394 Background: Tools to risk stratify localized prostate cancer are limited. We sought to conduct a robust, individual participant data (IPD) meta-analysis of the performance of Prolaris from diagnostic biopsy in localized prostate cancer. Prolaris is based on a combined clinical risk (CCR) score that categorizes individual patient risk into low, intermediate, or high risk based on locked and validated active surveillance (AS) and multi-modal therapy thresholds. Methods: A systematic literature search was performed, and IPD were collected where possible to perform a two-step IPD analysis. The primary endpoint was a composite of distant metastasis (DM) and prostate cancer specific mortality (PCSM), also analyzed individually. Within each cohort Cox proportional hazards models were fit adjusting for treatment received. Random-effects meta-analyses with Knapp-Hartung adjustment were used to create combined hazard ratio (HR) estimates across studies. Results: Fourteen eligible studies included 8,480 total patients, of which 7,926 had IPD. The cohort consisted of 20.0%, 33.9%, 32.5%, and 13.6% NCCN Low-, Favorable Intermediate-, Unfavorable Intermediate-, and High-Risk disease, respectively. Initial management was 42.9% non-interventional (e.g. AS), 23.6% surgery, 16.4% radiation therapy (RT), and 13.0% RT plus androgen deprivation therapy. CCR was prognostic for composite DM-PCSM after accounting for treatment received (HR 2.28 (95% CI 1.92, 2.62), p=9.14x10-9) with insignificant heterogeneity (I2 =14%, p=0.3) and was also individually prognostic for DM (p=1.87x10⁻⁶) and PCSM (p=3.14x10⁻⁴). Influence analyses demonstrated that results were not materially influenced by any one study. Additional meta-analyses demonstrated that CCR adds independent prognostic information to Gleason, CAPRA, or NCCN (all p<10) and that Prolaris Risk Groups are prognostic for composite DM-PCSM, as well as for individual endpoints (all p<0.05). Conclusions: Prolaris improves prognostication across NCCN Risk Groups and treatment strategies in localized prostate cancer. Prognostic value persists after adjusting for initial treatments and established clinical factors, highlighting utility in supplementing conventional risk models.
Monda et al. (2026) studied this question.