Computed tomography (CT) in predicting the risk of malignancy in gastrointestinal stromal tumors (GIST) in correlation with mitotic index

Abstract

Background: Despite being an uncommon form of abdominal neoplasia, gastrointestinal stromal tumors are the most prevalent mesenchymal tumors of the gastrointestinal system.
Objective: This research aimed to identify the predictors of malignancy on computed tomography (CT) for the evaluation of gastrointestinal stromal tumors (GIST) by correlating CT findings with the mitotic index.
Patients and methods: We retrospectively reviewed the radiological and pathological results of forty-five patients who had a histopathological diagnosis of GIST. All underwent post contrast CT for the abdomen and pelvis.
Results: There was a significant correlation among mitotic count and sex, area of necrosis, contour, mesenteric fat infiltration, and distant metastasis with p value < 0.05. The sensitivity of the 20% cut-off value was 74.2%, the specificity was 85.7%, and the accuracy was 77.8%. The sensitivity of 10% cut-off value was 93.5%, the specificity was 78.6%, and the accuracy was 88.9%. Out of 45 patients with GIST lesions, 39 (86.7%) had radiological findings that agreed with the pathological mitotic count. Six patients (13.3%) had mismatches (four favoring pathology and two favoring
radiology).
Conclusions: Although, pathology remains the cornerstone in having the final verdict regarding the malignant potential of GIST, the CT-based approach seems to be effective for risk stratification prior to surgery and treatment. This approach in conjunction with other risk variables such as patient age and morbidities could be helpful for a better management of GIST patients, to address them to the best tailored treatment at the time of diagnosis.