Gastrointestinal stromal tumor: consistent CD117 immunostaining for diagnosis, and prognostic classification based on...

Gain-of-function c-kit gene mutations and immunoreactivity of the c-kit protein CD117 in many gastrointestinal stromal tumors (GISTs) seem to support the idea that GISTs form a biologically distinct entity. In this study, the clinicopathologic features of 171 cases of GIST at a single institution were investigated for accurate diagnosis, and their relative risk for mortality was estimated by multivariate analysis. A GIST was defined diagnostically as a mesenchymal spindle or epithelioid cell lesion arising in the wall of the gastrointestinal tract with consistent immunoreactivity for CD117. The 171 patients with GISTs comprised 96 males (56.1%) and 75 females (43.9%), with a mean age of 59.4 years. One hundred and forty-five tumors (84.8%) occurred in the stomach, 18 (10.5%) in the small intestine, 6 (3.5%) in the rectum, and 2 (1.2%) in the esophagus. The median tumor size was 4.5 cm (range, 1.2 to 38 cm). Spindle-cell GISTs were present in 132 cases (77.2%); mixed GISTs, in 25 cases (14.6%); and epithelioid GISTs, in 14 cases (8.2%). Ten cases (55.6%) of spindled small intestine GIST contained eosinophilic skeinoid fibers. Immunoreactivity for CD34, h-caldesmon, alpha-smooth-muscle actin (SMA), desmin, and S-100 was observed in 156 (91.2%), 131 (76.6%), 46 (26.9%), 7 (4.1%), and 14 (8.2%) tumors, respectively. The percentage of CD34 positivity (38.8%) was low, in contrast with the high percentage of reactivity for SMA (77.8%) and S-100 (44.4%) in small intestine GISTs. By our histologic grading system using tumor differentiation, MIB-1 score, and necrosis, 129 tumors (75.4%) were classified as low grade and 42 tumors (24.6%) were classified as high grade. With a median follow-up period of 83.5 months for 122 living patients, the 5-year and 10-year survival rates were 81.7% and 67.4%, respectively. Multivariate analysis showed that both tumor size >10 cm and high grade were significantly associated with a poor outcome. As a result, GISTs >10 cm or high grade, 5 to 10 cm and low grade, and < or =5 cm and low grade were regarded as high risk, intermediate risk, and low risk for mortality, respectively. In conclusion, it is important to recognize GISTs that have a specific molecular pathogenesis and to separate them from other mesenchymal tumors with optimal immunostaining for CD117 when making a diagnosis and prognostic classification based on tumor size and MIB-1 grade.