Comparing Prostatype P-score and traditional risk models for predicting prostate cancer outcomes in Spain

Abstract

Introduction

Prostate cancer (PCa) shows varied aggressiveness, complicating personalised treatment decisions. Traditional risk stratification systems rely on clinical parameters but may miss crucial genetic insights. The Prostatype® score (P-score) integrates gene expression with clinical data to improve PCa risk assessment precision.

Objectives

To validate the P-score’s predictive performance for prostate cancer-specific mortality (PCSM) and metastasis in a Spanish cohort, comparing it with NCCN, D’Amico, and EAU systems.

Materials and methods

This study was multicentre, retrospective and included seven Spanish hospitals. Of 154 core needle biopsies, 93 met RNA criteria, and for those, P-score was calculated based on IGFBP3, VGLL3, and F3 genes expression and clinical data.

The primary endpoint was PCa-specific mortality (PCSM), with secondary endpoints being development of metastasis, adverse pathology (AP), and International Society of Urological Pathology (ISUP) grading.

Results

The P-score demonstrated superior accuracy in predicting 10-year PCSM, with an AUC of 0.81 and a C-index of 0.75, outperforming NCCN (AUC 0.77, C-index 0.69) and D’Amico/EAU (AUC 0.70, C-index 0.62). For metastasis prediction, the P-score achieved a C-index of 0.77, significantly higher than NCCN, D’Amico, and EAU (0.58). Kaplan–Meier analysis underscored the P-score’s ability to better stratify patients by risk, especially high-risk groups. Additionally, the P-score correlated with tumour burden, showing significant associations with positive biopsy cores (p = 0.017) and ISUP grade at radical prostatectomy (p = 0.0028).

Conclusions

In this Spanish cohort, the P-score outperformed traditional clinicopathological systems in predicting PCSM, development of metastasis, and pathological markers, supporting its clinical utility for more personalised PCa management.