PrE0807: A phase Ib feasibility trial of neoadjuvant nivolumab (N) without or with lirilumab (L) in cisplatin-ineligible patients (pts) with muscle-invasive bladder cancer (MIBC).

4518 Background: Neoadjuvant cisplatin-based chemotherapy (CT) prior to radical cystectomy (RC) improves overall survival (OS) in MIBC, but about half of pts are cisplatin-unfit or refuse it. Neoadjuvant immune checkpoint inhibitors can induce high pathologic complete response rate (ypT0N0). The combination of anti-PD-1 (N) and anti-KIR (L) is hypothesized to be safe and have significant activity based on the complementary and possibly synergistic roles in regulating adaptive and innate immune response in MIBC. Methods: This is a phase Ib multi-institutional trial in pts with localized MIBC treated with 2 neoadjuvant doses (4 weeks apart) of N alone (480 mg) in cohort 1 or N (480 mg) + L (240 mg) in cohort 2 prior to RC without adjuvant therapy (NCT03532451). Cohorts were enrolled sequentially and were not randomized. Key eligibility criteria included stage cT2-4aN0-1M0, ≥20% tumor content at TURBT and cisplatin-ineligibility (Galsky criteria) or refusal. Primary endpoint was safety manifested as rate of ≥G3 treatment related adverse events (TRAE) assessed in each cohort with CTCAE v5.0. Key secondary endpoints included the % of pts who had RC > 6 weeks after last neoadjuvant dose due to TRAE, CD8+ T cell density at RC, ypT0N0 and < ypT2N0 rates, CD8+ T cell density change between TURBT and RC, recurrence-free survival (RFS) and biomarkers in tumor tissue, blood and urine. Results: Among 43 pts enrolled (13 cohort 1, 30 cohort 2), median age was 75 (51-89), 67% were men, all had PS ECOG 0-1. Pts were cisplatin-ineligible due to impaired renal function (47%) and hearing loss (37%), while 14 % refused cisplatin. At baseline, 37 pts had cT2 stage, 2 had cN1 and 3 cNx. In cohort 1 and 2, 13 and 29 pts, respectively, completed intended neoadjuvant treatment, and 41/43 underwent RC (12/13 cohort 1, 29/30 cohort 2). One pt progressed to metastatic disease prior to RC (cohort 1) and 1 withdrew consent prior to being treated (cohort 2). Additionally, 1 patient was found to have cervical cancer at RC. Median time from last neoadjuvant dose to RC was 27 (95%CI: 24-29) days. There was no RC delayed > 6 weeks from treatment completion due to TRAE. G3 TRAEs occurred in 0% with N and 6.7% (90%CI 1.2-19.5%) in N+L (1: arthralgia, 1: gout, 2: hip pain) that all resolved. No G4/5 TRAEs occurred. Of 40 pts with MIBC and RC, ypT0N0 rates for N and N+L were 8% and 18%, while < ypT2N0 rates were 17% and 29%, respectively. Data on RFS and OS, and biomarker data were not yet mature. Conclusions: Neoadjuvant N alone and N+L combination prior to RC were safe, feasible and well tolerated in cisplatin-ineligible pts with MIBC, but ypT0N0 rates were unexpectedly low, especially with N alone. Two phase 3 trials (NCT03661320; NCT04209114) are evaluating the peri-operative role of N + chemotherapy +/- Linrodostat in cisplatin-fit and N +/- Bempeg in cisplatin unfit patients and are also assessing biomarkers. Clinical trial information: NCT03532451.