A Retrospective Study of Feline Gastric Lymphoma in 16 Chemotherapy-Treated Cats
The purposes of this study were to describe cases of feline gastric lymphoma with regards to signalment, clinical presentation, laboratory and ancillary study findings, response to therapy, and outcomes and to identify prognostic variables. Sixteen cats with stage I and II gastric lymphoma treated with chemotherapy were included in this study. Seventy-five percent of cats experienced remission. Overall, first remission duration was 108 days. Response to treatment was prognostic as in other types of feline lymphoma. Cats with a complete remission (CR) had longer survival times compared with cats with a partial remission (PR). Sex and treatment with a rescue protocol were found to be prognostic with castrated males having longer survivals than spayed females. Cats that received rescue chemotherapy had shorter first remission durations than those that did not. Prior treatment with steroids and stage were not found to be significant prognostic variables. This study characterizes gastric lymphoma treated with chemotherapy in cats. Further studies are needed to determine the comparative efficacy of surgical and chemotherapeutic treatments for feline gastric lymphoma.
Introduction
Lymphoma is the most common neoplasm in cats, representing 30% of feline tumors.1 Lymphoma in cats is associated with a bimodal age distribution, with a peak occurring at < 4 yr of age and another occurring at 8 yr of age.2 Siamese cats and male cats have also been associated with a higher risk of development of lymphoma.3,4 Unlike canine lymphoma, which is typically multicentric involving peripheral lymph nodes, feline lymphoma has a variety of anatomic and histologic presentations. Anatomically, feline lymphoma may be mediastinal, gastrointestinal, nodal, multicentric, or extranodal (with renal and nasal most common). Histological classification involves immunophenotype (B cell or T cell), cell size (large or small), and grade (high, intermediate, or low). Renal and nasal forms are predominantly B cell, while mediastinal and leukemic forms tend to be T cell.5–7 Immunophenotype of gastrointestinal lymphoma is fairly evenly distributed between B cell and T cell, with 54–65% B cell.8–10 A recent study, however, reports a majority of T-cell immunophenotype (83%) in feline gastrointestinal lymphoma using antigen receptor gene rearrangement analysis and immunohistochemistry, suggesting that previous studies may have underestimated the incidence due to difficulty in distinguishing mucosal T-cell lymphoma from lymphoplasmacytic inflammatory bowel disease.11 Small intestinal lymphoma has a slight predominance of T-cell phenotype (52%), whereas large intestinal lymphoma is comprised mainly of B-cell disease (88%).10 Most feline lymphoma is classified as either intermediate- or high-grade.6,12 A large study showed a higher prevalence of low-grade small cell phenotype in gastrointestinal lymphoma compared with a higher prevalence of high-grade tumors in mediastinal lymphoma.12
Treatment of feline gastrointestinal lymphoma is variable with protocols involving either oral prednisolone and chlorambucil or a variety of multiagent chemotherapy protocols. The reported median survival times for cats treated for lymphoma is 2–8 mo.2,5,13–17 Anatomic site plays a role in prognosis of feline lymphoma. In cats achieving a complete remission (CR), nasal lymphoma has a longer survival of 749 days, and central nervous system lymphoma has a very short survival of 70 days.18 Also, mediastinal and extranodal lymphoma has a significantly longer remission duration compared with gastrointestinal lymphoma in one study.17 Within gastrointestinal lymphoma, median survival times vary widely from 2 mo to 24 mo and specific location has not been shown to be prognostic.19–24 Across all studies, the most consistent prognostic factor in feline lymphoma is response to treatment. Survival times for cats with various anatomic forms of lymphoma treated with the University of Wisconsin–Madison chemotherapy protocol are 210 days overall, with a median of 654 days for cats achieving a complete response and 122 days for cats with a partial response.16 In a study directly comparing low-grade and high-grade gastrointestinal lymphoma, low-grade lymphoma, as expected, has a better prognosis than high-grade, with a median survival of 17 mo for cats with low-grade lymphoma compared with a median survival of 2.7 mo for cats with high-grade lymphoma.21 Other prognostic factors classically used in canine lymphoma, such as substage, immunophenotype, and pretreatment with steroids, inconsistently predict outcome in feline lymphoma.5,9,17,18 One study shows that substage is prognostic, with a 9.5 mo survival for substage “a” cats and a 3.5 mo survival for substage “b” cats; however, that classification is less useful in cats with gastrointestinal lymphoma because almost all cats show signs of illness.5 Steroid treatment prior to diagnosis does not correlate with prognosis in most studies; however, in one study, if CR is achieved, prior steroid treatment significantly reduces survival time.2,18
Gastric lymphoma is a relatively uncommon presentation of feline lymphoma. A recent report found 24% of feline gastrointestinal lymphoma cases involved gastric tumors, but only 18% were exclusively in the stomach, with a predominance of large B-cell lymphoblastic lymphoma diagnosed.10 To the authors’ knowledge, there are no previous reports on the prognosis of feline lymphoma confined to the stomach. The purposes of this study are to describe cases of feline gastric lymphoma in regards to signalment, clinical presentation, laboratory and ancillary study findings, response to therapy, and outcomes and to identify prognostic variables.
Materials and Methods
Patient Selection and Evaluation
Medical records of all cats with a histologic or cytologic diagnosis of gastric lymphoma at the Animal Medical Center, New York, New York, between 2003 and 2007 and at the Animal Cancer Care Clinic, Fort Lauderdale, Florida, between 2008 and 2010 were reviewed. A total of 62 cases of feline lymphoma with stomach involvement diagnosed either cytologically or histologically were identified. Cats were excluded if there was evidence of lymphoma in organs other than stomach and intra-abdominal lymph nodes based on ultrasonographic appearance, endoscopic biopsy, or surgical biopsy. Cases in which ultrasonographic appearance suggested involvement of another organ, but histopathology was also performed and did not support involvement of that organ, were included. If ultrasonographic appearance suggested involvement of another organ and either histopathology or cytology of that organ was not performed, that cat was excluded. Twenty-seven cases met the criteria. Cases were also excluded if they did not receive chemotherapy. In total, 16 cats were included in the study that had either cytologically or histologically diagnosed lymphoma confined to the stomach and lymph nodes and were treated with chemotherapy.
Signalment, historical, and physical examination findings were recorded in all cases. Staging tests included complete blood count, serum biochemical analysis, urinalysis, retroviral testing, abdominal ultrasound, thoracic radiographs and, in few cases, either bone marrow aspiration or immunophenotyping. No cats were completely staged with all tests, but all cats had either an abdominal ultrasound or exploratory laparotomy with biopsies. Histologic grade, chemotherapy protocol, remission status, remission duration, and survival time were recorded. The stage of disease was determined based on the system reported in Mooney et al. (1987), used at the Donaldson-Atwood Cancer Clinic.25
Chemotherapy
Chemotherapy protocols used included the following: 25 wk cyclophosphamide, doxorubicin, vincristine, and prednisolone (CHOP); l-CHOP (i.e., the CHOP protocol with l-asparaginase [400 IU/kg subcutaneously] given with or within 1 wk before the first vincristine); cyclophosphamide, vincristine, prednisone (COP); the Animal Medical Center (AMC) protocol; and chlorambucil and prednisone. The chemotherapy agents used in the CHOP protocol included vincristine (0.5–0.7 mg/m2 IV bolus), cyclophosphamide (200 mg/m2 IV bolus), and doxorubicin (1 mg/kg in 0.9% saline IV over 5–10 min). Steroid administration varied with methylprednisolone (4 mg per os [PO] q 12 hr), either prednisolone or prednisone (10 mg PO daily), and either dexamethasone (0.8 mg subcutaneously q 24 hr) or methylprednisolone acetate (20 mg intramuscularly once monthly). The COP protocol alternated vincristine (0.5–0.7 mg/m2 IV) and cyclophosphamide (200 mg/m2 IV). The AMC protocol used vincristine, cyclophosphamide, and doxorubicin as above with long-term maintenance chemotherapy treatment with methotrexate (1 mg/kg IV q fourth treatment) substituting for doxorubicin. In one cat, chlorambucil (2 mg PO q 72 hr) was administered in combination with prednisone (10 mg/cat PO q 24 hr).
Statistical Analysis
Remission duration was defined as the time from the date of beginning treatment to either progression or recurrence of clinical signs related to lymphoma. CR was defined as regression of all clinical signs for at least 30 days. Partial remission (PR) was defined as a > 50% but < 100% resolution of clinical signs for at least 30 days. No response was defined as either a < 50% resolution of clinical signs or a response for < 30 days. Remission status was determined by clinicians based on clinical signs reported by owners and physical exam findings. Survival was defined as the time from the date of histopathologic diagnosis of gastric lymphoma until death from any cause.
Descriptive data were subjected to a univariate analysis, and median values were calculated for each of the variables. To evaluate the prognostic significance of signalment (sex, age, and breed), clinical presentation (weight loss, anemia, inappetence, local versus diffuse disease, and vomiting), and specific treatment variables (clinical stage, tumor grade, chemotherapy protocol, rescue protocol, pretreatment with steroid, and treatment response), a Kaplan-Meier method was performed with standard statistical softwarea. Values of P < 0.05 were considered significant.
Results
Sixteen cats with gastric lymphoma were identified for inclusion in this study based on retrospective search of either surgical biopsy or necropsy reports in medical records. There were a variety of breeds represented, including eight domestic shorthairs, three Oriental breeds, two American shorthairs, one Turkish Van, one Ragdoll, and one Abyssinian. The mean and median ages were 11.4 yr and 12.8 yr (range, 4.6–16.1 yr), respectively. There were nine castrated males and seven spayed females.
Clinical signs have been summarized in Table 1. Vomiting was the most common clinical sign, present in 94% of cats. That was in contrast to previous reports for gastrointestinal lymphoma, having weight loss as the most common presenting clinical sign.2,19,21,23,24 However, decreased appetite, weight loss, and lethargy were also common. Physical examination and laboratory findings were also summarized in Table 1. A majority of cats were in good body condition (50%), and a palpable abdominal mass was only appreciated in five cats (31%). Abdominal pain, as appreciated by the attending clinician at the time of presentation, was present in two cats (13%), one of which was due to septic peritonitis secondary to gastric perforation. Two cats with thickened intestines on physical exam did not have intestinal involvement based on abdominal ultrasound or surgical biopsy, respectively. Anemia was a common laboratory finding, present in eight cats (50%). Feline leukemia virus and feline immunodeficiency virus testing was negative in all nine cats tested.
FeLV, feline leukemia virus; FIV, feline immunodeficiency virus; n, number of cases.
Ancillary studies included abdominal ultrasound, thoracic radiographs, and in a few cases, a bone marrow aspirate. A bone marrow aspirate was performed in only three cats, and in all cases a normal aspirate was obtained. Thoracic radiographs were performed in four cases, and were normal in all cases. Abdominal ultrasonography was used in 12 cases. Abnormalities in the stomach were observed in all 12 cases and varied from thickening to a discernible mass. Eight cats (50%) had a mass detected on ultrasound. One cat had ultrasonographic abnormalities observed in the spleen that were determined to be nodular hyperplasia upon biopsy. The same cat had ultrasonographic abnormalities observed in the liver that were determined to be consistent with cholangiohepatitis on biopsy. Three cats (19%) had enlarged intra-abdominal lymph nodes on abdominal ultrasound. These lymph nodes were not biopsied because diagnosis was obtained through endoscopy or cytology but were attributed to lymphoma. Two additional cats were found to have lymph node involvement on surgical biopsy. Endoscopy of the stomach was performed in eight cats (50%), with masses visualized in the stomach of six of those cases. Surgical biopsies were obtained in six cats (38%), with masses visualized in five cases. Two cases were diagnosed by cytology alone. Overall through either imaging or surgery, 81% of cats were determined to have a stomach mass. Based on those ancillary diagnostics, 11 cats were diagnosed with stage I lymphoma, and 5 cats were diagnosed with stage II lymphoma (Table 2).
Tumors were characterized histologically by grade in 14 cases. High-grade tumors predominated (n = 12), with the others being graded as intermediate (n = 2). Immunophenotype was only performed in two cases, with the result of B-cell lymphoma.
The 16 cats included in the study were treated with combination chemotherapy. Two of those 16 cats underwent surgery involving partial gastrectomy prior to chemotherapy treatment. Combination chemotherapy consisted of the 25 wk CHOP protocol with l-asparaginase (n = 11) or without l-asparaginase (n = 1), the AMC protocol with long-term maintenance chemotherapy (n = 2), the COP protocol (n = 1), or chlorambucil and prednisolone (n = 1). Nine of the 16 cats treated with chemotherapy achieved a CR and three obtained a PR, with an overall response rate of 75%. Eight cats were treated with rescue protocols, including mustargen, vincristine, prednisone, and procarbazine (n = 4); 1-(2-chloroethyl)3-cyclohexyl-1-nitrosurea (n = 2); or l-asparaginase (n = 2). Only two of those eight cats had a PR to treatment (25%). The other six cats did not respond to rescue therapy. Of the 16 cats treated with combination chemotherapy, 6 had received corticosteroids prior to diagnosis, including methylprednisolone (n = 4), methylprednisolone acetate (n = 1), and an unrecorded type of steroid (n = 1). Two of those cats had previously been diagnosed with inflammatory bowel disease for 1 yr and 8 yr.
Median overall clinical remission duration was 108 days. In cats achieving a PR, median remission duration was 125 days. In cats achieving a CR, median remission duration was 189 days. Median overall survival was 171 days. Median survival in cats achieving only a PR was 138 days, significantly reduced compared with those achieving a CR, with a median survival of 431 days (P < 0.001). Cats that did not respond to treatment had a median survival of 33 days (Figure 1).
Citation: Journal of the American Animal Hospital Association 50, 1; 10.5326/JAAHA-MS-5989
In this small population, sex had a significant (P = 0.049) impact on survival time but not remission duration, with castrated males having longer survival times (240 days) than spayed females (109 days) as shown in Figure 2. Age, breed, body weight, presenting clinical signs, and presence of anemia had no significant impact on either remission duration or survival time in this study. Pretreatment with steroids was not found to significantly reduce either remission duration or overall survival (P = 0.13; Figure 3). There was also no difference between remission duration and survival between cats with either stage I or stage II lymphoma. Cats that did not receive rescue chemotherapy had longer first remission durations compared with those that did (P = 0.019; Figure 4). Overall survival, however, was not significantly affected by rescue chemotherapy.
Citation: Journal of the American Animal Hospital Association 50, 1; 10.5326/JAAHA-MS-5989
Citation: Journal of the American Animal Hospital Association 50, 1; 10.5326/JAAHA-MS-5989
Citation: Journal of the American Animal Hospital Association 50, 1; 10.5326/JAAHA-MS-5989
Discussion
Lymphoma in cats is a diverse disease. Although it is a systemic disease, it localizes to a variety of organs in cats. This study sought to describe cases of lymphoma localized to the stomach in cats, in regards to signalment, clinical presentation, laboratory and ancillary study findings, response to therapy, and prognostic variables. The median age of cats with gastric lymphoma was 12.8 yr, similar to previous reports of gastrointestinal lymphoma.2,5,21–24 Vomiting was the most common clinical sign reported in this study, in contrast to most reports of feline gastrointestinal lymphoma in which weight loss was more common.2,19,21,23,24 Lymphoma of the gastrointestinal tract can present as either diffuse infiltration or a discrete mass. In this study of primary gastric lymphoma, 81% of cats had a mass present. Eight cats (50%) had endoscopic biopsies to confirm lymphoma, and endoscopy has previously been shown to be adequate for diagnosis of gastric lymphoma. Involvement of other organs was based on ultrasound examination in the eight cases biopsied endoscopically in this study.26 Therefore, there was the possibility of inclusion of cases with more extensive abdominal involvement of lymphoma in organs that were not biopsied in this study. Previous reports have shown a predominance of high-grade lymphoma arising in the stomach, and that was supported in this study, with 75% identified as high-grade and 12% identified as intermediate-grade.10,21 No low-grade tumors were identified.
Cats in this study received either multiagent chemotherapy (i.e., CHOP, l-CHOP, COP, or the AMC protocol) or oral chemotherapy (chlorambucil and prednisolone). The overall response rate was 75%. The overall median CR duration was 108 days, and overall survival was 171 days. Those remission and survival times of cats with gastric lymphoma were comparable to those in previous reports of high-grade feline lymphoma.16,19–24 Although a chlorambucil and prednisolone protocol is not typically used in the treatment of high-grade lymphoma, the one cat receiving that protocol in this study did so due to owner preference. That cat had a CR to treatment for 86 days and an overall survival time of 103 days. In this study, sex was found to significantly affect overall survival time, with castrated males having a more favorable prognosis than spayed females. Due to small sample size of this study, that finding may not be clinically relevant. None of the cats in this study were intact. In fact, no intact cats were present in the original 62 cats identified with gastrointestinal lymphoma involving the stomach. That was consistent with a recent epidemiologic study on feline intestinal neoplasia that showed a decreased risk in intact animals.27 It was possible that increased age played a role in an increased likelihood of neutered status; however, further investigation of a protective effect in intact felines is warranted. Treatment with a rescue protocol was found to be negatively associated with first remission duration in the cats in this study. Due to small sample size and heterogeneity of treatment, many confounding variables could have affected that finding. However, it is possible the cats that responded well to first-line chemotherapy for a prolonged time period were less likely to be treated with rescue protocols. That may be due to owner satisfaction with initial treatment outcome, financial reasons after prolonged treatment, a sudden clinical decline upon relapse, or death due to another cause. Counterintuitively, the majority of cats treated with a rescue protocol had achieved remission with first-line therapy, with four of eight cats (50%) achieving CR and two of eight cats (25%) achieving PR. Stage and pretreatment with steroids did not significantly affect either remission duration or survival in the cats studied. The authors acknowledge that such an analysis was underpowered, and the lack of a significant difference cannot be ruled out due to the small sample size. Previously, pretreatment with steroids has been shown to reduce survival times in cats obtaining a CR.18 In this study, only two cats achieving a CR received pretreatment steroids, limiting interpretation of the effect in this population. Response to therapy was significantly associated with both remission duration and overall survival in this study, consistent with previous studies.16 Survival times in those cats achieving CRs and PRs in this study were very similar to those reported by Milner et al. (2005), suggesting that stomach anatomic location is not a significant negative prognostic factor in feline lymphoma.16 Only two cats were treated with surgical excision of their gastric lymphoma with incomplete margins in both cases. Therefore, meaningful conclusions about the benefit of surgery in feline gastric lymphoma cannot be drawn from this study. However, in human patients with gastric and intestinal non-Hodgkin’s lymphoma, surgery improves overall survival.28,29 Further prospective studies evaluating surgery in the treatment of feline gastric lymphoma are needed.
Shortcomings of this study included the lack of complete staging, including immunophenotyping, in many of the cats and the reliance of clinicians on clinical signs to signify remission. The small number of cases in this study may hinder the value of the statistical analysis. Due to the limited sample size, a type I error cannot be ruled out for the differences in outcomes associated with sex, response to therapy, and rescue protocol. There was also inconsistency in treatment of the cats in this study, and, due to low numbers receiving individual protocols, direct comparisons could not be made. Those limitations, however, are unfortunately inherent in a retrospective study. Even with those limitations, this study supports that cats with gastric lymphoma have a similar prognosis to cats with other forms of high-grade gastrointestinal lymphoma and that response to treatment is seen in approximately 75% of cats with gastric lymphoma.
This study supports the fact that future studies to investigate the benefit of surgery in treatment of feline gastric lymphoma and the hormonal impact on the risk and response to treatment in feline gastric lymphoma are warranted. Additionally, prospective studies to identify a more favorable chemotherapy protocol, either first-line or rescue, based on remission duration and associated toxicities are very much needed for this disease.
Conclusion
The results of this retrospective study indicate that cats diagnosed with gastric lymphoma can experience survivals comparable with other types of feline lymphoma with chemotherapy treatment. These results show that response to therapy, sex, and rescue chemotherapy treatment had prognostic value. Further comparative prospective studies are needed to evaluate the most effective surgical and chemotherapeutic approach to the treatment of high-grade feline gastric lymphoma.
Overall survival for cats with gastric lymphoma treated with chemotherapy based on response to therapy. Overall survival was significantly different between cats that achieved a complete remission (CR), a partial remission (PR), or no response (NR) based on the log-rank test (P < 0.001).
Overall survival in cats with gastric lymphoma treated with chemotherapy based on sex. Castrated males (M/C) had significantly longer survival times than spayed females (F/S) (P = 0.049).
Overall survival based on pretreatment with steroids in cats with gastric lymphoma treated with chemotherapy. Difference in survival is not significant (P = 0.13).
First remission duration based on rescue chemotherapy protocol in cats with gastric lymphoma treated with chemotherapy. Rescue protocols included mustargen, vincristine, prednisone, and procarbazine (MOPP); 1-(2-chloroethyl)3-cyclohexyl-1-nitrosurea (CCNU); or l-asparaginase (l-aspar) alone. Cats that did not receive rescue chemotherapy had significantly longer remission durations compared with cats that did receive rescue chemotherapy (P = 0.019).
Contributor Notes
T. Gustafson's present affiliation is Flint Animal Cancer Center, Colorado State University, Fort Collins, CO.
B. Taylor's updated credentials since article acceptance are DVM, DACVIM (Oncology).
