Rad2Nivo: Phase IB study of radium-223 (Rad) with nivolumab (Nivo) in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC).

5060 Background: Rad is an α-emitting calcium-mimetic radiopharmaceutical that is FDA approved for pts with mCRPC with symptomatic bone metastasis (mets) and no visceral mets. Preclinical data suggest that radiation might enhance the anticancer effect from PD-1 therapy [PMID 25274032, 29137877]. Nivo is a PD-1 immune checkpoint antagonist. Few pts experience PSA reduction with Rad, and objective responses in bone mets cannot be assessed. We hypothesized that Rad+Nivo would be safe and result in decreased ctDNA variant allele frequency (VAF) as a marker of clinical efficacy. Methods: In this phase IB, prospective, open-label, single-center, single-arm study (NCT04109729), pts with symptomatic bone-metastatic mCRPC without visceral mets were eligible. Participants received intravenous Rad (55 kBq/kg IV) monotherapy for 6 cycles. Nivo (480mg IV) was introduced starting with cycle 3 and continued for up to 2 years. ctDNA collected at baseline, cycle 3 and 6 weeks after. Primary endpoints: (1) frequency of grade ≥3 non-hematologic treatment-emergent adverse events (TEAEs) and (2) change in ctDNA VAF from baseline compared to after 6 weeks of nivolumab therapy. Mean VAF was defined as average VAF of all somatic mutations present at baseline at a VAF 1%. Null hypothesis: proportion of pts with a non-zero ctDNA reduction is ≤ 20%. A one-sided exact binomial test was used to evaluate whether the proportion of pts demonstrating a decrease in mean VAF on-treatment exceeded 20%. With 36 pts, there will be 83% power at one-sided α = 0.05 to detect an alternative proportion of subjects with ctDNA reduction = 40%. Results: Overall, 39 pts were enrolled trial between 9/2020 and 1/2025, and 35 pts were treated with both Rad+Nivo. Median age was 71 years (54-84), median PSA at study entry was 13.9 ng/dL (0-3245). 22 pts (56.4%) had Gleason score ≥8 (unavailable for 3 pts). Median number of prior lines of therapy was 3 (1-7). 53.9% (n = 21) of pts had received prior chemotherapy. 61% (n = 24) of pts completed 6 doses of Rad (range 2-6). The median number of Nivo cycles was 4 (0-25). Dose holds or delays occurred in 26% (n = 10) and 31% (n = 12) of pts receiving Rad and Nivo, respectively. 38% (n = 15) of pts experienced at least one grade 3-5 TEAE regardless of attribution. Grade 3-4 TEAE related to Rad and Nivo occurred in 15% (n = 6) and 18% (n = 7) of pts, respectively. No dose-limiting toxicities were observed. Serious adverse events related to Rad and Nivo occurred in 5% (n = 2) and 2.6% (n = 1) of pts, respectively. Paired molecular data was available for 25 pts. VAF was measurable in all patients who had samples collected and reduction was observed in 15 of 25 pts (60%, one-sided 95% CI, 41.7%-100%; p < 0.01). Conclusions: The study met its primary endpoints. Rad+Nivo appears to be safe and VAF was decreased in 60% of patients tested. Survival outcomes and clinical correlation data with VAF will be presented at the meeting. Clinical trial information: NCT04109729 .