ntegrated Prognostic Score in Metastatic Castration-resistant Prostate Cancer Treated with Cabazitaxel - A CABASTY Post Hoc Analysis Validated by Two International Prospective Phase 3 Trials

Background and objective
Optimal therapy after docetaxel and an androgen receptor pathway inhibitor (ARPI) in metastatic castration-resistant prostate cancer (mCRPC) remains debated. We aimed to build and validate the cabascore, a prognostic and predictive score for patients with mCRPC receiving cabazitaxel, to guide treatment strategy.
Methods
In this post hoc analysis of the randomized CABASTY trial, the cabascore was derived from multivariable Cox model hazard ratios for overall survival (OS). The population was stratified into three groups using log-rank optimization for score thresholds. Data from the phase 3 CARD and PROSELICA clinical trials were used for validation.
Key findings and limitations
A total of 194 patients from CABASTY were included. Model variables were baseline Eastern Cooperative Oncology Group performance status (0 versus 1–2), metastatic sites (<2 vs ≥2), prostate-specific antigen, lactate dehydrogenase, alkaline phosphatase, and hemoglobin. Median OS was 21.6, 12.5, and 6.2 mo for cabascore-determined favorable, intermediate, and poor prognostic subgroups, respectively (intermediate versus favorable: hazard ratio (HR) = 2.14; 95% confidence interval (CI) = 1.41–3.25; p < 0.001; poor versus favorable: HR = 5.41; 95% CI = 3.42–8.72; p < 0.001). Validation of the model in the PROSELICA (n = 1175) and CARD (n = 126) cabazitaxel-exposed populations confirmed significantly prolonged median OS in the favorable versus intermediate and poor groups (both comparisons p < 0.001). In the overall CARD population (n = 250), cabazitaxel improved OS versus ARPI in the favorable group (p = 0.003), but not in the intermediate or poor groups.
Conclusions and clinical implications
The cabascore may help tailor treatment strategy in patients with mCRPC. Exploratory analyses suggest a greater benefit of cabazitaxel in patients with a favorable prognosis versus a second-line ARPI.