Gastrointestinal Lymphoma in 20 Dogs

Gastrointestinal Lymphoma in 20 Dogs A Retrospective Study C. Guillermo Couto, DVM, H. Carolien Rutgers, DVM, MS, Robert G. Sherding, DVM, and Jennifer Rojko, DVM, PhD The records of 20 dogs with histopathologically diagnosed gastrointestinal (GI) lymphoma (LSA) evaluated between 1970 and 1984 were reviewed. Fifteen dogs were considered to have primary GI LSA, while five dogs had GI involvement in association with the multicentric form. Most clinical and laboratory findings were nonspecific, but positive-contrast upper GI radiography was suggestive of GI LSA in all of the dogs evaluated. Nine dogs had extensive lymphocytic-plasmacytic inflammatory infiltrates around the neoplastic foci, resulting in difficulty in obtaining a diagnosis of GI LSA when samples were obtained by endoscopy. (Journal of Veterinary Internal Medicine 1989; 3:73-78) LYMPHOMAS (LSA) of the gastrointestinal (GI) tract account for approximately 5% to 7% of all canine lymphomas, representing the most common extranodal form.1-9 Except for isolated reports, case series of canine GI LSA emphasize the pathologic and epidemiologic aspects of this disea~e.~?~ Our report describes the clinicopathologic features of 20 dogs with GI LSA. Materials and Methods A review of the case records of The Ohio State University Veterinary Teaching Hospital from 1970 through 1984 revealed 20 dogs with histologically confirmed GI LSA. Dogs with neoplastic changes limited to the GI tract or with concurrent extra-gi involvement confined to the abdominal cavity or bone marrow were interpreted as having primary GI LSA." Dogs with involvement of thoracic or peripheral sites were interpreted as having GI LSA as a manifestation of multicentric LSA. Diagnostic evaluation included complete blood count, serum biochemical determinations, and plain abdominal radiographs in all dogs; upper GI positive-contrast From the Department of Veterinary Clinical Sciences (Couto, Rutgers, Sherding) and Veterinary Pathobiology (Rojko), College of Veterinary Medicine, The Ohio State University, Columbus, Ohio. Supported in part by DHHS ROI CA-35747-04. Dr. Rojko is a scholar of the Leukemia Society of America, Inc. Dr. Rutger's present address is Department of Veterinary Clinical Sciences, Small Animal Hospital, The University of Liverpool, Liverpool, England. Reprint requests: C. Guillermo Couto, DVM, Department of Veterinary Clinical Sciences, College of Veterinary Medicine, The Ohio State University, Columbus, OH 432 10. radiographs were obtained in 12 dogs. Abdominal ultrasonography was done in two dogs. The histologic diagnosis of LSA was made from tissues obtained by exploratory laparotomy in 11 dogs, at necropsy in three dogs, by gastroscopic biopsy in three dogs, and by colonoscopic biopsy in three dogs. Histologic classification of lymphoid tumors was done according to the functional system of Lukes and Collins" as modified by Valli et a1.i2 for bovine, feline, and canine lymphomas. Tumors were evaluated by cnteria established by Rappaport and Braylan'' and Jubb and Kennedy.14 When possible, the tunica of origin was determined. Clinical Findings Results Based on the criteria described above, 15 dogs had primary GI LSA, and five dogs had multicentric LSA with GI involvement. Age at presentation ranged from 17 months to 13 years (mean f standard deviation [SD], 6.7 -t 0.7 years). Five of the affected dogs (25%) were of mixed breeding. Breeds in which there were more than one affected dog included Scottish Terrier (n = 3), Poodle (n = 3), Bassett Hound (n = 2), Boxer (n = 2), and German Shepherd (n = 2). Breeds with one affected dog each included Golden Retriever, Doberman Pinscher, and soft-coated Wheaten Terrier. Eighteen dogs (90%) were males ( 13 intact, five castrated), and two (10%) were females (one intact, one spayed), compared with a hospital population from 1979 to 1984 of 48.8% males and 51.2% females. Both females were Bassetts. Body weight ranged from 4 to 35 kg (mean t SD, I8 * 2. I kg). 73

74 COUTO ET AL. Journal of Veterinafy Internal Medicine FIGS. 1A AND 1B. Appearance ofjejunal lymphomas (LSA) at surgery in the dog of Fig. 2; (A) outward appearance and (B) incisional view. Clinical signs included depression (n = 18; 90%), vomiting (n = 17; 85%), anorexia (n = 14; 70%), diarrhea (n = 13; 65%), weight loss (n = 10; 50%), icterus (n = 1; 5%), and tenesmus (n = 1; 5%). Vomiting and diarrhea usually occurred together; however, vomiting alone was noted in five dogs (25%), and diarrhea alone in one (5%). Vomiting was reported in 11 of 12 dogs (92%) with gastric involvement, and in seven of 14 dogs (50%) with intestinal involvement. The vomitus or feces contained visible blood in nine dogs (45%). The duration of clinical signs ranged from 3 days to 13 weeks (mean k SD, 6.1 f 0.8 weeks). Historically, the signs in most dogs gradually increased in frequency and severity, and most responded poorly to symptomatic treatment. Acute exacerbation of clinical signs in two dogs was caused by acute peritonitis resulting from perforation of tumor-associated gastric ulcers. Physical examination was abnormal in 17 dogs (85%). Abnormalities included poor body condition (n = 8; 40%), presence of a midabdominal or cranial abdominal mass (n = 5; 25%), abdominal pain (n = 4; 20%), pyrexia (39 to 4OOC) (n = 4; 20%), and hepatomegaly (n = 2; 10%). Enlarged tonsils, splenomegaly, and diffusely thickened bowel (Fig. 1) were each noted once. Hematologic and Biochemical Findings Hematologic findings were abnormal in 17 dogs (85%). Six dogs (30%) were anemic (packed cell volume range, 19 to 34%; mean _+ SD, 27.2 f 2.7%; reference range, 37 to 52%). With only one exception, anemic dogs were also hypoproteinemic. Hypoproteinemia and anemia were seen mostly in dogs with diffuse GI LSA. The anemia was normochromic and normocytic in four dogs and microcytic and hypochromic in two dogs. The anemia was nonregenerative in four dogs. In two others, mild reticulocytosis was attributed to chronic intestinal blood loss. Morphologic red blood cell abnormalities included target cells (n = 5), fragments (n = 4), and mild-to-moderate anisocytosis and polychromasia (n = 4). Leukocytosis due to neutrophilia occurred in ten dogs (50%) (range, 19,295 to 62,800 cells/pi; mean k SD, 30,340 * 426 cells/pl; reference range, 6000 to 18,000 cells/pl), and a left shift occurred in six dogs. Two of these dogs had acute peritonitis. Neutropenia ( 1390 cells/pl) and absolute lymphocytosis (6300 cells/pl; reference range, 1200 to 5200 cells/pl) were each noted once. Serum total protein and albumin concentrations were decreased in six dogs (total protein: range, 4.0 to 5.7 g/dl; mean k SD, 4.8 -t 0.3 g/dl; reference range, 5.7 to 7.8 g/dl; albumin: range, 1.2 to 2.3 g/dl; mean f SD, 1.8 f 0.2 g/dl; reference range, 2.4 to 3.6 g/dl). Serum biochemical abnormalities suggesting liver disorders were found in ten of 13 (77%) dogs with hepatic LSA and in one dog with nonneoplastic vacuolar hepatopathy. Serum alkaline phosphatase activity was elevated in four dogs (range, 167 to 1842 IU/k, mean f SD, 670 f 345 IU/l; reference range, 0 to 110 IU/l), serum alanine aminotransferase activity was increased in six dogs (range, 132 to 362 IU/l; mean f SD, 184 f 35 IU/l; reference range, 7 to 70 IU/l), and aspartate aminotransferase activity was elevated in four dogs (range, 135 to 263 IU/k mean f SD, 202 f 30 IU/l). Retention of bromsulfonphthalein was increased in the two dogs evaluated (9.5% and 41% retention at 30 minutes, respectively; normal, less than 5% retention). Electrolyte changes compatible with vomiting and diarrhea were seen in eight dogs (40%), and mild azotemia was seen in four dogs (20%). Peak plasma concentration of xylose at 60 and 90 minutes after oral administration (0.5 g/kg) was abnormally low in the two dogs evaluated (i.e., under 60 mg/dl). Feces from eight of 18 dogs (44%) were positive for occult blood. Radiographic Findings Plain abdominal radiography revealed abnormalities in nine of the 20 dogs (45%), including hepatomegaly (n = 3), splenomegaly (n = 3), midabdominal mass (n = 3),

Vol. 3. NO. 2 GASTROINTESTINAL LYMPHOMA IN DOGS 75 free abdominal gas and peritonitis (n = l), and decreased liver size attributed to unrelated chronic liver disease (n = 1). Enlarged adrenal glands and gastric wall thickening were each noted once. Upper GI positive-contrast radiography revealed abnormalities in all 12 dogs examined. Mucosal irregularities, luminal filling defects, and irregular thickening of the wall suggested infiltrative disease of the stomach in six (50%) and small intestine in six (50%) dogs (Fig. 2). Ultrasonography showed hepatic masses with hypoechoic centers in two dogs with abnormal liver tests and hepatic LSA. Clinical Course and Treatment The mean time for diagnosis after admission was 3.5 days. However, three dogs had an atypical presentation leading to diagnostic intervals of 2 to 3 months. Biopsy specimens in two of these dogs initially showed lymphocytic-plasmacytic enteritis (LPE) without LSA. In the two dogs with LPE, contrast radiography revealed infiltrative duodenal disease in one and gastric filling defects in the other. These dogs responded transiently to treatment with glucocorticosteroids, azathioprine,* and metronidazolet but then developed severe vomiting, diarrhea, weight loss, and hypoproteinemia. The first dog died after spontaneous bowel perforation; the second dog was euthanatized. At necropsy, both dogs had GI LSA, which also involved the liver in the second dog. The third dog had an abdominal mass and marked leukocytosis, eosinophilia, and basophilia; surgical biopsy specimens revealed eosinophilic pyogranulomatous gastroenteritis and lymphadenitis. This dog responded well to treatment with mebendazolet for 5 days and glucocorticosteroids for 7 weeks but then developed severe protein-losing enteropathy. A second surgical biopsy specimen revealed intestinal LSA. The dog died postoperatively; necropsy showed LSA involving the intestines, mesenteric lymph nodes, liver, and adrenal glands. Eight of the 20 dogs were treated. Five dogs treated with glucocorticosteroids alone or in combination with azathioprine or vincristinetj either responded only briefly or not at all, and all but one died or were euthanatized within 3 to 14 weeks of diagnosis. Multiple agent chemotherapy with cyclophosphamide, 11 vincristine, L- asparaginase,t and prednisone (CLOP protocol) was used in one dog and cyclophosphamide, vincristine, cytosine arabinoside,# and prednisone (COAP protocol) in two dogs. One of these had an objective response to * Imuran, Burroughs Wellcome, Research Triangle Park, NC. t Flagyl, Searle, Chicago, IL. # Telmintic, Pittman Moore, Washington Crossing, NJ. 0 Oncovin, Eli Lilly, Indianapolis, IN. 11 Cytoxan, Meade Johnson, Evansville, IN. ll Elspar, Merck, Sharp, & Dohme, West Point, PA. ## Cytosar U, Upjohn, Kalamazoo, MI. FIG. 2. Barium contrast radiograph ofa dog with small intestinal LSA, showing irregular thickening and filling defects of the small intestinal wall. therapy but was euthanatized 5 weeks later due to thrombosis of the iliac arteries. One dog was lost to follow-up, and one dog with colorectal LSA is still in remission (5 years after initiation of therapy). Four of the dogs not treated with chemotherapy died, three from postoperative bowel dehiscence. One dog was lost to follow-up; the other seven dogs were euthanatized at the time of diagnosis. Gross and Microscopic Pathology Tumor origin, cell type, and location are summarized in Tables 1 and 2. Fifteen dogs had primary GI LSA based on biopsy (n = 13) or necropsy (n = 2) findings. Tumor type and distribution were confirmed by necropsy in 13 dogs. All the LSAs appeared to originate in the submucosa with six arising in the stomach (40%), seven in the small intestine (47%), and two in the large intestine (1 3%). Most lesions involved diffuse infiltrates of noncleaved cells in the submucosa and lamina propria. Tumor cell size was highly variable. Of the six tumors

76 COUTO ET AL. Journal of Veterinary Internal Medicine TABLE 1. Origin, Type, and Tumor Location in Dogs With Primary Gastrointestinal Lymphomas Tumor Location No. of Dogs with Lymphocytic- Tumor No.of Small Large Lymph Bone Plasmacytic Tumor Origin Type* Dogs Stomach Intestine Intestine Node Liver Marrow K-ldney Inflammation Gastric D, S, NC 2 2 0 0 2 2 1 0 0 submucosa D, L, NC 3 3 1 0 2 3 0 0 2 N, L, NC I 1 1 0 1 1 0 0 0 Subtotal 6 6 2 0 5 6 I 0 2 Intestinal D, S, NC 2 0 2 0 1 2 0 0 1 submucosa D, M. NC 2 0 2 0 1 1 1 0 1 D. M-L, NC 2 1 1 2 I 0 0 0 2 D, L, NC 3 0 3 0 2 1 I 0 2 Subtotal 9 I 8 2 5 4-7 0 6 Total 15 7 10 2 10 10 3 0 8 * Tumor cell type classified according to Valli et al." where D = diffuse, N = nodular, S = small, M = medium, L =large, NC = noncleaved, C = cleaved. arising in the gastric submucosa, all involved the liver, and five invaded local lymph nodes, two the small intestine, and one the bone marrow. Of the nine tumors which originated in the intestinal submucosa, five also involved the local lymph nodes, four the liver, two the bone marrow and large intestine, and one the stomach. While lymphocytic-plasmacytic inflammation occurred in conjunction with tumors of both gastric and intestinal origin, it was found more frequently in association with tumors involving the intestine (Table 1). Specifically, marked-to-severe lymphocytic-plasmacytic inflammation was present adjacent to, or occasionally distant from, the neoplastic foci in eigbt of 15 dogs with primary GI LSA and in one of five dogs with multicentric LSA. The junctional region between neoplastic and nonneoplastic tissue was not sharply demarcated (Fig. 3). Often an inflamed mucosa overlaid a sublslucosal lymphomatous focus (Fig. 3). The other five dogs with clinically diagnosed GI LSA were interpreted as having GI LSA secondary to or in association with multicentric LSA. In these dogs, the mediastinal (n = 4), peripheral nodal (n = I), or pulmo- nary (n = I) lesions were of sufficient bulk to exclude a diagnosis of primary GI LSA alone. Lymphocytic-plasmacytic enteritis was a concurrent finding in only one dog. Discussion Based on the small number of dogs with GI LSA that could be retrieved in this 14-year computer search in our hospital and on the paucity of cases reported in the literature, it appears that GI LSA is less common in dogs than in cats or human^.',^,'^ The high prevalence of this neoplasm in male dogs confirms a previous observation that male dogs are at a higher risk for the development of GI LSA than female dogs' and is similar to observations in humans.i5 Clinical signs and physical findings were mostly nonspecific, although vomiting and/or diarrhea, often bloody, were frequent features of GI LSA. In most dogs, the onset of clinical signs was gradual and insidious, and in some cases, symptomatic treatment and/or corticosteroid therapy resulted in transient improvement. Ap- TABLE 2. Tumor Location in Dogs With Gastrointestinal and Multicentric Lymphomas Tumor Location No. of Dogs with Lymphocytic- Tumor Tumor No. of Small Large Lymph Plasmacytic Origin Type* Dogs Stomach Intestine Intestine Node Mediastinum Other Inflammation Multicentric D, M, NC 4 4 3 1 4 4 I each-kidney, 1 liver, tonsil, pancreas, peripheral nodes D, L, NC 1 0 1 0 1 0 kidney, lung 0 Total 5 4 4 1 5 4 1 * Tumor cell type classified according to Valli et at." where D = diffuse, N = nodular, S = small, M = medium, L = large, NC = noncleaved, C = cleaved.

Vol. 3. NO. 2 GASTROINTESTINAL LYMPHOMA IN DOGS 77 preciable weight loss was detected in half of the dogs. In addition, intraabdominal masses (enlarged mesenteric lymph nodes or lymphomatous masses involving the intestine) were detected by palpation in approximately 25% of the dogs. In two dogs, sudden exacerbation of chronic GI signs resulted from perforation of tumor-associated gastric ulcers. Hematologic abnormalities included anemia (30%), leukocytosis due to neutrophilia (50%), and a left shift (30%). The anemia, which was normocytic, normochromic in four dogs and microcytic, hypochromic in two dogs, was attributed to chronic disease in the first group of dogs and to chronic gastrointestinal blood loss with secondary iron deficiency in the latter. The leukocytosis was attributed primarily to tumor necrosis or to secondary bacterial infection. Liver enzyme elevations occurred in approximately 30% and were attributed to hepatic infiltration with neoplastic lymphoid cells or to prior corticosteroid therapy. Hypoproteinernia, the most frequent serum biochemical abnormality and usually associated with hypoalbuminemia, was observed in three fourths of the dogs for which serum or plasma protein concentrations were determined. In most the lesions were characterized by intestinal involvement with neoplasia and associated inflammation. Therefore, the hypoproteinemia was attributed to protein-losing enteropathy. Radiographic studies were valuable in suggesting a diagnosis of GI LSA. Plain abdominal radiographs revealed abnormalities compatible with neoplasia in approximately half of the dogs. Moreover, upper GI positive-contrast radiography contributed to the diagnosis of GI LSA in all 12 dogs on which this procedure was done. Hepatic ultrasonography in two dogs disclosed the presence of intrahepatic masses with hypoechoic centers; these ultrasonographic changes are similar to those previously reported by Nyland et a1.16 From the pathologic standpoint, difficulty arose when trying to differentiate between GI LSA and LPE. In two dogs with a previous diagnosis of LPE, subsequent biopsy or necropsy specimens revealed the presence of GI LSA. Furthermore, eight of 15 dogs with primary GI LSA and one of five dogs with multicentric LSA had extensive mucosal lymphocytic-plasmacytic inflammation. This suggests that either LPE is a prelymphomatous change in the GI tract or, alternatively, that some plasma-cell-rich areas within a heterogenous lymphomatous infiltration may resemble lesions of LPE. A syndrome of immunoproliferative small intestinal disease characterized by LPE has been described in three Basenji dogs that subsequently developed GI LSA.17 The problem is highlighted by recent reports which indicate that immunohistochemical stains are the only reliable method for differentiating benign (pseudolymphomatous) GI lymphoplasmacytic infiltrations from GI lymphomas in hurnans.l8, With these techniques, lympho- FIG. 3. Nonneoplastic lymphocytes in the gastric mucosa overlying an invasive lymphomatous lesion (H & E, X430). The superficial lymphocytic-plasmacytic infiltrate usually results in a diagnosis of lymphocytic-plasmacytic gastritis (or enteritis) when biopsies are obtained by endoscopy. cytic-plasmacytic infiltrates can be characterized as either polyclonal (pseudolymphomas) or monoclonal A similar approach to the investigation of untreated dogs with LPE may contribute further data in this respect. In some of these dogs the occurrence of neoplastic and lymphocytic-plasmacytic infiltrates in adjacent areas in the same dog emphasizes the fact that the diagnosis of LSA is dependent upon the site selected and depth of the biopsy. Exploratory laparotomies yielded diagnostic samples in all dogs in which the procedure was done. Furthermore, direct visualization and biopsy of the abdominal organs contributed to staging the disease. However, biopsy of GI LSA must be undertaken with care since complete effacement of the bowel architecture by neoplastic cells and concurrent hypoalbuminemia may lead to early postoperative dehiscence of the intestinal suture line. Since most lymphomas originate in the submucosa, endoscopic biopsy may be inadequate for diag-

78 COUTO ET AL. Journal of Veterinary Internal Medicine nosis because most of the specimens thus obtained are usually superficial..12 In many dogs with GI LSA, an inflammatory, nonneoplastic infiltrate (LPE) was the predominant superficial mucosal lesion (Fig. 3). Although results of therapy were discouraging, most dogs in this study were treated with corticosteroids, either singly or in combination with azathioprine or vincristine. Multiple-agent chemotherapy protocols are more effective than single agents or than combinations of two different anticancer drugs in the treatment of lymph~rna.~ Only three dogs were treated with multiple-agent chemotherapy (cyclophosphamide, vincristine, prednisone, and L-asparaginase in one dog; cyclophosphamide, vincristine, prednisone, and cytosine arabinoside in two dogs), and objective responses were observed in each. One dog with colorectal LSA continues to be in complete remission after 54 months of continuous chemotherapy. On the basis of our data several generalizations can be made. First, canine GI LSA usually involves several segments of the GI tract. Second, although clinical signs are nonspecific, the presence of intractable vomiting and/or diarrhea (with or without blood) in a middle-aged or older dog with considerable weight loss and hypoproteinemia should prompt the clinician to evaluate the patient for potential diffuse GI neoplasm. Third, upper GI positive-contrast radiographic studies are valuable in the preliminary diagnosis of GI LSA. Fourth, dogs with GI LSA should also be evaluated for evidence of mediastinal or peripheral nodal LSA. Last, once a presumptive diagnosis of GI LSA is made on the basis of clinical and radiographic findings, a tissue specimen for histopathology should be obtained, preferably through an exploratory laparotomy. References 1. Moulton JE, Dungworth DL. Tumors of the lymphoid and hemopoietic tissues. In: Moulton JE, ed. Tumors in Domestic Animals. Berkeley, CA: University of California, 1978; 150-204. 2. Couto CG. Gastrointestinal neoplasia. Proceedings of the 8th Kal Kan Symposium, 1984; 17-24. 3. Theilen GH, Madewell BR. Leukemia-sarcoma disease complex. In: Theilen GH, Madewell BR, eds. Veterinary Cancer Medicine. Philadelphia: Lea and Febiger, 1979; 204-288. 4. Patnaik AK, Hurvitz AI, Johnson CF. Canine gastrointestinal neoplasms. Vet Pathol 1977; 14:547-555. 5. Feeney DA, Klausner JS, Johnston GR. Chronic bowel obstruction caused by primary intestinal neoplasia: A report of five cases. J Am Anim Hosp Assoc 1982; 18:67-77. 6. Evans SM, De Frate LA. Gastric lymphosarcoma in a dog: A case report. Vet Radio1 1980; 21:55-56. 7. Weiser G, OGrady M. Erythrocyte volume distribution analysis and hematologic changes in dogs with iron deficiency anemia. Vet Pathol 1983; 20:230-241. 8. Holmberg CA, Manning JS, Osburn BI. Canine malignant lymphomas: Comparison of morphologic and immunologic parameters. JNCI 1976; 56:125-135. 9. Priester WA, McKay FW. The occurrence of tumors in domestic animals. Natl Cancer Inst Monogr 1980 541-210. 10. Squire RA, Bush M, Melby EC, et al. Clinical and pathologic study of canine lymphoma: Clinical staging, cell classification, and therapy. JNCI 1973; 51:565-574. 1. Lukes RJ, Collins RD. A functional classification of malignant lymphomas. In: Rebuck, Berard, Abell, eds. The Reticuloendothelial System. Baltimore: Williams and Wilkins, 1975; 2 13-242. 2. Valli VE, McSheny BJ, Dunham BM, et al. Histocytology of lymphoid tumors in the dog, cat and cow. Vet Pathol 1981; 18~494-512. 3. Rappaport H, Braylan RC. Changing concepts in the classification of malignant neoplasms of the hematopoietic system. In: Rebuck, Berard. Abell, eds. The Reticuloendothelial System. Baltimore: Williams and Wilkins, 1975; 1-19. 14. Jubb KV, Kennedy PC. Pathology of Domestic Animals I. New York: Academic, 1970; 386-392. 15. Skudder PA, Schwartz SI. Primary lymphoma of the gastrointestinal tract. Surg Gynecol Obstet 1985; 160:5-8. 6. Nyland TG, Kantrowitz BM. Ultrasound in diagnosis and staging of abdominal neoplasia. In: Gorman NT, ed. Oncology. New York: Churchill Livingstone, 1986; 1-24. 7. Breitschwerdt EB, Waltman C, Hagstad HV, et al. Clinical and epidemiologic characterization of a diarrheal syndrome in Basenji dogs. J Am Vet Med Assoc 1982; 180:9 14-920. 8. Saraga P, Hurlimann J, Ozzello L. Lymphomas and pseudolymphomas of the alimentary tract: An immunohistochemical study with clinicopathologic correlations. Hum Pathol 198 I; 12:713-723. 19. Kahn LB, Mir R. Lymphoid proliferations of the gastrointestinal tract. In: Levin B, Riddell RH, eds. Frontiers in Gastrointestinal Cancer. New York: Elsevier, 1984; 19-39.