PSA proportion of index lesion: a novel index for predicting prostate pathology outcomes in biopsy-naïve patients with PSA < 10 ng/ml and PI-RADS 3 lesions.

objectiveTo validate a novel marker, called prostate-specific antigen (PSA) proportion of index lesion (PPIL), that could predict prostate pathology outcomes in biopsy-naïve patients with PSA levels < 10 ng/ml and Prostate Imaging Reporting and Data System (PI-RADS) 3 lesions.

methodsThis study included 356 men retrospectively from 2019 to 2024 in cohort 1 (training set and internal validation set, The First Affiliated Hospital of Soochow University). Between 2020 and 2024, 158 men were included in cohort 2 (external validation set, The Fourth Affiliated Hospital of Soochow University). The volume of all index lesions was calculated using the Slicer-3D software. Logistic regression analysis was performed for variables including PPIL and classic parameters of prostate cancer (PCa) in cohorts 1 and 2. Receiver operating characteristic curve analysis was performed, and the area under the curve (AUC) was calculated to assess utility.

resultsIn total, 100 (28.1%) patients were diagnosed with PCa in cohort 1, and 51 (51.0%) had clinically significant prostate cancer (CsPCa). In cohort 2, 55 patients (34.8%) were diagnosed with PCa. Among these, 33 (60.0%) had CsPCa. Age, prostate volume, prostate-specific antigen density, lesion volume and PPIL were predictive factors for PCa and csPCa. The highest AUC of PPIL was noted in the training set and internal and external validation sets for PCa (0.743, 0.767, and 0.750) and csPCa (0.778, 0.927, and 0.766), respectively.

conclusionPPIL may be a useful marker for predicting prostate pathology outcomes in biopsy-naïve patients with PSA level < 10 ng/ml and PI-RADS 3 lesions. This novel index may assist urologists in biopsy decision making for these so-called "double gray zone" patients.