Diagnostic value of texture analysis of apparent diffusion coefficient maps for differentiating fat-poor angiomyolipoma from non-clear-cell renal cell carcinoma

Purpose To investigate the feasibility of texture analysis apparent diffusion coefficient (ADC) maps for differentiating fat-poor angiomyolipomas (fpAMLs) from non-clear-cell renal cell carcinomas (non-ccRCCs). Methods In this bi-institutional study, we included two consecutive cohorts different institutions with pathologically confirmed solid masses: 67 patients (fpAML = 46; non-ccRCC 21) model development and 39 24; 15) validation. Patients underwent preoperative magnetic resonance imaging (MRI), including diffusion-weighted imaging. We extracted 45 features using a software volumes interest on ADC maps. Receiver operating characteristic curve was performed to compare diagnostic performance between random forest (RF) (derived features) conventional subjective evaluation computed tomography MRI by radiologists. Results RF revealed that grey-level zone length matrix long-zone high emphasis dominant feature diagnosing fpAML. The area under (AUC) distinguish fpAMLs non-ccRCCs not significantly validation (p .19). cohort, AUC similar board-certified radiologists .46) higher than radiology residents .03). Conclusions Texture demonstrated discriminating non-ccRCCs. Diagnostic performances in were comparable despite data device manufacturers institutions.