PSMA PET/CT for Response Assessment and Overall Survival Prediction in Patients with Metastatic Castration-Resistant Prostate Cancer Treated with Androgen Receptor Pathway Inhibitors

We aimed to evaluate the role of prostate-specific membrane antigen (PSMA) PET/CT for response assessment and outcome prediction in patients with metastatic castration-resistant prostate cancer (mCRPC) treated with androgen receptor pathway inhibitors (ARPIs), including abiraterone acetate or enzalutamide. Methods: We retrospectively analyzed 30 ARPI-treated mCRPC patients who underwent ⁶⁸Ga-PSMA-11 PET/CT within 8 wk before (baseline) and 12 ± 4 wk after treatment initiation. Total PSMA tumor volume was calculated using the fixed threshold method (SUV ≥ 3). Patients were categorized as PSMA responders (PSMA-Rs) or PSMA nonresponders (PSMA-NRs) on the basis of both European Association of Urology/European Association of Nuclear Medicine (EAU/EANM) criteria and Response Evaluation Criteria in PSMA PET/CT (RECIP) 1.0. PSMA-R included patients with a complete response, a partial response, or stable disease, and PSMA-NR included those with progressive disease. On the basis of prostate-specific antigen (PSA), patients were classified as biochemical responders if PSA decreased by at least 50% and as nonresponders if it did not. The Φ-coefficient was used to evaluate the correlation of PSMA- and PSA-based responses. Survival analysis was performed using the Cox regression hazard model and the Kaplan–Meier method. Predictive accuracy was tested for both response criteria. Results: On the basis of PSMA PET/CT, 13 (43%) patients were PSMA-NR according to the EAU/EANM criteria and 11 (37%) patients were PSMA-NR according to RECIP 1.0. Significant correlations were observed between PSMA- and PSA-based responses for both criteria (Φ = 0.79 and 0.66, respectively). After a median follow-up of 25 mo (interquartile range, 21–43 mo), the median overall survival was significantly longer for PSMA-R than PSMA-NR (54 vs. 22 mo) for both the EAU/EANM criteria and RECIP 1.0, with hazard ratios of 6.9 (95% CI, 1.9–26; P = 0.004) and 5.6 (95% CI, 1.69–18.26, P = 0.005), respectively. No significant difference in predictive accuracy was found between the 2 criteria (C-index, 0.79 vs. 0.76, respectively, P = 0.54). Flare phenomena at the second PSMA PET study were not observed in our cohort. Conclusion: Our results demonstrate that PSMA PET/CT is a valuable imaging biomarker for response assessment and overall survival prediction when performed at 3 mo after ARPI treatment initiation in mCRPC patients. Both proposed PSMA response criteria (EAU/EANM and RECIP 1.0) seem to perform equally well. No PSMA flare was observed. Prospective validation of these findings is strongly needed.