A scoring model for differentiating gastric calcifying fibrous tumors from gastrointestinal stromal tumors less than 2 cm based on CT features

OBJECTIVES: This study aimed to compare the CT features of calcifying fibrous tumors (CFTs) and gastrointestinal stromal tumors (GISTs) less than 2 cm, and to establish a scoring system to differentiate them. METHODS: A total of 85 patients were included, comprising 22 CFTs and 63 GISTs. Clinical and CT imaging features were compared between the two groups. Independent predictors were identified using univariate and multivariate logistic regression analyses. A predictive model was constructed and converted into a simplified scoring system based on regression coefficients. Model performance was evaluated using receiver operating characteristic (ROC) analysis and the Hosmer-Lemeshow goodness-of-fit test. RESULTS: The analysis revealed eight variables with statistically significant differences (P < 0.05). Four variables were included in the score model, including long diameter/short diameter (LD/SD) ratio, calcification (mild), calcification (moderate), and degree of enhancement (moderate), and identified as independent predictors of CFT. The scoring system ranged from - 2 to 5 points, where higher scores correlated with a greater likelihood of CFT. The AUCs for the predictive model and scoring model were 0.833 and 0.822, respectively, with no significant difference between the two (P = 0.595). To facilitate clinical application, the scoring system was divided into four ranges, with corresponding probabilities of CFT of 7.14%, 14.29%, 64.71%, and 80.00%. CONCLUSION: The proposed CT-based predictive model and scoring system demonstrate good diagnostic performance and may enhance diagnostic confidence, potentially reducing unnecessary invasive procedures when CFT is suspected.