Performing Precise Biopsy in Naive Patients With Equivocal PI-RADS, Version 2, Score 3, Lesions: An MRI-based...

The primary objective of the present study was to avoid unnecessary prostate biopsy in biopsy-naive patients with Prostate Imaging Reporting and Data System, version 2 (PI-RADS v2), score 3, lesions.

We reviewed our prospectively maintained database from January 2012 to July 2018. Logistic regression analyses were performed to test different clinical factors as predictors of clinically significant prostate cancer (CSPCa) and build nomograms. Calibration curves were used to assess the concordance between the predictive value and the true risk. Decision curves were created to measure the overall net benefit.

The prostate cancer (PCa) and CSPCa detection rates were 37.2% (81 of 218) and 23.9% (52 of 218) in the PI-RADS v2, score 3, cohort. More PCa cases (61.7%; 50 of 81) and CSPCa cases (75%; 39 of 52) were found in the peripheral zone than in the transitional zone. Multivariate analysis showed that age, prostate-specific antigen density, lesion region, and apparent diffusion coefficient (ADC) were predictive factors for CSPCa and PCa. Internally validated calibration curves showed that the predicted risk of CSPCa was closer to the actual probability when the threshold was > 60%. Decision curves showed that a better net benefit was achieved when the model was used to guide clinical practice.

More cases of PCa and CSPCa were seen in the peripheral zone than in the transitional zone among patients with PI-RADS v2, score 3. The positive predictive value for a positive ADC (< 900 μm²/s) for the detection of CSPCa and PCa improved with an increasing prostate-specific antigen density. Biopsy can be avoided if the equivocal lesion has a negative ADC (> 900 μm²/s) and was in the transition zone.