Case 8 -

Gastric Stromal Tumors Elizabeth Montgomery, MD Johns Hopkins Medical Institute

Clinical history: Female, 61-year-old, This slide was prepares from a 7 cm mass excised from the stomach. Case 8 - Slide 1 Click to view with ImageScope Click to view with a Web-Based Viewer Gastric Stromal Tumors The stomach is the most common site for gastrointestinal stromal tumors (GIST, efig 325) and they are occasionally diagnosed on mucosal biopsies. Typically those diagnosed on biopsies are aggressive lesions that have invaded the mucosa. GISTs are mesenchymal tumors arising in the GI tract and occasionally within the abdomen with no GI connection. The earlier literature attempted to classify them as smooth muscle or nerve sheath tumors, but even in the benign tumors evidence for such differentiation was difficult to find. Mazur and Clark coined the term stromal tumor in 1983 [1] . GISTs show differentiation towards (and supposedly arise from a precursor of) interstitial cells of Cajal which are normally concerned with motility of the gut [2, 3] . This category is considered to subsume the ultrastructurally defined rare lesions previously known as gastrointestinal autonomic nerve tumors (GANT). The availability of specific antibodies and clarification of their immunohistochemical profile has facilitated diagnosis. However, careful morphologic examination and clinicopathological correlation remain essential for excluding mimics and for assessment of likely behavior in this heterogeneous group of neoplasms. GISTs comprise 5-10% of all sarcomas (for comparison, retroperitoneal = 15% and extremity 42%), and around 25% of GISTs are malignant, representing about 1% of all GI malignancies. They are most common in adults aged 50-60. These tumors vary in differentiation and prognosis according to their location within the GI tract. Esophageal GISTs are rare, but 50-70% involve the stomach, 25-40% the small intestine (of which 10-20% arise in duodenum, 27-37% in jejunum, 27-53% in ileum) and < 10% are colorectal (50% colonic, 50% rectal). They are least frequent in the esophagus (where smooth muscle tumors are more usually found), colon and rectum. GIST-type tumors arising in omentum, peritoneum and retroperitoneum have been identified, comprising 6.7% of the large AFIP series of 1008 GISTs [4] . Tumors present with primary mass, pain or bleeding (46%) or metastasis (47%) and two thirds exceed 5cm at presentation. Malignant GIST are rare with about 5 per million of population (compared with 25 per million for soft tissue sarcomas). There is an increased incidence of GIST including multiple tumors in small intestine in patients with NF-1, possibly associated with interaction between the NF-1 gene product and c-kit) and GIST is occasionally familial (associated with a germline c-kit mutation, see below). Some patients have second cancers, and some epithelioid gastric stromal tumors are associated with paraganglioma and pulmonary chondroma in Carney's triad. About 20% are malignant. The consensus is that 5 mitoses per 50 hpf and >5cm are adverse prognostic factors [5] . The five-year survival of gastric GISTs is about 40%, with improvement in completely resected cases. There is no evidence that radical surgery improves survival, so that the least extensive surgical procedure compatible with complete excision is advisable. Gastric GISTs are more frequent in males, but young patients, especially female, have an improved outcome; Persson et al (1992) reported 4 patients less than 26 years old who had metastasizing gastric epithelioid leiomyosarcomas, involving liver, nodes, peritoneum, and who all survived for 17-48 years. Epithelioid GISTs are the old leiomyoblastomas of Stout (the term epithelioid leiomyoma was introduced in the 1969 WHO). They comprise about 11% of gastric GISTs, of which 73-81% are benign. Large tumors in the fundus or cardiac area and posterior wall are more likely to be malignant. 47-60% of tumors metastasize, usually within 2 years and survival after metastasis is about 8 months. Lee et al found tumors larger than 6 cm had a worse outcome, and in the series of Appelman et al, no tumor <5.5 cm metastasized, and only one < 6cm. In the latter series, 1 case with no mitoses metastasized, 12% with 1-5 mitoses per 50 hpf metastasized and all cases with >10 mitoses per 50 hpf. (None had 6-10). Benign tumors had 0-5 mitoses per 50 hpf. Metastasis occurred from 55% of well differentiated and 75% of poorly differentiated epithelioid leiomyosarcomas. Ma et al also found that tumors with >5 mitoses per 50 hpf had malignant potential. In our series, necrosis had significant negative impact [6] . Most GISTs are spindle cell tumors with variable palisading, peculiar paranuclear vacuoles, and collagen fibrils. On a practical note, it is epithelioid GISTs that are most likely to cause diagnostic problems on mucosal biopsies as they are readily mistaken for a host of epithelioid and epithelial neoplasms. It is advisable to perform an immunohistochemical panel in assessing them to include CD117/c-kit, S100 protein (to address melanoma) and a cytokeratin stain to address signet cell carcinoma. The majority of GIST have kit mutations and are CD117/c-kit stain positive, but 5-10% lack kit mutations and some in the c-kit negative subset have alternate mutations of platelet derived growth factor alpha instead [7, 8, 9] . Since about 70% of GISTs express CD34, this can also be included in a diagnostic panel. On mucosal biopsies, it is difficult to assess tumor size and mitotic counts to prognosticate but it is possible to make a diagnosis in many cases and allow the patient to be scheduled for an operation to remove the tumor. On resection specimens a "recipe" for this has been published [10] , Table 1. Some observers took issue with this as it assumes that all GISTs are potentially malignant and does not take site into account but is still a useful construct for attempting a prognostic "forecast". Trupiano et al have published criteria that they believe allow separation of benign gastric stromal tumors from those they regard as not benign [11] (efig 333). We have attempted to apply these to a series of gastric GISTs and had one "outlier" case [6] so we continue to add a note of caution to reports on large gastric stromal tumors that appear otherwise benign. The AFIP group has the largest accumulated series of gastric stromal tumors, and has reported a large series of 1,765 GISTs confined to the stomach with good follow-up information [12] . There was a slight male predominance (55%) and a median age of 63 years. Only 2.7% of tumors occurred before the age of 21 years and 9.1% before the age of 40 years. The tumors varied from 0.5 to 44 cm (median, 6.0 cm) and most commonly presented with GI bleeding; 12% were incidentally detected. Histologic variants recognized among the spindle cell tumors included sclerosing, palisaded-vacuolated, hypercellular, and sarcomatous and among the epithelioid tumors, sclerosing, dyscohesive, hypercellular, and sarcomatous. Outcome was strongly dependent on tumor size and mitotic activity. Only 2% to 3% of tumors <10 cm and <5 mitoses/50 HPFs metastasized, whereas 86% of tumors >10 cm and >5 mitoses/50 HPFs metastasized. However, tumors >10 cm with mitotic activity <5/50 HPFs and those <5 cm with mitoses >5/50 HPFs had a relatively low metastatic rate (11% and 15%). A small number of patients survived intra-abdominal metastasis up to over 20 years. Tumor location in fundus or gastroesophageal junction, coagulative necrosis, ulceration, and mucosal invasion were all unfavorable factors (P < 0.001), whereas tumor location in antrum was favorable (P < 0.001). Probably the key feature of this very large series is that it allowed separating out a "benign" category of gastric GISTs based on large numbers of cases [12] . This group has proposed malignancy criteria specific to the stomach which are shown in Table 2. Another important feature of the AFIP study is that mucosal extension by gastric GISTs was evidence that the tumor would behave badly. This feature remained equivocal based on several earlier smaller studies (including our own! [6] ). However, based on these data from Miettinen et al [12] , if we encounter a gastric GIST on a mucosal biopsy, it is likely malignant and we should suggest this in our reports. Based on the large AFIP series of gastric GISTs, KIT/CD117 is found in 91% of gastric GISTs, CD34 in 82%, SMA in 18%, and desmin in 5%; the latter 2 tend to be focal. As a last diagnostic "tip", pathologists should be aware that a number of lesions other than GIST display CD117. Perhaps fibromatoses are most likely to result in diagnostic errors [13, 14] . We have had some success with using beta catenin for this separation [only fibromatoses show nuclear staining] [14, 15] . Table 1. Suggested Prognostic Criteria for Resected GISTs FROM ALL ANATOMIC SITES [10] Risk Category Size (cm) Mitotic counts/50 high power fields Very low risk <2 <5 Low risk 2-5 <5 Intermediate risk <5 5-10 6-10 <5 High risk >5 >10 Any size >5 Any >10 Table 2. Suggested Prognostic Criteria for Resected GISTs SPECIFIC TO STOMACH [12] Risk Category Size Mitotic Counts (per 50 high power fields) Benign No larger than 2 cm No more than 5/50 Probably Benign >2, < 5 cm >5, < 10 cm No more than 5/50 No more than 5/50 Uncertain No larger than 2 cm >5/ 50 Low to Moderate Malignant Potential >10 cm >2, < 5 cm Not more than 5/ 50 >5/ 50 High Malignant potential >5, < 10 cm > 10 cm > 5/ 50 > 5/ 50 References. Mazur MT, Clark HB. Gastric stromal tumors. Reappraisal of histogenesis. Am J Surg Pathol. 1983;7(6):507-19. Kindblom LG, Remotti HE, Aldenborg F, Meis-Kindblom JM. Gastrointestinal pacemaker cell tumor (GIPACT): gastrointestinal stromal tumors show phenotypic characteristics of the interstitial cells of Cajal [see comments]. Am J Pathol. 1998;152(5):1259-69. Sircar K, Hewlett BR, Huizinga JD, Chorneyko K, Berezin I, Riddell RH. Interstitial cells of Cajal as precursors of gastrointestinal stromal tumors. Am J Surg Pathol. 1999;23(4):377-89. Emory TS, Sobin LH, Lukes L, Lee DH, O'Leary TJ. Prognosis of gastrointestinal smooth-muscle (stromal) tumors: dependence on anatomic site. Am J Surg Pathol. 1999;23(1):82-7. Miettinen M, Lasota J. Gastrointestinal stromal tumors--definition, clinical, histological, immunohistochemical, and molecular genetic features and differential diagnosis. Virchows Arch. 2001;438(1):1-12. Montgomery E, Abraham SC, Fisher C, et al. CD44 loss in gastric stromal tumors as a prognostic marker. Am J Surg Pathol. 2004;28(2):168-177. Heinrich MC, Corless CL, Duensing A, et al. PDGFRA activating mutations in gastrointestinal stromal tumors. Science. 2003;299(5607):708-10. Heinrich MC, Corless CL, Demetri GD, et al. Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J Clin Oncol. 2003;21(23):4342-9. Yamamoto H, Oda Y, Kawaguchi K, et al. c-kit and PDGFRA mutations in extragastrointestinal stromal tumor (gastrointestinal stromal tumor of the soft tissue). Am J Surg Pathol. 2004;28(4):479-88. Fletcher CD, Berman JJ, Corless C, et al. Diagnosis of gastrointestinal stromal tumors: A consensus approach. Hum Pathol. 2002;33(5):459-65. Trupiano JK, Stewart RE, Misick C, Appelman HD, Goldblum JR. Gastric stromal tumors: a clinicopathologic study of 77 cases with correlation of features with nonaggressive and aggressive clinical behaviors. Am J Surg Pathol. 2002;26(6):705-14. Miettinen M, Sobin LH, Lasota J. Gastrointestinal Stromal Tumors of the Stomach: A Clinicopathologic, Immunohistochemical, and Molecular Genetic Study of 1765 Cases With Long-term Follow-up. Am J Surg Pathol. 2005;29(1):52-68. Yantiss RK, Spiro IJ, Compton CC, Rosenberg AE. Gastrointestinal stromal tumor versus intra-abdominal fibromatosis of the bowel wall: a clinically important differential diagnosis. Am J Surg Pathol. 2000;24(7):947-57. Montgomery E, Torbenson MS, Kaushal M, Fisher C, Abraham SC. Beta-catenin immunohistochemistry separates mesenteric fibromatosis from gastrointestinal stromal tumor and sclerosing mesenteritis. Am J Surg Pathol. 2002;26(10):1296-301. Bhattacharya B, Dilworth HP, Iacobuzio-Donahue C, et al. Nuclear beta-catenin expression distinguishes deep fibromatosis from other benign and malignant fibroblastic and myofibroblastic lesions. Am J Surg Pathol. 2005;29(5):653-9.