The effects of prostate volume and PI-RADS category on optimal PSA-density thresholds for biopsy decision-making

OBJECTIVES: To evaluate the effect of prostate volume on the risk of clinically significant prostate cancer (csPCa) across a range of PSA-density (PSAd) values, and to explore the relationship between PI-RADS category and PSAd in predicting csPCa. MATERIALS AND METHODS: We retrospectively analyzed 2190 patients undergoing mpMRI for suspected PCa. Patients were classified as csPCa and clinically insignificant (negative and insignificant PCa). Logistic regression was performed to assess the csPCa risk at different PSAd cut-offs across different prostate volume subgroups (< 40, 40-60, 60-80, > 80 mL) and PI-RADS categories. The effect of prostate volume on PSAd performance was evaluated using ROC analysis. To assess robustness, we performed an 80:20 split-sample internal validation. RESULTS: 747/2190 (34.1%) patients had PCa, including 571 (26.1%) with csPCa. Regardless of PSAd, csPCa risk exceeded 30% for PI-RADS 4 and 50% for PI-RADS 5. At a 10% csPCa risk threshold, the optimal PSAd cut-offs were 0.20 ng/mL² for PI-RADS 1-2 and 0.12 ng/mL² for PI-RADS 3. Logistic regression showed a significant inverse correlation between prostate volume and csPCa probability. Notably, 79% of csPCa patients with prostate volume ≤ 40 mL had a PSAd ≥ 0.15 ng/mL², compared to only 22.4% with volumes ≥ 60 mL. PSAd performed significantly worse for larger glands (≥ 60 mL), with AUCs of 0.70 versus 0.84 (≤ 40 mL) and 0.82 (40-60 mL), both p < 0.001. CONCLUSION: The optimal PSAd cut-offs for guiding biopsy decisions were 0.20 ng/mL² for PI-RADS 1-2 and 0.12 ng/mL² for PI-RADS 3. When using PSAd to guide biopsy decision for PI-RADS 1-3 patients with large prostates (> 60 mL), caution is warranted, as PSAd becomes significantly less accurate. KEY POINTS: Question The optimal PSA-density thresholds for biopsy decisions in PI-RADS 1-3 patients remain uncertain, and data on the impact of prostate volume on its performance are limited. Findings Optimal PSA-density thresholds were 0.20 ng/mL² for PI-RADS 1-2 and 0.12 ng/mL² for PI-RADS 3. Diagnostic performance of PSA-density decreased significantly in men with larger glands (≥ 60 mL). Clinical relevance PSA-density cut-offs of 0.20 (PI-RADS 1-2) and 0.12 ng/mL² (PI-RADS 3) can guide biopsy decisions. In PI-RADS 1-3 patients with large prostate (≥ 60 mL), PSA-density becomes significantly less predictive, and low values may not reliably exclude csPCa.