Diagnostic value of MRI-based PSA density in predicting transperineal sector-guided prostate biopsy outcomes

Purpose Prostate-specific antigen (PSA) density (PSAD) has potential to increase the diagnostic utility of PSA, yet has had poor uptake in clinical practice. We aimed to determine the diagnostic value of magnetic resonance imaging-derived PSAD (MR-PSAD) in predicting transperineal sector-guided prostate biopsy (TPSB) outcomes. Materials and methods Men presenting for primary TPSB from 2007 to 2014 were considered. Histological outcomes were assessed and defined as: presence of any cancer or significant cancer defined as presence of Gleason 4 and/or maximum tumour core length (MCCL) ≥ 4 mm (G4); or Gleason 4 and/or MCCL ≥ 6 mm (G6). Sensitivity, specificity and positive and negative predictive values were calculated, and receiver operating characteristics (ROC) curves were generated to compare MR-PSAD and PSA. Results Six hundred fifty-nine men were evaluated with mean age 62.5 ± 9 years, median PSA 6.7 ng/ml (range 0.5–40.0), prostate volume 40 cc (range 7–187) and MR-PSAD 0.15 ng/ml/cc (range 0.019–1.3). ROC area under the curve (95% CI) was significantly better for MR-PSAD than PSA for all cancer definitions ( p < 0.001): 0.73 (0.70–0.76) versus 0.61 (0.57–0.64) for any cancer; 0.75 (0.71–0.78) versus 0.66 (0.62–0.69) for G4; and 0.77 (0.74–0.80) versus 0.68 (0.64–0.71) for G6. Sensitivities for MR-PSAD < 0.1 ng/ml/cc were 85.0, 89.9 and 91.9% for any, G4 and G6 cancer, respectively. Conclusion MR-PSAD may be better than total PSA in determining risk of positive biopsy outcome. Its use may improve risk stratification and reduce unnecessary biopsies.