Prognostic and predictive value of baseline PSMA-PET total tumour volume and SUVmean in metastatic castration-resistant prostate cancer in ENZA-p (ANZUP1901): a substudy from a multicentre, open-label, randomised, phase 2 trial.

BACKGROUND: Quantitative parameters derived from gallium-68 [68Ga]Ga-prostate-specific membrane antigen (PSMA)-11 PET-CT (PSMA-PET-CT) such as whole-body standardised uptake value (SUV)mean and total tumour volume (PSMA-TTV) have shown prognostic value for response to lutetium-177 [177Lu]Lu-PSMA-617 monotherapy in patients with prostate cancer. Adding [177Lu]Lu-PSMA-617 to enzalutamide improved overall survival compared with enzalutamide in patients with metastatic castration-resistant prostate cancer in the ENZA-p trial. This prespecified substudy of ENZA-p evaluated baseline PSMA-PET quantitative parameters as predictive and prognostic biomarkers for enzalutamide plus [177Lu]Lu-PSMA-617 and enzalutamide monotherapy. METHODS: ENZA-p was an open-label, randomised, phase 2 trial done in 15 hospitals in Australia. Participants were aged 18 years or older with progressive metastatic castration-resistant prostate cancer who had not previously been treated with docetaxel or androgen receptor pathway inhibitors (abiraterone permitted) for metastatic castration-resistant prostate cancer, had [68Ga]Ga PSMA-PET-CT-positive disease, an Eastern Cooperative Oncology Group performance status of 0-2, and at least two risk factors for early progression on enzalutamide. Patients were randomly assigned (1:1) by a centralised, web-based system using minimisation with a random component to either enzalutamide 160 mg daily (oral) or enzalutamide 160 mg daily plus adaptive-dosed (two or four doses) intravenous [177Lu]Lu-PSMA-617 7·5 GBq every 6-8 weeks. The primary endpoint was prostate-specific antigen (PSA) progression-free survival, which has been reported previously. All participants underwent baseline [68Ga]Ga-PSMA-11 PET-CT to assess eligibility (SUVmax >15 at a single site and SUVmax >10 at all larger tumour sites). PSMA-PET parameters were quantified with semi-automated software to derive PSMA-TTV and SUVmean and correlated with overall and PSA progression-free survival in a prespecified analysis, with the primary endpoint of this substudy being overall survival. Thresholds were based on SUVmean highest quartile (Q4 vs Q1-3) and PSMA-TTV median at baseline. We used the Kaplan-Meier method and Cox regression models and analysed patients on a treatment received basis. The trial is registered with ClinicalTrials.gov, NCT04419402, and follow-up is complete. FINDINGS: Between Aug 17, 2020, and July 26, 2022, 162 participants were randomly assigned to enzalutamide (n=79) or enzalutamide plus [177Lu]Lu-PSMA-617 (n=83). This substudy included the 160 of the 162 randomly assigned patients who received study treatment (79 in the enzalutamide group and 81 in the enzalutamide plus [177Lu]Lu-PSMA-617 group). Median follow-up at the final data cutoff (July 31, 2024) was 34 months (IQR 29-39), with 96 overall survival events (53 with enzalutamide and 43 with enzalutamide plus [177Lu]Lu-PSMA-617). Baseline median SUVmean was 7·7 (IQR 6·5-9·8) and median PSMA-TTV was 234 mL (76-687). Median overall survival for PSMA-TTV below or above the median in the enzalutamide group was 39 months (95% CI 31-not estimable) versus 20 months (13-24; HR 0·23 [95% CI 0·13-0·42], log-rank p<0·0001). The corresponding median overall survival for PSMA-TTV below or above the median in the enzalutamide plus [177Lu]Lu-PSMA-617 group was 35 months (95% CI 32-37) versus 28 months (26-34; HR 0·66 [0·36-1·21], log-rank p=0·18). The test for interaction between PSMA-TTV and treatment group for overall survival was p=0·0078. Median overall survival for SUVmean Q4 versus Q1-3 in the enzalutamide group was 29 months (95% CI 17-39) versus 25 months (21-31; HR 0·84 [0·44-1·60], log-rank p=0·59). For enzalutamide plus [177Lu]Lu-PSMA-617, median overall survival for SUVmean Q4 versus Q1-3 was 32 months (95% CI 21-not estimable) versus 34 months (27-35; HR 0·80 [0·38-1·68], log-rank p=0·56). The test for interaction between SUVmean (Q4 vs Q1-3) and treatment group for overall survival was p=