The prevalence of gastrointestinal tumors in dogs is low and has been reported as 3% of all neoplasms in that species.1 Leiomyosarcomas are reported to be the most common sarcoma of the canine intestinal tract and comprise approximately 10% to 30% of all intestinal tumors in dogs.2–6 Recently, the advent of advanced immunohistochemical staining techniques has led to reclassification of GILMSs in human medicine. It has been recognized that tumors previously categorized as GILMSs are not composed entirely of smooth muscle but are principally derived from ICCs.7–9 Interstitial cells of Cajal are the pacemaker cells of the gastrointestinal tract and regulate intestinal motility and peristalsis. The ICC-derived tumors have been renamed GISTs. Gastrointestinal stromal tumors are immunohistochemically distinguishable from GILMSs or true smooth muscle tumors on the basis of expression of c-kit (CD-117). Leiomyosarcomas do not express the c-kit protein. The c-kit protein is a transmembrane receptor with a tyrosine kinase component. Most GISTs have activating c-kit mutations that are thought to be the driving force in GIST oncogenesis.10–16 Gastrointestinal stromal tumors have not only been identified in humans but have also been recently detected in equids, nonhuman primates, and dogs.17–22 Some canine mast cell tumors also express c-kit mutations.23,24 The purpose of the study reported here was to reexamine (via immunohistochemical techniques) canine tissue samples that had been previously classified as GILMSs, identify and differentiate GISTs from GILMSs, and compare the biological behavior and clinical course of GISTs and GILMSs in dogs. To the authors' knowledge, no reports of naturally occurring GISTs in a hospital population of client-owned dogs have been published previously.
Criteria for Selection of Cases
Medical records of dogs with histologically confirmed GILMSs that were evaluated at the Matthew J. Ryan Veterinary Hospital of the University of Pennsylvania between January 1990 and June 2003 were reviewed. Dogs were included in the study if a GILMS (located from stomach to rectum) had been detected and a complete medical record was available. Dogs were excluded from the study if the tumor originated outside the gastrointestinal tract or the medical record was incomplete.
Procedures
Details of signalment, clinical signs, physical examination findings, results of initial diagnostic tests, surgical findings, adjunctive treatment, and outcome after surgery were recorded. Surgery reports were reviewed to determine the location of the tumor and whether there was evidence of macroscopically visible metastatic disease. Reports of pathologic findings were reviewed to evaluate completeness of tumor resection.
Sections of formalin-fixed, paraffin-embedded biopsy specimens that had previously been classified (on the basis of their histologic appearance) as GILMSs were reevaluated by 1 pathologist (MHG). Immunohistochemical staining involving antibodies against c-kit (CD-117), SMA, desmin, S100 protein, and vimentin was performed on each specimen. Specimens that yielded positive results after staining for CD-117 were considered GISTs. Specimens that yielded negative results after staining for CD-117 and positive results after staining for SMA or desmin (or both) were considered GILMSs. Tumors that did not react with stains for CD-117, SMA, desmin, or S100 protein were classified as undifferentiated sarcomas. Staining for vimentin was used to differentiate mesenchymal (GIST, GILMS, and undifferentiated sarcomas) from epithelial tumors. Canine tissue sections known to express the relevant antigens were used as positive control specimens; for c-kit detection, a canine mast cell tumor was used, and for SMA detection, vascular and intestinal smooth muscle tissues were used. Intestinal epithelial cells were used as negative control specimens.
The number of dogs that received chemotherapy was recorded as well as the chemotherapeutic drug used and number of doses administered. Outcome after surgery was determined from medical records and telephone interviews with owners and referring veterinarians. Survival time was calculated from the time of surgery to the date of death or last follow-up contact.
To determine differences between GISTs and other tumors with regard to tumor location, sex, metastasis, recurrence, or whether an affected dog was discharged alive from hospital, the χ2 test or Fisher exact test was used. To estimate survival for the different histopathologic groups (eg, GISTs and GILMSs), the Kaplan-Meier product limit method was used. Difference in survival between histopathologic groups was assessed statistically by use of the log rank test. Dogs were not censored because the cause of death was not always known. Results are presented as means with ranges for continuous data and frequencies or as percentages for categoric data. All analyses were performed by use of statistical software.a A value of P < 0.05 was considered significant.
Results
Biopsy specimens from 42 dogs with a previously confirmed histopathologic diagnosis of GILMS were included in the study. Specimens were representative of 1 tumor/dog. Follow-up information was available for 26 dogs. Mean age at time of diagnosis was 10.4 years (range, 5 to 15 years). Of the 42 dogs, 23 were male and 19 were female. Breeds included mixed breed (n = 19), Golden Retriever (6), Labrador Retriever (5), Basset Hound (2), Cocker Spaniel (2), Miniature Poodle (2), Toy Poodle (2), West Highland White Terrier (1), Maltese (1), Rhodesian Ridgeback (1), and Boston Terrier (1).
Clinical signs at initial evaluation were nonspecific and included lethargy (n = 19 dogs), vomiting (19), inappetence (12), melena (12), hematochezia (7), weight loss (7), and hematemesis (3). Complete blood counts and serum biochemical analyses were performed for all dogs. Abnormalities identified via CBC included anemia (n = 28 dogs), leukocytosis (20), neutrophilia (18), and thrombocytopenia (8). Abnormalities identified via serum biochemical analyses included hypoalbuminemia (n = 7 dogs), high liver enzyme activities (5), azotemia (4), and hypoglycemia (3). The 3 dogs that were hypoglycemic all had GISTs; hypoglycemia resolved in all dogs after removal of the tumor.
Thoracic radiography was performed in 37 dogs before surgery and revealed no gross evidence of metastasis in any dog. Thirty-seven dogs underwent abdominal ultrasonography before surgery, and a mass in the gastrointestinal tract was detected in all dogs. Ultrasonographically, regional metastasis was suspected in 3 dogs on the basis of anechoic to hypoechoic lesions in the liver and spleen, although none of these lesions were evaluated cytologically or histologically (via examination of fine-needle aspirate or biopsy specimens) before surgery.
The expression of c-kit (CD-117) was used to distinguish GISTs from GILMSs. On the basis of immunohistochemical results, 28 of 42 tumors were reclassified as GISTs (Figure 1). Among these 28 GISTs, all also yielded positive results after staining for vimentin and 11 yielded positive results after staining for SMA; none of the GISTs reacted positively with stains for desmin or S100 protein. By definition, GILMSs yield negative results after staining for c-kit, and hence, 10 of the 42 tumors remained classified as GILMSs. Among the 10 GILMSs, all yielded positive results after staining for SMA and vimentin and 3 yielded positive results after staining for desmin; none of the GILMSs reacted positively with stain for S100 protein (Figure 2). Four of the 42 tumors yielded positive results after staining for vimentin, but did not react with any of the other immunohistochemical stains. These 4 tumors were reclassified as undifferentiated sarcomas.
Photomicrograph of a section of a GIST removed from a dog. The section underwent immunohistochemical staining for c-kit (CD117). Unlike the mucosa and muscularis, which are not stained, the underlying neoplastic mass (consisting of interwoven bundles of fusiform cells) has reacted with the stain. In the circular muscular layer, scattered c-kit–positive ICCs are evident. Bar = 100 μm.
Citation: Journal of the American Veterinary Medical Association 230, 9; 10.2460/javma.230.9.1329
Anatomic location of the 28 GISTs included cecum (n = 16), duodenum (4), jejunum (4), colon (2), and stomach (2). The GILMSs were located in the jejunum (n = 5), stomach (3), ileum (1), and colon (1). The undifferentiated sarcomas were located in the duodenum (n = 2), jejunum (1), and rectum (1). In dogs, GISTs developed significantly (P = 0.01) more frequently in the cecum and large intestine than did GILMSs, and GILMSs developed significantly (P = 0.01) more frequently in the stomach and small intestine than did GISTs.
Exploratory laparotomy and complete resection of the tumor were attempted in 40 dogs. Biopsy samples were collected from only 2 dogs because those tumors were incidental findings during exploratory surgery. Four GISTs, all of which were located in the cecum, had perforated the cecal wall, resulting in septic peritonitis. Two GISTs and 1 undifferentiated sarcoma were associated with macroscopic evidence of metastasis at the time of surgery, which was subsequently confirmed histologically. The metastatic GISTs were detected in the liver in 1 dog and as a mass in the mesentery of another dog; the undifferentiated sarcoma had metastasized to the liver and omentum. Among the 40 dogs in which resection of the tumor was attempted, findings of histologic examination of the margins of the resected tumor were reported in the records of 27. Tumors in 22 (81%) dogs were considered to be resected with clean margins. Of the 5 tumors resected without a clean margin, 2 were GISTs, 2 were undifferentiated sarcomas, and 1 was a GILMS.
None of the dogs were treated with adjuvant chemotherapy or radiation therapy after the initial diagnosis was made. However, 2 dogs were treated with adjuvant chemotherapy following tumor recurrence. One dog with recurrence of GILMS at the mesenteric root received 4 doses of doxorubicin hydrochloride,b 2 doses of cisplatinc administered IV, and 1 dose of cisplatin administered intralesionally prior to euthanasia. The second dog received chemotherapy for recurrence of a gastric undifferentiated sarcoma. This dog received 4 doses of doxorubicin and was euthanized after the fourth dose because of persistent anorexia. Postmortem examinations were not performed on either of these dogs.
Four dogs, each with a GIST, died during the perioperative period as a result of tumor-associated perforation of the gastrointestinal wall and sepsis. Median survival time among the 26 dogs for which follow-up information was available (including those dogs that died during the perioperative period) was 11.6 months (range, 0 to 96 months) for dogs with GISTs (n = 17), 7.8 months (range, 1 to 16 months) for dogs with GILMSs (5), and 2.9 months (range, 1 to 5 months) for dogs with undifferentiated sarcomas (4). There was no significant (P = 0.50) difference in survival time among the 3 tumor groups when all dogs were considered (Figure 3). However, when data from only dogs that survived the perioperative period (n = 22) were examined, median survival time was 37.4 months for dogs with GISTs, 7.8 months for dogs with GILMSs, and 2.9 months for dogs with undifferentiated sarcomas (Figure 4). In this latter analysis, dogs with GISTs appeared to survive longer than dogs with GILMSs or undifferentiated sarcomas, but those differences were not significant (P = 0.08).
Photomicrograph of a section of intestinal wall (a positive control specimen) removed from a dog in which GILMS had been previously diagnosed. The section underwent immunohistochemical staining for SMA. The muscularis mucosa and vascular smooth muscle are stained, as is the underlying neoplastic mass that consists of interwoven bundles of fusiform cells. Bar = 100 μm.
Citation: Journal of the American Veterinary Medical Association 230, 9; 10.2460/javma.230.9.1329
Kaplan-Meier curve depicting survival times in 26 dogs for which follow-up information was available after treatment for GISTs (n = 17 dogs; dotted line), GILMSs (5 dogs; solid line), and undifferentiated sarcomas (4 dogs; dashed line). Four dogs with GISTs died during the perioperative period.
Citation: Journal of the American Veterinary Medical Association 230, 9; 10.2460/javma.230.9.1329
Kaplan-Meier curves depicting survival times in 22 dogs that were discharged alive from the hospital following treatment for GISTs (n = 13 dogs; dotted line), GILMSs (5 dogs; solid line), or undifferentiated sarcomas (4 dogs; dashed line).
Citation: Journal of the American Veterinary Medical Association 230, 9; 10.2460/javma.230.9.1329
Discussion
Results of the present study have identified GISTs in a hospital population of dogs. In our hospital, these tumors were previously classified as GILMSs. As applied in human medicine, immunohistochemical staining techniques should be used to more accurately classify mesenchymal tumors of the gastrointestinal tract in dogs. A previous study20 revealed the presence of GISTs in dogs undergoing necropsy at a military facility. However, in that study, all malignant gastrointestinal mesenchymal tumors were reclassified as GISTs, irrespective of results of staining for CD-117. In 2 other studies21,22 in dogs, mesenchymal tumors were reclassified as GISTs on the basis of positive results of staining for CD-117, but data regarding clinical treatment or outcome were not reported. In our study, 28 of 42 GILMSs were reclassified as GISTs on the basis of immunohistochemical findings.
The most common clinical signs among dogs of the present study were nonspecific. Inappetence, weight loss, and lethargy were frequently detected in dogs with both GISTs and GILMSs. Melena, hematemesis, and hematochezia were also commonly associated with both tumor types. Perforation of the gastrointestinal wall subsequent to tumor invasion occurred only in the cecum and only in 4 dogs with GISTs. None of the study dogs with GILMSs had a perforation of their tumor through the gastrointestinal wall. It is possible that compared with GILMSs, GISTs are more locally invasive and result in perforation of the gastrointestinal wall in a greater proportion of dogs. However, the number of cases in the present study was too small to confirm this hypothesis. Another explanation may be that the cecal wall (the site of GIST-associated gastrointestinal wall perforation) is more susceptible to perforation than other areas of the gastrointestinal tract.
In the veterinary literature there is confusion with regard to the nomenclature of GISTs. In a previous report,23 the term GIST was used to describe all gastrointestinal tumors of mesenchymal origin, including GILMSs. In humans, GISTs and GILMSs are separate entities and are distinguished on the basis of results of staining for CD-117. In our study, the immunohistochemical staining pattern was modeled after that associated with human GISTs.9 The only tumors classified as GISTs were those that yielded positive results after staining for CD-117; all those that did not react with that stain were classified as either GILMSs or undifferentiated sarcomas. Coexpression of c-kit and SMA is detected in 20% to 30% of human GISTs.9 Similarly, in the dogs of this report, 40% of GISTs yielded positive results after staining for SMA. We speculate that the overlap of smooth muscle staining results is related to the distribution of ICCs in the gastrointestinal tract. Interstitial cells of Cajal can be associated with Auerbach's plexus as well as interspersed between the intestinal circular muscle layer and submucosa.24 It is possible that the ICCs located along the muscular layer express SMA. Another possibility is that GISTs originate from the primitive stem cells that can differentiate into either ICCs or smooth muscle cells, enabling them to express both CD-117 and SMA.9 Positive results of staining for S100 protein are typically associated with neuronal tumors such as schwannomas. The GISTs in the present study yielded negative results after staining for S100 protein, which confirms their nonneuronal nature and helps to differentiate them from neuronal tumors.9 All tumors in the present study yielded positive results after staining for vimentin. This immunohistochemical stain was used as a positive control for mesenchymal tissue; thus, all tumors were of mesenchymal origin. Four of the 42 tumors did not react with any of the stains, and consequently, they were reclassified as undifferentiated sarcomas. We speculate that this may be a result of suboptimal antigen preservation or loss of antigen expression with malignant transformation.
In dogs, GISTs appear to have a different biological behavior than GILMSs. In the study of this report, GISTs developed most commonly in the cecum and large intestine (18/28 dogs), whereas fewer were found in the stomach and small intestine (10/28). The GILMSs were more commonly located in the stomach and small intestine (9/10 dogs) and less commonly located in the large intestine (1/10). In earlier literature,1 GILMSs were reported to be most common in the jejunum and cecum of dogs. On the basis of our study findings, it seems likely that many tumors located in the cecum that were classified as GILMSs were actually GISTs. It is possible that GISTs are the most common tumor of the cecum in dogs, although specific analysis of a cohort of dogs with cecal tumors remains to be performed. There appears to be a difference in the site predilection of GISTs between dogs and humans; in humans, GISTs are commonly located in the stomach and rarely located in the colon.9,20
In our study, only 3 of the 42 dogs had confirmed metastatic disease: 2 had GISTs, and 1 had an undifferentiated sarcoma. None of those dogs were treated with chemotherapy. Two dogs had recurrence of the tumor: 1 had a GILMS, and 1 had an undifferentiated sarcoma. Both of those dogs were treated with chemotherapy, but were euthanized during or shortly after the completion of their treatment protocol because of persistent anorexia in 1 dog and unknown reasons in the other dog. In humans, conventional chemotherapy and radiation therapy have been largely unsuccessful in the control of GISTs.25 Treatment options for metastatic and unresectable GISTs in humans include oral administration of tyrosine kinase inhibitors. One such tyrosine kinase inhibitor is imatinib mesylate.d Administration of this drug to humans with metastatic or unresectable GISTs has resulted in prolonged survival times, compared with survival times associated with other treatments.26 Imatinib mesylate has been used in dogs, but the treatment, even at low doses, has been associated with hepatotoxicosis.27,28 Other tyrosine kinase inhibitors are currently being evaluated for use in animals with c-kit–expressing tumors such as mast cell tumors.29 Studies should be undertaken to evaluate these treatments in animals with metastatic or unresectable GISTs.
In the present study, treatment of this cohort of dogs was relatively uniform because only 2 of the 42 dogs were treated with chemotherapy. This allowed reliable assessment of the outcome of surgical management without the effect of major bias attributable to the use of adjuvant treatments. With surgery alone, median survival time for all dogs with GISTs was 11.6 months, which was not significantly different from the survival time for dogs with GILMSs. However, if the 4 dogs that did not survive the perioperative period are excluded, the median survival time increases to 37.4 months. For dogs with GILMSs in our study, the median survival time was 7.8 months, and none of these dogs died during the perioperative period. Follow-up information for all dogs was not available, and a significant difference in survival time between these 2 groups might have been detected with greater case numbers. It is also likely that in a study population of dogs with a mean age of 10.4 years at initial evaluation, many of the dogs with GISTs might have died from causes other than that tumor, given that the median survival time of these dogs exceeded 3 years. Other studies3,4,30–32 of dogs with gastrointestinal mesenchymal tumors have revealed median survival times of as long as 12 months; however, in the reports of those studies, no distinction between GISTs and GILMSs was made. Limitations of our study include a small number of cases, lack of follow-up information for some dogs, and lack of necropsy findings or known causes of death. It appears that, compared with GILMSs, GISTs in dogs are more locally aggressive and more likely to result in perioperative death as a result of gastrointestinal wall perforation and sepsis; however, dogs with GISTs that are resected with clean margins have a better long-term prognosis. To confirm these hypotheses, studies with greater numbers of cases are required.
ABBREVIATIONS
| GILMS | Gastrointestinal leiomyosarcoma |
| ICC | Interstitial cell of Cajal |
| GIST | Gastrointestinal stromal tumor |
| SMA | Smooth muscle actin |
SAS software, version 9.1, SAS Institute Inc, Cary, NC.
Adriamycin, Ben Venue Laboratory Inc, Bedford, Ohio.
Cisplatin, Sicor Pharmaceutical Inc, Irvine, Calif.
Gleevec, Novartis International AG, Basel, Switzerland.
References
- 1↑
Phillips BS. Tumors of the intestinal tract. In:Withrow SJ, MacEwen EF, ed.Small animal clinical oncology. 3rd ed.Philadelphia: WB Saunders Co, 2001;335–346.
- 2
Chen HC, Parris LS, Parris RG. Duodenal leiomyosarcoma with multiple hepatic metastases in a dog. J Am Anim Hosp Assoc 1986;22:533–537.
- 3
Bruecker KA, Withrow SJ. Intestinal leiomysarcoma in six dogs. J Am Anim Hosp Assoc 1988;24:281–284.
- 4
Kapatkin AS, Mullen HS, Mathiesen DT, et al. Leiomyosarcoma in dogs: 44 cases (1983–1988). J Am Vet Med Assoc 1992;201:1077–1079.
- 5
Patnaik AK, Hurvitz AI, Johnson GF. Nonlymphoid intestinal neoplasia in 32 dogs and 14 cats. Vet Pathol 1977;14:547–555.
- 6
Cohen M, Post GS. Nephrogenic diabetes insipidus in a dog with intestinal leiomyosarcoma. J Am Vet Med Assoc 1999;215:1818–1820.
- 7
Sircar K, Hewlett BR, Huizinga JD. Interstitial cells of Cajal as precursors of gastrointestinal stromal tumors. Am J Surg Pathol 1999;23:377–389.
- 8
Nishida T, Hirota S. Biological and clinical review of stromal tumors in the gastrointestinal tract. Histol Histopathol 2000;15:1293–1301.
- 9↑
Miettinen M, Lasota J. Gastrointestinal stromal tumors—definition, clinical, histological, immunohistochemical, and molecular genetic features and differential diagnosis. Virchows Arch 2001;438:1–12.
- 10
Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 1998;279:577–580.
- 11
Kitamura Y, Hirota S, Nishida T. Molecular pathology of c-kit proto-oncogene and development of gastrointestinal stromal tumors. Ann Chir Gynaecol 1998;87:282–286.
- 12
Kindblom LG, Remotti HE, Aldenborg F, et al. Gastrointestinal pacemaker cell tumor (GI-PACT): gastrointestinal stromal tumors show phenotypic characteristics of the interstitial cells of Cajal. Am J Pathol 1998;152:1259–1269.
- 13
Lasota J, Jasinkski M, Sarlomo-Rikala M, et al. C-kit mutations occur preferentially in malignant versus benign GISTs and do not occur in leiomyomas and leiomyosarcomas. Am J Pathol 1999;154:53–60.
- 14
Lasota J, Wozniak A, Sarlomo-Rikala M, et al. Mutations in exon 9 and 13 of KIT gene are rare events in gastrointestinal stromal tumors. A study of 200 cases. Am J Pathol 2000;157:1091–1095.
- 15
Lux ML, Rubin BP, Biase TL, et al. KIT extracellular and kinase domain mutations in gastrointestinal stromal tumors. Am J Pathol 2000;156:791–795.
- 16
Taniguchi M, Nishida T, Hirota S, et al. Effect of c-kit mutation on prognosis of gastrointestinal stromal tumors. Cancer Res 1999;154:1643–1647.
- 17
Del Piero F, Summers BA, Cummings JF, et al. Gastrointestinal stromal tumors in equids. Vet Pathol 2001;38:689–697.
- 18
Hafner S, Harmon BG, King T. Gastrointestinal stromal tumors of the equine cecum. Vet Pathol 2001;38:242–246.
- 19
Banjerjee M, Lowenstine LJ, Munn RJ. Gastric stromal tumors in two Rhesus macaques (Macaca mulatta). Vet Pathol 1991;28:30–36.
- 20↑
Frost D, Lasota J, Miettinen M. Gastrointestinal stromal tumors and leiomyomas in the dog: a histopathologic, immunohistochemical, and molecular genetic study of 50 cases. Vet Pathol 2003;40:42–54.
- 21
Bettini G, Morini M, Marcato PS. Gastrointestinal spindle cell tumours of the dog: histological and immunohistochemical study. J Comp Pathol 2003;129:283–293.
- 22
Kumagai K, Uchida K, Miyamoto T, et al. Three cases of canine gastrointestinal stromal tumors with multiple differentiations and c-kit-expression. J Vet Med Sci 2003;65:1119–1122.
- 23↑
LaRock RG, Ginn PE. Immunohistochemical staining characteristic of canine gastrointestinal stromal tumors. Vet Pathol 1997;34:303–311.
- 24↑
Epperson A, Hatton WJ, Callaghan B, et al. Molecular markers expressed in cultured and freshly isolated interstitial cells of Cajal. Am J Physiol Cell Physiol 2000;279:C529–C539.
- 25↑
Demetri GD. Targeting the molecular pathophysiology of gastrointestinal stromal tumors with imatinib. Mechanisms, successes, and challenges to rational drug development. Hematol Oncol Clin North Am 2002;16:1115–1124.
- 26↑
Demetri GD, von Mehren M, Blanke CD, et al. Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors. N Engl J Med 2002;347:472–480.
- 27
Cohen MH, Williams G, Johnson JR, et al. Approval summary for imatinib mesylate capsules in the treatment of chronic myelogenous leukemia. Clin Cancer Res 2002;8:935–942.
- 28
Druker BJ, Lydon NB. Lessons learned from the development of an Abl tyrosine kinase inhibitor for chronic myelogenous leukemia. J Clin Invest 2000;105:3–7.
- 30
Birchard SJ, Couto CG, Johnson S. Nonlymphoid intestinal neoplasia in 32 dogs and 14 cats. J Am Anim Hosp Assoc 1986;22:533–537.
- 31
Crawshaw J, Berg J, Sardinas JC, et al. Prognosis for dogs with nonlymphomatous, small intestinal tumors treated by surgical excision. J Am Anim Hosp Assoc 1998;34:451–456.
- 32
Cohen M, Post GS, Wright JC. Gastrointestinal leiomyosarcoma in 14 dogs. J Vet Intern Med 2003;17:107–110.
