Monitoring of large B-cell lymphoma and T-zone lymphoma in a dog via flow cytometry

Abstract

A 12-y-old, castrated male Pomeranian dog was presented because of mandibular lymph node (LN) enlargement. Physical examination and a complete blood count revealed generalized lymphadenopathy and moderate lymphocytosis. Fine-needle aspirate cytology revealed expansion of medium lymphocytes in the right mandibular LN and expansion of large lymphocytes in the left popliteal LN. Flow cytometry identified 2 aberrant lymphocyte populations in both LNs, namely a CD5+CD45− T-cell population, and a large CD21+ B-cell population. Flow cytometry of the peripheral blood revealed an identical population of aberrant CD45− T cells. The patient was diagnosed with concurrent T-zone lymphoma and leukemia, and B-cell lymphoma. Multi-agent chemotherapy was instituted, and serial clinical and flow cytometric analysis revealed complete remission of the neoplastic B cells, but persistence of the neoplastic T cells and persistent lymphadenopathy. This case affirms the diagnostic value of flow cytometry and reveals a unique limitation of the RECIST criteria.

Lymphoma is the most common hematologic neoplasm in dogs and represents 6% of all canine cancers.¹⁰ According to a canine-adapted WHO classification approach, the most frequent nodal subtypes are diffuse large B-cell lymphoma, marginal-zone lymphoma, peripheral T-cell lymphoma, nodal T-zone lymphoma, and T lymphoblastic lymphoma.¹⁵ T-zone lymphoma (TZL), accounts for ~10% of all newly diagnosed canine lymphoma cases and is the most common form of clinically indolent canine lymphoma.^6,15 Affected dogs commonly have slowly developing, peripheral lymphadenopathy (reported in up to 100% of cases) with peripheral lymphocytosis (reported in up to 63% of cases).^{1,3,5,6,11,15} B-cell lymphoma (BCL) represents 62–75% of canine lymphoma.^6,10 The most common form of canine nodal BCL is diffuse large BCL (DLBCL), which represents 93% of canine BCLs.^7,15 Most dogs with DLBCL have rapidly developing peripheral lymphadenopathy.^6,15 We describe here the diagnosis and monitoring of a dog with concurrent BCL and TZL.

A 12-y-old, castrated male Pomeranian was presented to the University of Minnesota Veterinary Medical Center (UMN-VMC; St. Paul, MN, USA) because of a 9-mo history of progressive mandibular lymph node (LN) enlargement. Physical examination revealed enlargement of the right and left mandibular, retropharyngeal, prescapular, popliteal, inguinal, and axillary LNs. The only significant abnormality on a complete blood count (CBC; Advia 2120, Siemens) was mild leukocytosis (14.9 × 10⁹/L; reference interval [RI]: 3.9–14.6) as a result of moderate lymphocytosis (10.5 × 10⁹/L; RI: 0.8–3.4) and a minimal increase in band neutrophils (0.15 × 10⁹/L; RI: 0–0.13). The peripheral blood lymphocytes were medium (1.0–1.5× the diameter of a neutrophil) with a slightly increased volume of basophilic cytoplasm and a round nucleus containing condensed chromatin (Fig. 1A).

Figure 1.

Photomicrographs from the peripheral blood and right mandibular and left popliteal lymph nodes from the affected dog at presentation. A. In the peripheral blood, the morphologically atypical lymphocytes were medium sized with a slightly increased volume of basophilic cytoplasm and round nuclei with condensed chromatin. In the mandibular (B) and popliteal (C) lymph nodes, 2 morphologically aberrant lymphoid populations were identified: 1) large lymphocytes with scant cytoplasm (black arrows), and 2) intermediate lymphocytes (white arrows). Bars = 20 µm.

Fine-needle aspirate (FNA) samples from the right mandibular and left popliteal LNs were collected, and slides were stained with an aqueous Romanowsky stain (7120 Aerospray hematology slide stainer cytocentrifuge; Wescor). Evaluation of the samples from the right mandibular LN revealed large numbers of a minimally mixed lymphoid population (Fig. 1B). A 100-nucleated cell differential revealed 84% medium lymphocytes (1.0–1.5× the diameter of a neutrophil), 14% small lymphocytes (< 1.0× the diameter of a neutrophil), and 2% large lymphocytes (> 1.5× the diameter of a neutrophil). The medium lymphocytes contained a small amount of lightly basophilic cytoplasm and an eccentrically positioned, round or cleft nucleus with variably smooth chromatin. Occasionally, these cells had a prominent cytoplasmic projection. The large lymphoid cells contained a small amount of deeply basophilic cytoplasm and a round-to-reniform nucleus with stippled chromatin and 1–3 nucleoli. The small lymphocytes were unremarkable. In contrast, the left popliteal LN (Fig. 1C) contained 82% large lymphocytes, 11% medium lymphocytes, and 7% small lymphocytes. The large and medium lymphocytes were morphologically similar to those described in the right mandibular LN. Numerous cytoplasmic fragments, frequent ruptured cells, and scattered tingible body macrophages were noted from both sites. Based upon the morphologically atypical, medium lymphocytes, the right mandibular LN was interpreted as possible lymphoma. The left popliteal LN was interpreted as lymphoma based upon the expanded population of large lymphocytes. To provide additional diagnostic certainty and immunophenotype, flow cytometry was performed.

Aspirates from the left popliteal and right mandibular LNs and an additional sample of EDTA-anticoagulated whole blood were submitted for flow cytometry.¹¹ Analysis of the right mandibular LN identified a phenotypically aberrant population of small-to-medium CD5+CD45−CD4−CD8− T cells (94% of all viable events) and a minority population (1% of all viable events) of large CD21+ B cells (Fig. 2A). Flow cytometry of the left popliteal LN revealed populations of cells that were phenotypically identical, but proportionally different, to those in the right mandibular LN. This was reflected by 44% CD5+CD45−CD4−CD8− T cells and 49% large CD21+ B cells (Fig. 2B). Flow cytometry of the peripheral blood revealed phenotypically identical populations of CD5+CD45−CD4−CD8− T cells (89% of all viable events) and large CD21+ B cells (0.4% of all viable events; Fig. 2C). Based upon the cytologic and flow cytometric findings, the patient was diagnosed with: 1) TZL involving the right mandibular and left popliteal LNs and peripheral blood, and 2) BCL involving the left popliteal LN, with probable involvement of the right mandibular LN and peripheral blood. To confirm these findings, PCR for antigen receptor rearrangement (PARR) analysis was performed on right mandibular and left popliteal LN aspirates and revealed clonal T- and B-cell populations in both sites. Lymphadenectomy was suggested, but not performed.

Figure 2.

Flow cytometry scatterplots from aspirates of: A. The right mandibular lymph node (LN). B. Left popliteal LN. C. Peripheral blood from the affected dog at presentation. In both LNs and peripheral blood, a phenotypically aberrant population of CD5+CD45− T cells (green) is identified. These cells are small by forward scatter, demonstrate low side scatter, and are CD21 negative. Also identified in all 3 samples are variable numbers of CD21+ B cells (blue) that, when examined according to their light scatter properties, segregate into 2 discrete populations; one population of CD21+ cells has low forward and side scatter (black arrows) and a second population of CD21+ cells has high forward scatter and intermediate side scatter (gray arrows).

The affected dog was treated with a multi-agent chemotherapeutic protocol consisting of vincristine, cyclophosphamide, doxorubicin, and prednisone. To monitor clinical response, the longest diameter of affected LNs was recorded during treatment in accordance with Response Evaluation Criteria for Solid Tumors (RECIST) guidelines (https://recist.eortc.org/). At the initiation of treatment, the longest diameters of the right mandibular and left popliteal LNs were 36.5 mm and 15 mm, respectively (Suppl. Fig. 1A, 1B).¹³ At the eighth chemotherapy session, halfway through the protocol and 4 mo after diagnosis, the right mandibular LN had not decreased in size and was classified as stable disease (Suppl. Fig. 1A). In contrast, at the same time, the left popliteal LN had decreased in size and lacked measurable disease based on physical examination (Suppl. Fig. 1B). These clinical observations were substantiated by repeat flow cytometry.

Within the stable right mandibular LN, repeat flow cytometry revealed stable proportions of both the CD5+CD45− T cell (94% at diagnosis vs. 97% at recheck) and CD21+ B cell (< 1% at both diagnosis and recheck) populations (Fig. 3A). Also paralleling its clinical status, repeat flow cytometric analysis of the left popliteal LN revealed a marked decrease in the proportion of large CD21+ B cells from 49% at diagnosis to < 1% at recheck (Fig. 3B).

Figure 3.

Flow cytometric scatterplots from aspirates of the: A. Right mandibular lymph node (LN). B. Left popliteal LN from the affected dog at the midpoint of chemotherapy. In both LNs, a phenotypically aberrant population of CD5+CD45− T cells (green) is identified. These cells, which are flow cytometrically identical to cells observed at the time of diagnosis, are small by forward scatter, demonstrate low side scatter, and are CD21 negative. In contrast to the pretreatment samples, very few CD21+ B cells (blue) are identified (black arrows).

In total, the recheck findings were interpreted as a lack of response of the TZL and complete response of the BCL to chemotherapy. The peripheral blood lymphocytosis persisted during treatment, but flow cytometry was not repeated (Suppl. Fig. 1C). Eighteen months after initial diagnosis, the dog was euthanized following a cardiac event and acute onset of azotemia of unknown etiology. The BCL relapsed 11 mo after diagnosis and, despite rescue chemotherapy, was not in remission at the time of death based on physical examination findings. An autopsy was not performed.

Although lymphoma is the most common hematologic neoplasm in dogs and a common neoplasia overall, the presence of 2 forms of lymphoma concurrently in 1 patient is uncommon.^2,4,12,15 Histology is traditionally considered the gold-standard tool for the diagnosis and subtyping of lymphoma, but the affected pet in our case was evaluated using FNA cytology and flow cytometry. This diagnostic approach may not always yield a defined WHO subtype; however, its ease, low invasiveness, increasing availability, and provision of phenotype-specific survival data increasingly promotes it as the future de facto clinical approach.

The initial FNA cytology from the right mandibular and left popliteal LNs revealed 2 morphologically distinct populations of lymphocytes, which was suggestive of 2 neoplastic lymphoid populations. In the right mandibular LN, there was a dominant population of medium lymphocytes with eccentrically positioned nuclei and prominent cytoplasmic projections (“hand-mirror cells”) that was, based on previous publications, suggestive of TZL.^1,9,10,15 In contrast, the left popliteal LN contain an expanded population of medium-to-large lymphocytes that contained minimal visible cytoplasm and large nuclei, which are features associated with BCL.¹⁰ Although cytology is widely accepted to be an accurate tool for the diagnosis of canine lymphoma, it is important to note that there can be significant cytomorphologic overlap of specific subtypes.^7,9 Determining the subtype of lymphoma can be critical in determining the best course of treatment to optimize prognosis. In cases of TZL, chemotherapy was associated with shorter median survival time by ~ 50%.⁶ In contrast, BCL responds well to chemotherapy, with median survival times of dogs increasing significantly with therapy.⁶

Flow cytometry is increasingly utilized in veterinary medicine given its rapid resulting, less invasive sampling requirements, prognostic capacity, and reduced costs (compared to LN surgical extirpation and histopathology). Using flow cytometry, lymphoid neoplasia is diagnosed through the identification of either an expanded population of phenotypically appropriate cells or the identification of a population of phenotypically aberrant cells. In our case, the right mandibular and left popliteal LNs contained a population of phenotypically aberrant medium CD5+ T cells that lacked expression of the pan-leukocyte antigen CD45 and either CD4 or CD8 (94% and 44% of viable events, respectively). These findings were consistent with TZL.¹¹ In addition, both LNs contained a distinct, although not numerically expanded, population of CD21+ B cells (1% and 49% of viable events, respectively). Although these cells did not demonstrate phenotypic aberrancy, a diagnosis of BCL was based upon the abnormally large size of the B cells. In our laboratory, in alignment with previous canine flow cytometry studies, B-cell size is defined according to the relationship of their median forward light scatter signal (FSC) to the median FSC of the same patient’s residual T cells.¹⁶ In our case, the B cells from the popliteal and mandibular LNs were defined as large given their ratios of 1.7 and 1.8×, respectively, exceeding the published cutoff of 1.6.^8,16

Although one of the most compelling aspects of our case is the presence of 2 lymphoid malignancies in a single patient, given the indolent course of TZL and the expanding accessibility of flow cytometry (which is uniquely positioned to diagnose indolent lymphoid neoplasia), this phenomenon is likely to become more frequent. Studies have found that 10% of dogs with TZL develop or are presented with a second malignancy of various types, including recent reports of dogs with TZL and a second lymphoid neoplasia.^2–5,12 The cause of secondary neoplasms is unknown; however, hypotheses include genetic predisposition or chemotherapy. Our report also prompts the question: what is the most appropriate systemic chemotherapy protocol for such patients? Although a CHOP-based protocol is intuitive for a dog with high-grade BCL, such a protocol may be contraindicated in dogs with TZL. According to one report, dogs with TZL experienced worse clinical outcomes with systemic chemotherapy (including protocols incorporating doxorubicin) versus dogs that were given no treatment.¹⁴ Although there is an insufficient number of similar cases, the increasing sensitivity of lymphoma diagnosis will likely expand this population and, as our case demonstrates, a unique shortcoming of the RECIST monitoring approach may prompt prospective analysis of the impact of systemic chemotherapy in dogs with 2 biologically unique lymphoid neoplasms.

Our case highlights the shortcomings of the RECIST criteria in a unique clinical scenario. In lymphoma, the RECIST criteria document response to treatment in cases in which peripheral lymphadenopathy is a primary component of disease presentation. The RECIST guidelines are based on finding measurable disease, namely a lesion (including LNs) > 10 mm in at least one dimension that can serve as the basis for evaluating response to therapy. Responses to therapy are defined as: 1) complete response (CR), in which the lesion is considered non-pathologic by the evaluator; 2) partial response (PR), with > 30% decrease in longest diameter; 3) progressive disease (PD), with > 20% increase in longest diameter; or 4) stable disease (SD), with neither a 30% decrease or 20% increase in longest diameter. RECIST criteria use the change in peripheral LN size to evaluate treatment efficacy,¹³ and are not specialized per subclassification of lymphoma nor used as a direct prognostic factor. At the midpoint of treatment, our affected dog’s right mandibular LN was classified as SD. This clinical assessment was supported by the repeat flow cytometric study, which revealed near-identical percentages of neoplastic T cells and B cells. This lack of clinical and flow cytometric response to chemotherapy is consistent with the low-grade nature of canine TZL and exemplifies a situation in which the RECIST criteria accurately assess the disease present. Simultaneously, the affected dog’s left popliteal LN was classified as CR. In contrast to the right mandibular LN, in which there was alignment of the clinical and flow cytometric findings, repeat flow cytometry of the left popliteal LN revealed a marked decrease in the percentage of neoplastic B cells (from 49% to < 1%), but a notable increase in the percentage of neoplastic T cells (from 44% to 86%). This disagreement highlights the shortcomings of the RECIST criteria in monitoring patients with combined chemosensitive and chemoresistant forms of lymphoma and illustrates the need for additional monitoring modalities to assess patient response to treatment more accurately.