has been documented in canine lymphoma tissues.^26-28

Infectious Factors

The hypothesis that a retrovirus may be involved in the pathogenesis of canine lymphoma has not been confirmed. However, serologic detection of Epstein-Barr virus infection, linked to some forms of lymphoma in humans, has been documented in dogs with lymphoma and is currently being investigated.²⁹

In humans, a direct association between Helicobacter sp. infections and development of gastric lymphoma has been made.³⁰ Although this has not been definitively shown in dogs, there is evidence of Helicobacter sp. infection in laboratory beagle dogs resulting in gastric lymphoid follicle formation that is considered a precursor of mucosa-associated lymphoid tissue (MALT) lymphoma in humans.³¹

Environmental Factors

In humans, evidence has accumulated implicating phenoxyacetic acid herbicides, in particular 2, 4-dichlorophenoxyacetic acid (2, 4-D), in the development of NHL. A published hospital-based case-control study of dogs indicated that owners in households with dogs that develop malignant lymphoma applied 2, 4-D herbicides to their lawn and/or employed commercial lawn care companies to treat their yard more frequently than owners of dogs without lymphoma.³² The risk of canine lymphoma was reported to rise twofold (odds ratio [OR] = 1.3) with four or more yearly owner applications of 2, 4-D. The results of this study have come under criticism, and three additional follow-up investigations have not validated assertions of increased risk.^33-35 In another study, dogs exposed to lawn treatment within 7 days of application were greater than 50 times more likely to have urine levels of 2, 4-D at 50 µg/L or higher.³⁶ The highest concentration was noted 2 days after application. In an environmental case-control study performed in Europe, two variables, residency in industrial areas and use of chemicals (defined as paints or solvents) by owners, modestly increased the risk of developing lymphoma; however, no link was found with pesticide use.³⁷

A weak association between lymphoma in dogs and exposure to strong magnetic fields was observed in a preliminary epidemiologic study.³⁸ In this hospital-based case-control study, the risk of developing lymphoma categorized into high or very high exposure was increased (odds ratio = 1.8). More thorough studies are necessary to evaluate this association further. Proximity to environmental waste was implicated in two European studies; however, it was felt to be a risk indicator rather than a risk factor and would require further case-control investigations.^39,40

Immunologic Factors

Impaired immune function has also been implicated in dogs with lymphoma. Immune system alterations in the dog such as immune-mediated thrombocytopenia, independent of age and sex, have been associated with a higher risk of subsequently developing lymphoma when compared to the normal population.^41,42 Additional evidence comes from observations in human and feline transplantation patients. In a case-control study of cats undergoing renal transplant, 24% of cases developed cancer (36% of those were lymphoma) while on cyclosporine immunosuppressive therapy compared to 5.1% of control cats, none of which developed lymphoma (OR, 6.1; p = 0.001).⁴³ A case of lymphoma developing in a dog following treatment with cyclosporine also exists.⁴⁴ One report suggests an association between the immunodysregulation observed in dogs with atopic dermatitis and the risk of developing epitheliotropic T-cell lymphoma; whether this is associated with the disease or the immunomodulatory treatments commonly applied is unknown.⁴⁵

Atypical Anatomic Forms of Lymphoma

Hepatosplenic lymphoma is a relatively uncommon, distinct presentation in the dog marked by a lack of significant peripheral lymphadenopathy in the face of hepatic, splenic, and bone marrow infiltration with malignant lymphocytes, usually of T-cell origin.^64,65 Biologically, this form of lymphoma is extremely aggressive and poorly responsive to therapy. In humans the tumor usually is composed of γδT-cells (i.e., T-cells that express the γδT-cell receptor), and this immunophenotype has been confirmed in at least one dog in the veterinary literature.⁶⁵

Intravascular (angiotropic, angioendotheliomatosis) lymphoma is a distinct form of lymphoma defined as proliferations of neoplastic lymphocytes within the lumen and wall of blood vessels in the absence of a primary extravascular mass or leukemia. It has been reported several times in the veterinary literature, and in most cases it involves the CNS and peripheral nervous system (PNS), including the eye.^66-71 The B-cell immunophenotype is most common in humans; however, in most reported cases in dogs, the origin is either T-cell or null cell (neither B- nor T-cell), although one case of a B-cell phenotype has been reported.

Pulmonary lymphomatoid granulomatosis (PLG) is a rare pulmonary infiltrative and/or nodular disorder characterized by a heterogenous accumulation of lymphocytes (both B and T, although some evidence suggests primarily a T-cell origin), neutrophils, plasma cells, and macrophages, often arranged angiocentrically.^72-75 Whether this syndrome is a true lymphoma or a prelymphoma state is debatable. Clinical signs are related to respiratory compromise, and various chemotherapeutic protocols have been used with reported results varying from rapid progression to long-term clinical remissions.

Histologic Classification Systems

Lymphomas arise from clonal expansion of lymphoid cells with distinctive morphologic and immunophenotypic features. Many histologic systems have been used to classify NHL in humans, and some of these have been applied to lymphoma in the dog and other species. The National Cancer Institute (NCI) Working Formulation⁷⁶ and the updated Kiel system⁷⁷ have been adapted to canine tumors with some success. The World Health Organization (WHO) also publishes a histologic classification scheme, which uses the revised European American lymphoma (REAL) system as a basis for defining histologic categories of hematopoietic and lymphoid tumors in domestic animals.⁷⁸ This system incorporates anatomic, histologic, and immunophenotypic criteria (B- and T-cell immunophenotype), with the goal of enabling accurate and reproducible diagnosis of specific neoplastic disease entities. This theoretically should assist in better tailoring of treatment protocols, better correlation of prognosis, and better comparative capabilities. Table 32-1 shows some of the WHO categories in three different surveys, including a 2-year survey (2008-2009) of canine necropsy and biopsy cases at the University of Wisconsin-Madison Veterinary Medical Teaching Hospital pathology service^79,80; some of the less common categories in the WHO system were not represented and are not listed. The WHO system provides accurate and consistent reproducible diagnostic results similar to the system used in human pathology; accuracy among a group of pathologists examining 300 cases was at 83% agreement, and accuracy in evaluating the six most common diagnoses (80% of the cases) was 87%.⁸¹ Clinical studies are needed to correlate the various categories of disease with biologic behavior, response to treatment, and prognosis. Preliminary results indicate dogs with indolent lymphoma (e.g., marginal zone lymphoma, follicular lymphoma, B- or T-cell small cell lymphoma, T-cell–rich B-cell lymphoma, and T zone lymphoma) maintain normal activity and appetite levels even during advanced stages of disease and experience long-term survival even with limited or no therapy.^81-84

The Working Formulation (WF) was developed to allow investigators to “translate” among the numerous classification systems so that clinical trials could be compared in humans. Most of the larger compilations agree that most canine lymphomas are intermediate or high grade; however, diffuse immunoblastic forms appear to predominate in the United States, whereas the follicular large cell variations predominate in Europe. A comparison of European and American classifications is warranted based on this discrepancy. The WF categorizes tumors according to pattern (diffuse or follicular) and cell type (e.g., small cleaved cell, large cell, immunoblastic), but it does not include information about the immunophenotype of the tumor.⁷⁶ The WF subtypes are related to the biology of the tumor and patient survival. The updated Kiel classification includes the architectural pattern, morphology (centroblastic, centrocytic, or immunoblastic), and immunophenotype (B-cell or T-cell) of the tumor cells.⁷⁷ In both systems, the tumors can then be categorized as low-grade, intermediate grade, or high-grade malignancies. Low-grade lymphomas composed of small cells with a low mitotic rate typically progress slowly and are associated with long survival times but are ultimately incurable. High-grade lymphomas with a high mitotic rate progress rapidly but are more likely to respond initially to chemotherapy and, in humans, are potentially curable.

Several features of canine lymphomas become apparent when these classification systems are applied. The most striking difference between canine and human lymphomas is the scarcity of follicular lymphomas in the dog.^79,80 Some diffuse lymphomas in the dog may initially be follicular, but these may progress to the more aggressive, diffuse form by the time of diagnostic biopsy. The most common form of canine lymphoma is diffuse large-cell lymphoma, a high-grade tumor most commonly of B-cell origin.^80,81,85 Only a small percentage of canine lymphomas (5.3% to 29%) are considered low-grade tumors.

High-grade lymphomas occur frequently if diffuse large-cell lymphomas, classified as intermediate grade in the WF, are considered high-grade, as in the updated Kiel classification (in which they are labeled as diffuse centroblastic lymphomas). A documented difference exists in the prevalence of the various immunophenotypes based on breed.²⁵ For example, cocker spaniels and Doberman pinschers are more likely to develop B-cell lymphoma, Boxers are more likely to have T-cell lymphoma, and golden retrievers appear to have an equal likelihood of B- and T-cell tumors.

To be clinically useful, these classification systems in the end must yield information about response to therapy, maintenance of remission, and survival. Some studies suggest that the subtypes in the WF can be correlated with survival, and the Kiel system may be useful for predicting relapse.^86,87 In most studies, high-grade lymphomas achieve a complete response (CR) to chemotherapy significantly more often than low-grade tumors. However, dogs with low-grade tumors may live a long time without aggressive chemotherapy.^83,84 Dogs with T-cell lymphomas have shown a lower rate of CR to chemotherapy and shorter remission and survival times than dogs with B-cell tumors (with the exception of low-grade T-cell subtypes).^53,54,86,88 Furthermore, T-cell lymphomas tend to be associated with hypercalcemia.^89,90

In the veterinary literature, 60% to 80% of canine lymphomas are of B-cell origin; T-cell lymphomas account for 10% to 38%; mixed B- and T-cell lymphomas account for as many as 22%; and null cell tumors (i.e., neither B-cell nor T-cell immunoreactive) represent fewer than 5%.^53,54,91-93 The development of monoclonal antibodies to detect specific markers on canine lymphocytes has made immunophenotyping of tumors in dogs routinely available in many commercial laboratories. Such techniques can be performed on paraffin-embedded samples, from tissue microarrays, on cytologic specimens obtained by fine-needle aspiration (FNA) of lesions, or by flow cytometric analysis of cellular fluid samples (e.g., peripheral blood, effusions) and lesion aspirates.

The Rappaport classification system, proposed in 1956 for human NHL, describes the architectural pattern (follicular or diffuse) and the cytologic features (well differentiated, poorly differentiated, or histiocytic) of lymphoma.^94,95 This system has not proved useful in providing prognostic information or in guiding therapy in dogs with lymphoma because of the low number of follicular lymphomas in dogs, the problematic “histiocytic” subgroup, and the failure to account for different morphologic and immunologic cell types.

One criticism of the Rappaport, Kiel, and WF classification systems is that they fail to include extranodal lymphomas as a separate category. The WHO system does include anatomic location as a factor in determining certain categories. Although differences between nodal and extranodal tumors in biologic behavior and prognosis are well recognized, comparative information about the histogenesis of these tumors is lacking. For example, in humans small-cell lymphomas arising from MALT are composed of cells with a different immunophenotype than that of other small-cell lymphomas (i.e., MALT lymphomas typically are negative for both CD5 and CD10). Except for cutaneous lymphoid neoplasms, detailed characterization of extranodal lymphomas in dogs has not been done. Although cutaneous lymphoma is a heterogeneous group of neoplasms that includes an epitheliotropic form resembling mycosis fungoides and a nonepitheliotropic form, most cutaneous lymphomas have a T-cell phenotype.^64,96

To summarize, it is important to determine the histologic grade of canine lymphomas as low (small lymphocytic or centrocytic lymphomas) or intermediate to high (diffuse large cell, centroblastic, and immunoblastic lymphomas) and the architecture as diffuse or follicular. Furthermore, determining the immunophenotype of the tumor provides useful information. Response rates to chemotherapy are, in general, better in animals with B-cell tumors and intermediate- to high-grade lymphomas. Dogs with low-grade lymphomas can have long survival times without aggressive therapy.

Physical Examination

A thorough physical examination should include palpation of all assessable lymph nodes, including a rectal examination; in the authors’ experience, a significant proportion of dogs will have rectal polyps consisting of aggregates of neoplastic lymphocytes. Inspection of mucous membranes for pallor, icterus, petechiae, and ulceration should be undertaken as these signs may indicate anemia or thrombocytopenia secondary to myelophthisis or immune-mediated disease or may be evidence of major organ failure or uremia. Abdominal palpation may reveal organomegaly, intestinal wall thickening, or mesenteric lymphadenopathy. The presence of a mediastinal mass and/or pleural effusion can be suspected following thoracic auscultation. An ocular examination, including funduscopic assessment, may reveal abnormalities (e.g., uveitis, retinal hemorrhage, ocular infiltration) in approximately one-third to one-half of dogs with lymphoma.^107,111

Complete Blood Count, Biochemistry Profile, and Urinalysis

Anemia, the most common lymphoma-related hematologic abnormality, is usually normochromic and normocytic (nonregenerative), consistent with anemia of chronic disease.¹⁰⁸ However, hemorrhagic and hemolytic anemias may also occur, and regenerative anemias may reflect concomitant blood loss or hemolysis. Additionally, if significant myelophthisis is present, anemia may be accompanied by thrombocytopenia and leukopenia.^112,113 In animals with anemia or evidence of bleeding, in addition to a platelet count, a reticulocyte count and coagulation testing may be indicated. Thrombocytopenia may be seen in 30% to 50% of cases, but bleeding is seldom a clinical problem. Neutrophilia can be seen in 25% to 40% of dogs and lymphocytosis occurs in approximately 20% of affected dogs.¹⁰⁸ Circulating atypical lymphocytes may be indicative of bone marrow involvement and leukemia. It is important to differentiate multicentric lymphoma with bone marrow involvement (i.e., stage V disease) from primary lymphoblastic leukemia (see later), as the prognosis for each is entirely different. Hypoproteinemia is observed more frequently in animals with alimentary lymphoma. In dogs with a high total protein or evidence of an increased globulin fraction on a chemistry profile, serum proteins may be evaluated by serum electrophoresis. Monoclonal gammopathies have been reported to occur in approximately 6% of dogs with lymphoma.¹¹⁴

Serum biochemical abnormalities often reflect the anatomic site involved, as well as paraneoplastic syndromes such as hypercalcemia. In cases of hypercalcemia of unknown origin, lymphoma should always be considered high on the differential disease list, and diagnostic testing directed at this possibility should be undertaken (see Chapter 5). In addition, the presence of hypercalcemia can serve as a biomarker for response to therapy and early recurrence. Increased urea nitrogen and creatinine concentrations can occur secondary to renal infiltration with tumor, hypercalcemic nephrosis, or prerenal azotemia from dehydration. Increases in liver-specific enzyme activities or bilirubin concentrations may result from hepatic parenchymal infiltration. Increased serum globulin concentrations, usually monoclonal, occur infrequently with B-cell lymphoma.

Urinalysis is part of the minimum database used to assess renal function and the urinary tract. For example, isosthenuria and proteinuria in the absence of an active sediment may indicate renal disease, and hematuria may result from a hemostatic abnormality. It is important to remember that isosthenuria in azotemic dogs with hypercalcemia is not necessarily indicative of renal disease as the high calcium levels interfere with tubular concentration capabilities through disruption of antidiuretic hormone (ADH) control.

Several abnormalities in serum have been explored as biomarkers of lymphoma in the dog. Examples include alpha-fetoprotein, alpha-1 glycoprotein levels, zinc, chromium, iron, endostatin, vascular endothelial growth factor (VEGF), lactate dehydrogenase, C-reactive protein haptoglobin, and antioxidants/oxidative stress markers.^115-122 The clinical, biologic, and prognostic significance of these alterations is yet to be definitively characterized.

Histologic and Cytologic Evaluation of Lymph Nodes

Morphologic examination of the tissue and cells that constitute the tumor is essential to the diagnosis of lymphoma. Care should be taken to avoid lymph nodes from reactive areas (e.g., mandibular lymph nodes), unless those nodes are the only ones enlarged; the prescapular or popliteal lymph nodes are preferable if also involved. Also, lymphoid cells are fragile, and in preparing smears of aspirated material only gentle pressure should be applied in spreading material on the slides. In most cases, a diagnosis of lymphoma can be made on evaluation of fine-needle aspirates of affected lymph nodes or other tissues. Typically, most of the cells are large lymphoid cells (>2 times the diameter of a red blood cell [RBC] or larger than neutrophils), and they may have visible nucleoli and basophilic cytoplasm (Figure 32-6, A) or fine chromatin with indistinct nucleoli. Because tissue architecture is not maintained in cytologic specimens, effacement of the node or capsular disruption cannot be detected. Therefore marked reactive hyperplasia characterized by increased numbers of large lymphoid cells may be difficult to distinguish from lymphoma, and small cell lymphomas may have few cytologic clues that point to malignancy. Also, classification of lymphoma, which has been attempted using cytologic appearance and immunophenotypic analysis,¹²³ into subcategories that make up the low-, intermediate-, and high-grade forms is performed most accurately on histologic sections (discussed previously).

For accurate histopathologic evaluation, an entire lymph node, including the capsule, should be removed, placed in buffered formalin, and submitted to a pathologist. Although needle core biopsies may be satisfactory, it is important to avoid crush artifact or inadequate sample size. Most pathologists prefer whole node biopsies because they provide the maximal amount of information. Effacement of normal nodal architecture by neoplastic lymphocytes and capsular disruption are characteristic findings (Figure 32-6, C and D). Diagnostic ultrasonography and ultrasound-guided FNA or needle biopsy have been useful for evaluation of involvement of the liver, spleen, or abdominal lymph nodes.^124-126 Aspiration of ultrasonographically normal splenic tissue is rarely contributory to a diagnosis.¹²⁴ If possible, the diagnosis should be made by sampling peripheral nodes, avoiding percutaneous biopsies of the liver and spleen. However, if there is no peripheral node involvement, it is appropriate to biopsy affected tissues in the abdominal cavity.

Histologic and Cytologic Evaluation of Extranodal Sites

When GI lymphoma is suspected, an open surgical wedge biopsy of the intestine is preferred in most cases to differentiate lymphoma from lymphocytic enteritis. If associated abdominal lymph nodes also appear involved, image-guided biopsies may be associated with less morbidity than intestinal biopsies. Multiple samples may be necessary to accurately diagnose segmental disease. Endoscopic biopsies may be inadequate as only a superficial specimen is obtained; however, more aggressive endoscopic biopsy techniques combined with more accurate histopathologic, immunophenotypic, and molecular assessments are improving the diagnostic yield of these less invasive techniques.^127-130 In many dogs with primary GI lymphoma, an inflammatory nonneoplastic infiltrate (i.e., LPE) may be misdiagnosed on biopsy specimens that are too superficial. The application of assays for clonal expansion (e.g., PARR—see next section on molecular diagnostic techniques) does not appear as yet to be as accurate for endoscopically derived intestinal biopsies as with other solid lymphoid tumors in dogs.

Cytologic examination of cerebrospinal fluid (CSF), thoracic fluid, or mass aspirates is indicated in animals with CNS disease, pleural effusion, or an intrathoracic mass, respectively. In one study of dogs with CNS involvement, CSF analysis was diagnostic in seven of eight dogs.¹⁰³ Characteristics of the CSF included an increased nucleated cell count in the seven dogs, and 95% to 100% of the cells were atypical lymphocytes. The CSF protein concentration was increased in five of the dogs, ranging from 34 to 310 mg/dL (reference interval: <25 mg/dL).

For cutaneous lymphoma, punch biopsies (4 to 8 mm) should be taken from the most representative and infiltrative, but not secondarily infected, skin lesions. Application of immunophenotypic and clonality assessments of cutaneous biopsies can aid in differentiating lymphoma from benign lymphocytic lesions.^58,131,132

Molecular Diagnostic Techniques

Molecular techniques can be used to establish a diagnosis of lymphoma or to further characterize the tumor after the initial diagnosis is made. Tissues and cells from peripheral blood, lymph nodes, nonlymphoid sites, and effusions can be analyzed by various molecular means to aid in cases that represent a more difficult diagnostic challenge, particularly in cases where reactive lymphocytosis and lymphoma are both possible based on standard histologic or cytologic assessment. These include histochemical and cytochemical, immunohistochemical and immunocytochemical, flow cytometric, and polymerase chain reaction (PCR) techniques. For example, the immunophenotype (B-cell versus T-cell),^93,133-140 proliferation rate (e.g., expression of Ki67, proliferating cell nuclear antigen [PCNA] expression, argyrophilic nucleolar organizer regions [AgNOR]),^{53,133,141-145} and clonality (PCR for antigen receptor gene rearrangement [PARR])^146-154 of the tumor can be determined. The availability of such analyses is increasing; however, at present, only immunophenotype and PARR clonality assays are routinely used in dogs to inform clinical decision making.

Immunophenotyping

Immunophenotyping is used to determine the type of cells that comprise the tumor, but this technique also can be helpful for making the initial diagnosis.^133-140,155 When a heterogenous population of lymphocytes is expected in a tissue, documentation of a homogeneous population of the same immunophenotype is supportive of a neoplastic process. The immunophenotype of a lymphocyte is identified by determining the expression of molecules specific for B-cells (e.g., CD79a, CD20) and T-cells (e.g., CD3). Although tumor cells sometimes have morphologic characteristics that typify a particular immunophenotype, exceptions occur, and morphologic appearance cannot be used as the sole determinant of immunophenotype. For example, in a series of nine high-grade T-cell lymphomas and leukemias in dogs, the cells had a plasmacytoid appearance, typically associated with B-cell lymphoma.¹⁵⁶ Similarly, anatomic location does not always predict the immunophenotype.

For accurate determination of immunophenotype, antibodies against lymphocyte markers are applied to tissue sections (immunohistochemistry), cytologic specimens (immunocytochemistry), or individual cells in a fluid medium (flow cytometry). Flow cytometric evaluation of cells obtained by needle aspiration is also feasible. For T-cells, markers include CD3 (pan T), CD4 (helper T), and CD8 (cytotoxic T); for B-cells, the markers are CD79a (Figure 32-6, B), CD20, and CD21. Increasingly, aberrant expression of CD molecules has been reported in canine lymphoma. In a study of 59 dogs with lymphoma, tumor cells from six dogs were positive for both T- and B-cell markers; however, a clonality assay (see later) revealed clonality either of the T-cell or the immunoglobulin receptor but not both. This indicates that in some cases, the malignant cells may co-express B- and T-cell markers.⁹³ Antibodies against these molecules are used to determine the immunophenotype; however, they also have potential utility as a therapeutic modality if tumor cells could be targeted using these antibodies.

Other Immunohistochemical and Immunocytochemical Assessments

Assessments of markers of multidrug resistance and apoptotic pathways (e.g., P-glycoprotein, p53, Bcl-2 proteins) have been evaluated in dogs with lymphoma.^{19,79,142,156a,156b} However, their clinical significance and utility await further evaluation.

Clonality Assays

Occasionally, diagnosis of lymphoma and differentiation of malignant versus benign proliferation of lymphocytes are not possible based on standard histologic and cytologic criteria. In these cases, advanced molecular analyses may be helpful to confirm a diagnosis. Clonality is the hallmark of malignancy; that is, the malignant cell population theoretically should be derived from expansion of a single malignant clone characterized by a particular DNA region unique to that tumor. For example, in a dog with T-cell lymphoma, all the malignant cells theoretically should have the same DNA sequence for the variable region of the T-cell receptor gene. Likewise, in a dog with B-cell lymphoma, the tumor cells should have identical DNA sequences in the variable region of the immunoglobulin (Ig) receptor gene. Conversely, in reactive lymphocytosis, the cells are polyclonal for their antigen receptors. Using this knowledge, investigators have used PCR technology to amplify the variable regions of the T-cell and immunoglobulin receptor genes to detect the presence of clonal lymphocyte populations in dogs (see Figure 8-4 of Chapter 8). These techniques are reviewed in Chapter 8 and elsewhere.¹⁵⁰ In physician-based medicine, such assays of clonality are approximately 70% to 90% sensitive and have a false-positive rate of approximately 5%, and recent studies report similar rates in dogs. False-negative and false-positive results can occur with clonality assays. For example, cells from a dog with lymphoma may be negative for clonality if the clonal segment of DNA is not detected with the primers used, if the malignant cells are natural killer (NK) cells (rare), or if the malignant cells are present in too low a frequency to be detected. False positives occur rarely in some infectious diseases (e.g., ehrlichiosis and Lyme disease). In these cases, a diagnosis should be made only after considering the results of all the diagnostic tests, including histologic/cytologic evaluation, immunophenotyping, and clonality studies in conjunction with signalment and physical examination findings. These molecular techniques, although helpful for diagnosis, could also have utility in detecting early recurrence and in determining more accurate clinical stage and so-called “molecular remission rates” because they are more sensitive than standard cytologic assessment of peripheral blood, bone marrow, or lymph nodes (covered subsequently in section on treatment response).

Proteomics

Proteomics comprises, simplistically, methodologies that analyze the entire protein component or protein signature of cells (the proteome). Protein components of a cell (normal or malignant) change over time with upregulation and downregulation of gene expression in response to varied stimuli (e.g., growth factors, environmental cues). It may therefore be possible to use the field of proteomics to identify serum biomarkers of malignancy (i.e., cancer-specific protein markers) and to further analyze response to therapy or even to predict which therapies are appropriate for an individual patient’s tumor. Although in its infancy in veterinary oncology, preliminary investigations of the proteome of dogs with lymphoma have been reported^157-160; however, they have yet to reach the level of sophistication in which useful output would have an impact on clinical decision making.

Bone Marrow Evaluation

A bone marrow aspirate or biopsy (from proximal humerus or iliac crest) is recommended for complete staging and prognostication and is indicated in dogs with anemia, lymphocytosis, peripheral lymphocyte atypia, or other peripheral cytopenias. In one study of 53 dogs with lymphoma, 28% had circulating malignant cells and were considered leukemic, whereas bone marrow examination indicated involvement in 57% of the dogs.¹⁶² The presence of a few prolymphocytes and large lymphocytes with nucleoli in the circulation of dogs with lymphoma may indicate bone marrow involvement. It is important to remember these cells also can be seen with GI parasitism, immune-mediated hemolytic anemia, and other immune-mediated and infectious diseases. As discussed previously, tumor cells within the peripheral and bone marrow compartments can also be identified using clonality assays (PARR) that are more sensitive than routine microscopic examination in detecting malignant cells; however, the prognostic significance of the knowledge gained with more sensitive staging methodologies is yet to be determined. Although bone marrow evaluation may offer prognostically valuable information, it is not necessary to perform the procedure if the client is committed to treat regardless of stage.

Imaging

Evaluation of thoracic and abdominal radiographs may be important in determining the extent of internal involvement (Figure 32-7). Approximately 60% to 75% of dogs with multicentric lymphoma have abnormalities on thoracic radiographs, with one-third having evidence of pulmonary infiltrates (see Figure 32-3) and two-thirds having thoracic lymphadenopathy (sternal and tracheobronchial lymph nodes [see Figure 32-7]) and widening of the cranial mediastinum (see Figure 32-2).^97,98 Pulmonary infiltrates usually are represented by an interstitial and/or alveolar pattern; however, nodules (rarely) and bronchial infiltrates can also occur.¹⁶³ Pleural effusion may also be present. Cranial mediastinal lymphadenopathy is detected in 20% of dogs with lymphoma.^97,163 Abdominal radiographs reveal evidence of involvement of medial iliac (sublumbar) and/or mesenteric lymph node, spleen, or liver in approximately 50% of cases. In the authors’ practice, for the typical cases of canine multicentric lymphoma, imaging is limited to thoracic radiographs as there is no prognostic difference between dogs with stage III and IV disease (i.e., liver/spleen involvement), whereas the presence of cranial mediastinal lymphadenopathy is of prognostic significance (see prognosis section). However, if there are clinical signs attributable to abdominal disease or if complete staging is necessary (e.g., for clinical trial inclusion), further imaging of the abdomen is warranted. Abdominal ultrasonography can be important for obtaining ultrasound-guided intraabdominal samples for diagnosis. It may also be useful for the diagnosis of GI, abdominal nodal, and hepatosplenic lymphoma.¹²⁵ Ultrasonographic (including Doppler ultrasound) assessment of peripheral lymph nodes has also been explored¹²⁶; however, its clinical applicability is questionable because cytologic assessment of peripheral nodes is easy, inexpensive, and of higher diagnostic utility.

Advanced imaging modalities, including computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), or PET/CT imaging, are becoming more commonplace in veterinary practice and their utility is only now being determined.^164-169 PET/CT imaging is the current standard of care for following and indeed predicting durability of treatment response in human patients with lymphoma, and both [18F]fluorothymidine (18FLT) PET/CT and [18F]fluoro-D-glucose (18FDG) PET imaging have been reported in dogs with lymphoma.^164-166 18FLT-PET/CT functional and anatomic imaging shows promise for the evaluation of response to cytotoxic chemotherapy in dogs with lymphoma and for predicting relapse before standard clinical and clinicopathologic confirmation (Figure 32-8).

Multidrug Combination Protocols

Many chemotherapeutic protocols for dogs with lymphoma have been developed over the past 15 to 20 years (Table 32-3).^91,170-190 Significant limitations arise when comparing efficacy studies in the veterinary literature for the various published protocols. Few of these studies include sufficient numbers for adequate statistical power and even fewer compare treatment protocols in a randomized prospective fashion. In addition, staging, inclusion, and response criteria vary considerably between reports. Therefore evaluations of efficacy among various protocols are subject to substantial bias, making direct comparisons difficult and indeed precarious. A recurring theme in the concluding statement in most of these published protocols is some variation of “prospective randomized trials will be required to confirm these suggestive findings.” In an attempt to better standardize response criteria and outcome reporting of future trials, the Veterinary Cooperative Oncology Group (VCOG) has recently published response evaluation criteria (v1.0)¹⁶⁹ (see subsequent response evaluation section). The greatest obstacle to the performance of prospective randomized comparative lymphoma trials in veterinary oncology is financial; that is, clinical trials are inherently costly, and because most of the known effective drugs are unregistered off-label human generic (i.e., off patent) drugs, the incentive for pharmaceutical-funded, sufficiently powered, randomized field trials is low, resulting in a general lack of comparative data.

Despite the plethora of available combination protocols, most are modifications of CHOP protocols initially designed for human oncologic use, and currently randomized prospective evidence does not exist to clearly recommend one over the other as long as the basic “CHOP” components are present. CHOP represents combinations of cyclophosphamide (C), doxorubicin (H, hydroxydaunorubicin), vincristine (O, Oncovin), and prednisone (P). Conventional CHOP-based chemotherapy induces remission in approximately 80% to 95% of dogs, with overall median survival times (MSTs) of 10 to 12 months. Approximately 20% to 25% of treated dogs will be alive 2 years after initiation of these protocols (Figure 32-9). Response rates and duration of response vary according to the presence or absence of prognostic factors discussed subsequently in the section on prognosis in this chapter. The relative cost of the various protocols to the client depends on the drug(s) selected, the size of the animal, the frequency of administration, and the laboratory tests required to monitor adverse events and response.

Dogs responding to chemotherapy and undergoing complete remission are usually free of clinical signs associated with lymphoma and subsequently return to a very good quality of life. Treating dogs with lymphoma is initially gratifying because a high percentage enjoy a complete response. Most dogs tolerate chemotherapy well, and although dose reductions and treatment breaks (“treatment holidays”) are not uncommonly required in individual cases, only a minority of dogs develop significant adverse events requiring hospitalization.^191,192 Studies assessing client perceptions of medical treatment for cancer in general and lymphoma in particular report a positive experience; most owners feel treatment was worthwhile, that it resulted in improvement in the well-being of their pet, and that quality of life during treatment was good.^193,194 Very few clients express regret about treating lymphoma using a multidrug protocol.

With lymphoma, the fundamental goals of chemotherapy are to induce a complete durable (>6 months) first remission (termed induction), to reinduce a remission when the tumor recrudesces (or the patient relapses) following achievement of a remission (termed reinduction), and, finally, to induce remissions when the cancer fails to respond to induction or reinduction using drugs not present in the initial protocols (termed rescue).

An unanswered question in the treatment of lymphoma has been whether long-term maintenance chemotherapy is useful following an initial course of aggressive induction chemotherapy lasting 6 months or less. Long-term maintenance chemotherapy has not been shown to be of significant value in humans with most forms of NHL; however, in humans, the initial induction course of chemotherapy is much more aggressive than that used in veterinary patients. Although no randomized prospective studies have been performed to address the therapeutic benefit of long-term maintenance chemotherapy in dogs, most comparisons of dogs treated with CHOP-based protocols do not show any clear advantage for a maintenance or consolidation phase after induction therapy.* Indeed, in most reports, dogs receiving shorter, less costly protocols that do not include a prolonged maintenance phase have comparable remission and progression-free survival (PFS) durations and appear to more readily achieve second remissions when they relapse following completion of chemotherapy than their counterparts receiving long-term maintenance. These data, taken together, suggest that maintenance therapy is not beneficial for most dogs with lymphoma. Until well-designed randomized prospective trials indicate otherwise, the author (DMV) prefers protocols that utilize an aggressive induction without maintenance.

Single-Agent Chemotherapy with Known Activity for Dogs with Lymphoma

The most effective, currently available chemotherapeutic agents for canine lymphoma include doxorubicin, L-asparaginase, vincristine, cyclophosphamide, and prednisone—most of which are represented to one degree or another in most first-line multiagent chemotherapy protocols. Other drugs that have documented activity are often considered second-line agents and include lomustine, vinblastine, actinomycin-D, mitoxantrone, mustargen, chlorambucil, methotrexate, dacarbazine (DTIC), 9-aminocamptothecin, ifosfamide, cytosine arabinoside, and gemcitabine. Of these, cytosine arabinoside,¹⁹⁹ ifosfamide,²⁰⁰ and gemcitabine²⁰¹ appear to have only minimal activity. With the exception of doxorubicin, induction therapy with single-agent chemotherapy does not typically result in durable remission durations when compared with standard combination protocols (see Table 11-2, Chapter 11). Incorporation of other standard cytotoxic drugs with single-agent activity into standard CHOP-based protocols has not resulted in significant gains, and most are reserved for subsequent rescue settings.

Induction in Treatment-Naïve Patients

It is now clearly established that “standard of care” combination protocols used in dogs with lymphoma are essentially variations of “CHOP” protocols (see Table 32-3). Specific details regarding dose and timing of the CHOP protocol currently preferred by the author (DMV) are outlined in Box 32-2. This protocol does not have a maintenance therapy arm, and all treatments cease at 19 weeks, provided the animal is in complete remission. Although several other CHOP-based protocols include L-asparaginase either at initiation or at varying times throughout the protocol, several studies suggest this does not result in clinically relevant increases in remission rate, speed of attaining remission, or first-remission duration, and therefore the author reserves its use for rescue situations.^{172,195,202,203}

If client or other considerations preclude a CHOP-based protocol, single-agent doxorubicin (30 mg/m², intravenous [IV], every 3 weeks for 5 total treatments) is offered along with a 4-week tapering oral prednisone regimen (same prednisone regimen in Box 32-2) as a less aggressive and less costly approach. The expected CR rate will range from 50% to 75%, with an anticipated median survival of 6 to 8 months.^{171,172,204,205} The addition of oral cyclophosphamide (50 mg/m² daily for 3 days starting on the same day as doxorubicin) to single-agent doxorubicin resulted in a numerically but not statistically superior outcome in a recent randomized trial²⁰⁵ comparing doxorubicin/prednisone with doxorubicin/cyclophosphamide/prednisone (PFS of 5.6 months versus 8.2 months, respectively). This trial was only powered to detect a threefold difference in PFS; therefore larger trials should be undertaken to confirm any benefit.

If clients are reticent to include IV medications, the author often recommends a protocol of oral lomustine (CCNU; 70 mg/m² by mouth [PO] every 3 weeks for 5 treatments) and prednisone. This protocol has been associated with short median remissions (40 days) in only one small case series²⁰⁶; however, in the author’s experience, a subset of dogs have remained in remission for several months on this protocol when clients decline IV medication.

If financial or other client concerns preclude the use of systemic chemotherapy, prednisone alone (2 mg/kg PO, daily) will often result in short-lived remissions of approximately 1 to 2 months. In these cases, it is important to educate clients that, should they decide to pursue more aggressive therapy at a later date, dogs receiving single-agent prednisone therapy are more likely to develop multidrug resistance (MDR) and experience shorter remission and survival durations with subsequent combination protocols.^207-209 This is especially true following long-term prednisone use or in dogs that have experienced a recurrence while receiving prednisone. Therefore the earlier that clients opt for more aggressive therapy, the more likely a durable response will result.

A CBC should be performed prior to each chemotherapy treatment. A minimum of 1500 neutrophils/µL (some oncologists use a cut-off of 2000 neutrophils/µL) and 50,000 platelets/µL should be present prior to the administration of myelosuppressive chemotherapy. If the neutrophil count is lower than 1500/µL, it is best to wait 5 to 7 days and repeat the CBC; if the neutrophil count has increased to more than 1500 cells/µL, the drug can be safely administered. A caveat to these restrictions is that for dogs presented prior to initiation of chemotherapy with low neutrophil and platelet counts due to bone marrow effacement, myelosuppressive chemotherapy is instituted in the face of cytopenias in order to clear the bone marrow of neoplastic cells and allow hematopoiesis to normalize.

In those breeds likely to have MDR1 gene mutations (e.g., collies; see Chapter 11) and therefore to be at risk for serious chemotherapeutic toxicity,²¹⁰ the author will initiate a CHOP protocol out of sequence, beginning with non-MDR1–associated drugs, such as cyclophosphamide. This ensures treatment of the lymphoma while allowing sufficient time for analysis of MDR1 gene mutations prior to initiating MDR1 substrate drugs. No specific protocols have been scrutinized for treating dogs that are double-mutant for MDR1; however, if using MDR1 substrate drugs, the author initiates at a 40% dose reduction. Subsequent dose modifications (increased or decreased dosage) can be implemented, depending on adverse event levels observed, particularly neutrophil counts at nadir.

Mechanisms of Drug Resistance

Drug resistance can be inherent in cancer cells or develop following exposure to selected chemotherapeutic agents and often is associated with increased expression of members of the adenosine triphosphate (ATP)-binding cassette (ABC) transporter superfamily (e.g., P-glycoprotein pump), many of which efflux various chemotherapeutic compounds from cells (see Chapter 11).²³⁹ P-glycoprotein is under the control of the MDR1 gene. MDR has been reported in canine lymphoma following exposure to chemotherapy.^156a,225,240 Expression levels of mRNA encoding the canine MDR1 gene have been characterized in canine cell lines and lymphomas. Although expression of MDR1 mRNA correlated with in vitro drug sensitivity, it did not correlate with in vivo doxorubicin sensitivity in dogs with lymphoma in this study. Additionally, quantitative analysis of mRNA for 10 different drug-resistance factors was performed in 23 dogs with lymphoma.²⁴¹ These dogs were divided into drug “sensitive” and “resistant” categories based on response to a CHOP-based protocol; however, significant differences in expression were not observed in this small study.

Altering Drug Pharmacokinetics

Methods of increasing the time that tumor cells are exposed to chemotherapeutics should theoretically enhance tumor killing. These methods could include long-term continuous infusions (impractical in many veterinary situations), increasing the frequency of treatments, or enhancing the circulation time of drugs used. In one study, dogs with lymphoma received lower dose doxorubicin weekly rather than a higher dose every 3 weeks (thereby decreasing C_max, which is associated with cardiotoxicity) in order to potentially increase the time of drug exposure.²⁴² No benefit was noted, and, in fact, remission rates were inferior. Studies evaluating pegylated long-circulating doxorubicin-containing liposome drug delivery systems in dogs with lymphoma have also been performed.^243,244 Although efficacy was established, enhancement of remission or survival durations over equivalent doses of native doxorubicin was not observed.

Monoclonal Antibody Approaches

In the past decade, enhanced durability of first remissions in humans with non-Hodgkin’s B-cell lymphoma has been achieved primarily through the institution of monoclonal antibody (MAb)-based therapies (so-called R-CHOP protocols); the “R” refers to rituximab, a recombinant chimeric murine/human antibody directed against the CD20 antigen, a hydrophobic transmembrane protein located on normal pre-B and mature B lymphocytes. Following binding, rituximab triggers a host cytotoxic immune response against CD20-positive cells. Unfortunately, rituximab does not have therapeutic activity in dogs due to a lack of external recognition of a similar antigen on canine lymphoma cells and the inherent antigenicity of human-derived antibodies in dogs.^245,246 Another immunotherapy approach involved MAb-231, a murine-derived anticanine MAb (IgG2a). It mediates antibody-dependent cellular cytotoxicity (ADDC) and complement-mediated cellular cytotoxicity (CMCC).^247-249 It also prevented outgrowth of canine lymphoma xenografts in nude mice. In a noncontrolled clinical study of 215 dogs treated with CHOP-based chemotherapy and MAb-231, enhanced overall survival was observed; however, the antibody was removed from the commercial market in the mid-1990s without definitive randomized trials being performed. Several laboratories throughout the world are currently working to characterize and develop effective MAb therapies for use in dogs.

Antitumor Vaccine Approaches

Several antitumor vaccine approaches have been applied in dogs with lymphoma. A tumor vaccine extract using killed lymphoma cells combined with Freund’s adjuvant was administered to a small number of dogs after remission induction with combination chemotherapy.²⁵⁰ Prolongation of median survival was noted in the treatment group; however, a subsequent study revealed that prolongation was likely due to the Freund’s adjuvant.²⁵¹ An autologous killed lymphoma tumor cell vaccine has been intralymphatically administered to dogs placed in remission using a combination chemotherapy protocol, and, although modest gains were reported in remission times, no survival advantage was found.²⁵² An exploratory vaccine study targeting telomerase²⁵³ (see Chapter 14, Section D) and one using RNA-loaded CD40-activated B cells²⁵⁴ in dogs with lymphoma have also been conducted. These studies involved small numbers of nonrandomized patients and lacked controlled populations for comparison. In a randomized study of 60 dogs with lymphoma comparing CHOP-based chemotherapy with CHOP-based chemotherapy and a human granulocyte-macrophage colony-stimulating factor (GM-CSF) DNA cationic-lipid complexed autologous whole tumor cell vaccine, a small measure of immunomodulation was documented by delayed-type hypersensitivity; however, significant improvement in clinical outcome was not noted.²⁵⁵ Although little well-supported activity is reported to date with these immunomodulatory approaches, our basic understanding of methodologies is expanding.

1. Curative intent therapy for stage I lymph node and solitary extranodal disease (i.e., nasal, cutaneous, spinal lymphoma).

2. Palliation for local disease (e.g., mandibular lymphadenopathy, rectal lymphoma, mediastinal lymphoma where precaval syndrome is present, localized bone involvement).

3. Total body radiation combined with bone marrow or stem cell transplantation.

4. Whole or staged half-body RT following chemotherapy-induced remissions.

Alimentary Lymphoma

Most dogs with alimentary lymphoma are presented with diffuse involvement of the intestinal tract, and involvement of local lymph nodes and liver is common. Chemotherapy in dogs with diffuse disease has been reported to be unrewarding for the most part^47,268,269; however, more aggressive CHOP-based protocols used extensively for multicentric lymphoma in dogs have resulted in durable remissions in a small subset of cases. Solitary alimentary lymphoma is rare in the dog; however, if the tumor is localized and can be surgically removed, results (with or without follow-up chemotherapy) can be encouraging.

Primary Central Nervous System Lymphoma

CNS lymphoma in dogs usually results from extension of multicentric lymphoma. However, primary CNS lymphoma (PCNSL) has been reported.^103-105 If tumors are localized, local RT should be considered. Few studies have reported the use of chemotherapy. In one study, cytosine arabinoside (Ara-C) at a dosage of 20 mg/m² was given intrathecally; this treatment was combined with systemic chemotherapy and CNS radiation.¹⁰³ Overall, the response rates are low and of short duration (several weeks to months).

Cutaneous Lymphoma

Treatment of cutaneous lymphoma depends on the extent of disease. Solitary lesions may be treated with surgical excision or RT. Fractionated RT (to a total dose of 30 to 45 Gy) has been associated with long-term control.²⁷⁰ Diffuse cutaneous lymphoma is best managed with combination chemotherapy, although the rate and durability of response is generally less than in multicentric lymphoma. The most widely used protocols for epitheliotropic cutaneous T-cell lymphoma include CCNU (60 to 70 mg/m² PO, every 3 weeks) along with prednisone.^271,272 Although response rates approach 80%, median remission is approximately 3 months; occasionally, durable remissions are encountered. The author has added L-asparaginase to this protocol (see Table 32-5), and although anecdotally it appears to improve response, comparative data are not available. Sporadic reports of other therapies for cutaneous lymphoma in small numbers of cases include the use of COAP (cyclophosphamide, vincristine [Oncovin], Ara-C, and prednisone),²⁷³ retinoic acid analogs (e.g., Accutane, etretinate),²⁷⁴ L-asparaginase and pegylated L-asparaginase,²⁷⁵ topical mechlorethamine (Mustargen),²⁷⁶ and recombinant human α-interferon.²⁷⁷ All of these reports involved small numbers of cases and resulted in limited response rates with short durations.

A form of cutaneous lymphocytic infiltration has recently been characterized as an indolent T-cell lymphoma based on clonality.¹³¹ It is associated with slow progression and long-term survival following corticosteroid management; however, it does have the potential to progress to high-grade lymphoma.