To evaluate the value of computed tomography (CT) morphological features, especially the "overflowing beer sign" and "angular interface", in differentiating fat-poor angiomyolipoma (fpAML) from clear cell renal cell carcinoma (ccRCC).

The clinical and CT data of 259 patients with pathologically confirmed fpAML (n=64) and ccRCC (n=195) at our hospital from January 2019 to December 2024 were retrospectively analyzed. All tumors had a maximum diameter of ≤4 cm. Univariate and multivariate logistic regression analyses were used to identify independent predictors for fpAML and ccRCC. Incremental diagnostic value of key imaging features was assessed by constructing progressive diagnostic models, and an optimal combined model was established.

Multivariate logistic regression analysis identified the overflowing beer sign (OR=25.127), non-round shape (OR=6.284), female gender (OR=3.445), and smaller maximum tumor diameter (OR=0.534) as independent predictors for fpAML (all P<0.05). The angular interface was significant in univariate analysis ( P<0.001) but was not included in the final multivariate model. Adding the "overflowing beer sign" (model 2) to the clinical model (model 1) (AUC=0.817) significantly improved its performance (AUC=0.930, P<0.001). The combined model (model 3) incorporating all independent predictors achieved the best diagnostic performance (AUC=0.940), which was superior to model 2 (P=0.023).

The "overflowing beer sign" is the most powerful independent CT predictor for differentiating fpAML from ccRCC. A combined model integrating the overflowing beer sign, tumor shape, gender, and maximum diameter can offer an efficient and reliable tool for preoperative differentiation.