Morphologic analysis with computed tomography may help differentiate fat-poor angiomyolipoma from renal cell carcinoma: a retrospective study with 602 patients

Abstract

PURPOSE: To assess whether morphologic analysis using computed tomography (CT) could differentiate between fat-poor angiomyolipoma (fpAML) and renal cell carcinoma (RCC). METHODS: A total of 602 patients with a histologically confirmed fpAML (n = 49) or RCC (n = 553) were evaluated. All renal lesions were less than 4 cm in size and had no gross fat on contrast-enhanced CT. For morphologic analysis, overflowing beer sign and angular interface were evaluated. Overflowing beer sign was defined as contact length between bulging-out portion of a mass and the adjacent renal capsule of 3 mm or greater. Angular interface was defined as the angle of parenchymal portion of a mass of 90 degrees or less. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were assessed. Multivariate analysis was conducted to determine which variable is predictive of fpAML. RESULTS: Sensitivity, specificity, PPV, NPV, and accuracy were 61.2% (30/49), 97.1% (537/553), 65.2% (30/46), 96.6% (537/556), and 94.2% (567/602) with overflowing beer sign, while they were 55.1% (27/49), 81.9% (453/553), 21.3% (27/127), 95.4% (453/475), and 79.7% (480/602) with angular interface for fpAML, respectively. Both CT variables were predictive of fpAML (overflowing beer sign, odds ratio = 132.881, p < 0.001; angular interface, odds ratio = 5.766, p = 0.010). The multivariate model with CT variables showed good performance for predicting fpAML (AUC, 0.871 with angular interface, 0.943 with overflowing beer sign, and 0.949 with both). CONCLUSION: Morphologic analysis with contrast-enhanced CT may be useful for differentiating fpAML from RCC. Overflowing beer sign has the potential as an imaging biomarker for fpAML.