Objective: The purpose of this study was to evaluate MDCT parameters for differentiating malignant (category T1 and T2) from benign gastric ulcers and to evaluate the performance characteristics of these predictors with optimal cutoff points determined in receiver operator characteristic analysis.
Subjects and methods: The subjects were 26 patients with gastric cancer (11 with T1 lesions, 15 with T2 lesions) and 26 patients with benign gastric ulcer. MDCT and virtual gastroscopic findings were analyzed according to four qualitative criteria: ulcer shape, base, and margin and changes in adjacent folds. The quantitative criteria ulcer size, thickness of the gastric wall around an ulcer, thickness of the enhanced ulcer base, and enhancement around an ulcer were measured on multiplanar reconstruction images. We calculated the sensitivity and specificity of each quantitative criterion. Receiver operator characteristic analysis was used to identify cutoff points yielding optimal sensitivity and specificity for the diagnosis of gastric cancer.
Results: On virtual gastroscopy, ulcer shape and margin and gastric fold changes had sensitivities of 80.8%, 84.6%, and 90.9% and specificities of 76.9%, 73.1%, and 77.8%, respectively, in the diagnosis of gastric cancer. On multiplanar reconstruction images, thickness of the enhanced ulcer base and enhancement around the ulcer had sensitivities of 80.8% and 73.1% and specificities of 100% and 100%.
Conclusion: MDCT combined with virtual gastroscopy and multiplanar reconstruction enhances the morphologic details of gastric ulcers and is a useful way to differentiate malignant (T1 and T2) and benign gastric ulcers.