Prostate Health Index Density Outperforms Prostate-specific Antigen Density in the Diagnosis of Clinically Significant Prostate Cancer in Equivocal Magnetic Resonance Imaging of the Prostate: A Multicenter Evaluation.

ABSTRACT

PURPOSE: We compare Prostate Health Index, Prostate Health Index density, and PSA density in predicting clinically significant prostate cancer in MRI-guided prostate biopsy. MATERIALS AND METHODS: This is a multicenter evaluation of prospectively maintained prostate biopsy databases at 10 urology centers. Men with Prostate Health Index and MRI-guided targeted and systematic prostate biopsy performed and without prior prostate cancer diagnosis were included. The additional value of PSA density, Prostate Health Index, and Prostate Health Index density to MRI PI-RADS (Prostate Imaging Reporting & Data System) score was evaluated with multivariable analyses, area under the curve, and decision curve analyses. The proportion of unnecessary biopsies that can be avoided are estimated for clinically significant prostate cancer (International Society of Urological Pathology group ≥2 prostate cancer). RESULTS: A total of 1,215 men were analyzed. Prostate cancer and clinically significant prostate cancer were diagnosed in 51% (617/1,215) and 35% (422/1,215) of men, respectively. Clinically significant prostate cancer was diagnosed in 4.4% (3/68), 15% (72/470), 39% (176/446), and 74% (171/231) of highest PI-RADS score of 2, 3, 4, and 5 lesions, respectively. In multivariable analyses, independent predictors for clinically significant prostate cancer detection included Prostate Health Index (OR 1.04), prostate volume (OR 0.97), and PI-RADS score 4 (OR 2.81) and 5 (OR 8.34). Area under the curve for clinically significant prostate cancer of PI-RADS + Prostate Health Index density (0.85) was superior to PI-RADS + PSA density (0.81), Prostate Health Index density (0.81), Prostate Health Index (0.78), PI-RADS (0.76), PSA density (0.72), and PSA (0.60) in the whole cohort, and the superiority of Prostate Health Index density was also observed in PI-RADS 3 lesions. Decision curve analysis showed Prostate Health Index density achieving the best net clinical benefit in PI-RADS 3 or 4 cases. Among PI-RADS 3 lesions, using cutoffs of PSA density 0.15, Prostate Health Index 38.0, and Prostate Health Index density 0.83 could reduce 58%, 67%, and 72% of unnecessary biopsies, respectively. CONCLUSIONS: Prostate Health Index density outperformed Prostate Health Index or PSA density in clinically significant prostate cancer detection in men with multiparametric MRI performed, and further reduced unnecessary biopsies in PI-RADS 3 lesions.