PHI for adverse pathology and PHI density for upgrading: two complementary predictors in localized prostate cancer

Background The Prostate Health Index (PHI) has emerged as a promising biomarker for improving prostate cancer (PCa) detection. However, its role in predicting pathological aggressiveness at radical prostatectomy (RP) requires further elucidation. We aimed to assess the ability of PHI and PHI density (PHId) to predict adverse pathological features and ISUP upgrading in men undergoing robot-assisted RP. Methods We conducted a prospective, single-center observational study including 130 patients who underwent RP. Total PSA, free PSA, and p2PSA were measured using the Access 2 platform, and PHI and PHId were calculated accordingly. Multivariable logistic regression models evaluated the association of PSA, PSA density, PHI, and PHId with two predefined endpoints: (i) adverse pathological features (pT ≥3 and/or pN1), and (ii) upgrading from biopsy to RP. Models were adjusted for DRE status, PI-RADS score, and biopsy ISUP grade. Discrimination was assessed using AUC and the DeLong test. Results Median PSA, PHI, PSA density, and PHId were 6.21 ng/mL, 58, 0.15 ng/mL/cc, and 1.40, respectively. Adverse pathological features were identified in 39% of cases, and upgrading occurred in 59% (after exclusion of biopsy ISUP 5 cases). PHI emerged as the strongest independent predictor of adverse pathology (OR 1.04; 95% CI 1.02–1.07), yielding the highest model accuracy (AUC 0.86) and showing a sharp risk increase above PHI > 40. PHId was the only marker independently associated with upgrading (OR 1.54; 95% CI 1.10–2.29), achieving the best discrimination (AUC 0.83), with risk escalation particularly evident in patients with PI-RADS 4–5 lesions. Conclusions PHI and PHId provide complementary information in localized PCa. PHI accurately identifies patients at risk of adverse pathological extension, whereas PHId effectively captures the probability of grade reclassification at RP. Integrating these biomarkers into patient assessment may support precision-based decision-making and helping to minimize undertreatment.