Objective: To characterize the CT features and to identify predictors of malignancy from CT of GISTs.

Material and Method: A retrospective review of CT images of 50 patients with pathologically and immunohistochemically proven GISTs was done by two radiologists and final interpretations were reached by consensus. Images were evaluated for site, size, contour, boundary, growth pattern, enhancement pattern, degree of enhancement, necrosis, calcification, ulceration, perilesional fat stranding, evidence of bowel obstruction, and signs of malignancy. Categorical variables were compared using Fisher’s exact test and continuous variables used the t-test. Univariate and multivariate logistic regression models were used to identify significant predictors of a high mitotic rate.

Results: Of the 50 patients, the most common location of GISTs was stomach (62%) The mean size was 10.2 cm (SD 5.2 cm). The contour was lobulated in 84%. The boundary was smooth in 84%. The growth pattern was exophytic in 68%. Most of tumors had heterogeneous density on post-contrast images (88%). Necrosis (84%), calcification (14%), ulceration (40%), perilesional fat stranding (44%), and bowel obstruction (2%) were present in the tumors. The CT signs of malignancy found were adjacent organ invasion (18%), ascites (18%), lymphadenopathy (6%), liver metastasis (20%), and peritoneal seeding (16%). Necrosis and peritoneal seeding were statistically significant independent predictors for high mitotic GISTs in multivariate logistic regression (p<0.05). The probability of a high mitotic rate was 1 (95% CI, 0.40-1.00) in the presence of both necrosis and peritoneal seeding.

Conclusion: The stomach was the most common site of GIST. The CT features of GIST were lobulated, smooth tumor margins, exophytic growth pattern, and heterogeneous enhancement on post-contrast CT images. Presence of both necrosis and peritoneal seeding were found to be a significant predictor of high mitotic rate of GISTs. The probability of a high mitotic rate was 1 (95% CI, 0.40-1.00).

Keywords: Computed tomography, Gastrointestinal stromal tumor (GIST)