Eur Urol Oncol. 2026 Apr 10:S2588-9311(26)00062-3. doi: 10.1016/j.euo.2026.03.003. Online ahead of print.
ABSTRACT
BACKGROUND AND OBJECTIVE: Cancer of the Prostate Risk Assessment (CAPRA) score and its postsurgical variant (CAPRA-S) predict biochemical recurrence (BCR) after radical prostatectomy (RP) but do not capture tumor biology. We developed two integrated clinical scores, CAPRA-G and CAPRA-SG, combining CAPRA or CAPRA-S with Genomic Prostate Score (GPS).
METHODS: We analyzed 955 patients with Oncotype DX testing before RP. Associations between GPS, CAPRA, and CAPRA-S scores, and BCR were assessed using Cox models. CAPRA-G and CAPRA-SG were derived from β-coefficient weighting. Model discrimination was evaluated using Harrell’s C-index (3000 bootstrap resamples); differences were assessed with a nonparametric U-statistic. Decision-curve analysis compared CAPRA with CAPRA-G and CAPRA-S with CAPRA-SG.
KEY FINDINGS AND LIMITATIONS: Median follow-up was 60 mo (IQR 36-72). GPS predicted BCR independent of CAPRA or CAPRA-S (p < 0.001). Using log-hazard β-coefficients, one CAPRA and CAPRA-S point corresponded to » 6 and 10 GPS points, respectively, yielding CAPRA-G = CAPRA + GPS/6 and CAPRA-SG = CAPRA-S + GPS/10. Compared with CAPRA and CAPRA-S, CAPRA-G and CAPRA-SG showed a statistically significant improvement in C-indices, increasing from 0.67 to 0.76 (p < 0.001) and from 0.81 to 0.84 (p = 0.041), respectively. Decision-curve analysis demonstrated higher net benefit for CAPRA-G and CAPRA-SG at clinically relevant thresholds (» 0.05-0.45). Limitations included lack of external validation and under-representation of high-risk disease due to selective GPS prescription.
CONCLUSIONS AND CLINICAL IMPLICATIONS: GPS integration enhances BCR risk stratification in the RP setting. CAPRA-G and CAPRA-SG scores quantify genomic risk and translate GPS into simple, clinically applicable tools with improved clinical utility. External validation is warranted.
PMID:41966955 | DOI:10.1016/j.euo.2026.03.003
