Impact of the Nadir Androgen Receptor Signaling Inhibitor-Derived Integrative Response Model in Real-World Patients With Metastatic Castration-Sensitive Prostate Cancer.

BACKGROUND

Early deep prostate-specific antigen (PSA) decline after the initiation of androgen receptor signaling inhibitor (ARSI) therapy is associated with favorable long-term outcomes in metastatic castration-sensitive prostate cancer (mCSPC). The Nadir ARSI-Derived Integrative Response (NADIR) model was recently developed using Phase III trial data to predict the probability of achieving a favorable early PSA response. However, its performance in real-world clinical settings remains unclear.

METHODS

We retrospectively reviewed 196 patients with mCSPC who received first-line ARSI therapy at Hiroshima University Hospital and affiliated institutions between 2018 and 2025; 186 had complete data required to calculate the NADIR score. An early favorable PSA response was defined as a decline in PSA to ≤ 0.2 ng/mL within 6 months of treatment initiation. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUC), and calibration was evaluated using calibration plots. Patients were stratified into tertiles based on the NADIR score, and time-to-event outcomes, including overall survival (OS) and time to castration-resistant prostate cancer (CRPC), were analyzed using the Kaplan-Meier method.

RESULTS

The proportion of patients with an early favorable PSA response decreased stepwise across the upper, middle, and lower tertiles (61.3%, 33.8%, and 17.7%, respectively; Cochran-Armitage trend test, p < 0.01). The NADIR model demonstrated acceptable discrimination for predicting an early favorable PSA response, with an AUC of 0.74 (95% confidence interval [CI] 0.66-0.81). Agreement between predicted probabilities and observed outcomes was maintained across NADIR-defined risk groups. In a univariable analysis treating the NADIR score as a continuous variable, higher NADIR scores were strongly associated with an early favorable PSA response (odds ratio per 10% increase, 1.74; 95% CI 1.36-2.23; p < 0.001). CRPC-free survival differed significantly across NADIR-defined tertiles, whereas OS did not differ significantly.

CONCLUSIONS

In this real-world cohort, the NADIR model retained acceptable predictive performance, suggesting applicability in routine clinical practice and potential utility in individualized pretreatment clinical management.