A radiologic diagnostic scoring model based on CT features for differentiating gastric schwannoma from gastric gastrointestinal stromal tumors.

We aimed to further explore the CT features of gastric schwannoma (GS), propose and validate a convenient diagnostic scoring system to distinguish GS from gastric gastrointestinal stromal tumors (GISTs) preoperatively. 170 patients with submucosal tumors pathologically confirmed (GS n=35; gastric GISTs n=135) from Hospital 1 were analyzed retrospectively as the training cohort, and 72 patients (GS=11; gastric GISTs=61) from Hospital 2 were enrolled as the validation cohort. We searched for significant CT imaging characteristics and constructed the scoring system via binary logistic regression and converted regression coefficients to weighted scores. The ROC curves, AUCs and calibration tests were carried out to evaluate the scoring models in both the training cohort and the validation cohort. For convenient assessment, the system was further divided into four score ranges and their diagnostic probability of GS was calculated respectively. Four CT imaging characteristics were ultimately enrolled in this scoring system, including transverse position (2 points), location (5 points), perilesional lymph nodes (6 points) and pattern of enhancement (2 points). The AUC of the scoring model in the training cohort were 0.873 (95% CI, 0.816-0.929) and the cutoff point was 6 points. In the validation cohort, the AUC was 0.898 (95% CI, 0.804-0.957) and the cutoff value was 5 points. Four score ranges were as follows: 0-3 points for very low probability of GS, 4-7 points for low probability; 8-9 points for middle probability; 10-15 points for very high probability. A convenient scoring model to preoperatively discriminate GS from gastric GISTs was finally proposed. AJCR