External validation of the P-Score in the prospective, pre-biopsy DEPROMP cohort: integrated mpMRI and PSMA PET/CT risk stratification for prostate cancer detection.

Abstract

This study aimed to externally validate the P-Score, a composite scoring system combining PI-RADS and PRIMARY score, for its diagnostic accuracy in detecting clinically significant (ISUP ≥ 2) and higher-grade (ISUP ≥ 3) prostate cancer (PCa). 230 biopsy-naïve men with suspected PCa were prospectively enrolled in the DEPROMP trial and underwent multiparametric MRI and [⁶⁸Ga]Ga-DKFZ-PSMA-11 PET/CT. Lesions were assessed using PI-RADS v2.1, the PRIMARY score, and the derived P-Score. Targeted biopsies were guided by imaging findings, and histopathology served as the reference standard. Diagnostic performance was evaluated using receiver operating characteristic curve analyses and DeLong's test to compare the area under the curve (AUC) between scoring systems. For clinically significant PCa (ISUP ≥ 2), no significant difference in diagnostic performance was observed between the P-Score (AUC 0.796) and PI-RADS (AUC 0.804; p = 0.72), whereas the P-Score significantly outperformed the PRIMARY score (AUC 0.730; p < 0.001). Similarly, for detection of ISUP ≥ 3 disease, no significant difference in diagnostic performance was observed between the P-Score (AUC 0.840) and PI-RADS (AUC 0.830; p = 0.66). In the PI-RADS 3 subgroup, diagnostic performance remained limited, with AUCs of 0.573 (95% CI: 0.407-0.739) for the P-Score and 0.556 (95% CI: 0.389-0.722) for the PRIMARY score. Overall, the P-Score demonstrated robust diagnostic performance across disease categories and showed comparable accuracy to PI-RADS. The P-Score demonstrated diagnostic performance comparable to PI-RADS in this external validation cohort. While providing a standardized framework integrating anatomical and molecular imaging findings, further studies are warranted to determine whether it has a role in selected clinical scenarios and prospective research settings.