Additivity and the efficacy of immune checkpoint blockade-based combination regimens in urothelial cancer.

Abstract

779 Background: Immune checkpoint blockade (ICB)-based combination regimens have transformed the treatment landscape of advanced solid tumors including urothelial cancer. Understanding whether such combinations are synergistic, additive, or less than additive is crucial for (a) refining combination therapy in urothelial cancer, (b) dissecting the role of concurrent versus sequential treatment, and (c) identifying which cytotoxic agents combine most favorably with ICB. Here we analyzed phase III trials in advanced urothelial cancer to determine if the efficacy of combination therapies were predictable by additivity or exhibited non-additive effects. Methods: We analyzed progression-free survival (PFS) from phase III trials of ICB-containing regimens for advanced urothelial cancer (Table). PFS distributions of constituent drugs were used to calculate the predicted PFS of combination therapy under the null hypothesis of additive PFS times (Hwangbo et al, Nature Cancer, 2023). For KEYNOTE-361 and IMvigor130, which allowed investigator’s choice of cisplatin or carboplatin, separate predictions were made for cisplatin- and carboplatin-treated cohorts. The Cox Proportional Hazards model tested whether observed PFS significantly differed from additivity. Results: PFS of combination therapies in EV-302, CheckMate901, and DANUBE were statistically indistinguishable from additivity (Table). Trials using choice of cisplatin or carboplatin (KEYNOTE-361, IMvigor130) showed that ICB with cisplatin produced additive PFS, but ICB with carboplatin was significantly antagonistic (inferior to additive PFS; P = 0.00004). Conclusions: Recently approved regimens enfortumab vedotin + pembrolizumab, and nivolumab + gemcitabine-cisplatin, were as effective as predicted by additivity, demonstrating that combining individually effective agents without synergy can make effective regimens. In advanced urothelial cancer, PD-1 or PD-L1 inhibitors are antagonistic with carboplatin, but confer additive PFS benefits with cisplatin. Modeling the PFS of combination regimens from single agent data could have predicted the success of most contemporary phase III trials in advanced urothelial cancer. Trial Combination Predicted HR for PFS (95% CI) Observed HR for PFS (95% CI) P value EV-302 Enfortumab vedotin + Pembrolizumab 0.49 (0.44 - 0.55) 0.45 (0.38 - 0.54) 0.93 CheckMate901 Nivolumab + Gemcitabine-Cisplatin 0.69 (0.60 - 0.79) 0.72 (0.59 - 0.88) 0.78 DANUBE Durvalumab + Tremelimumab 0.87 (0.78 - 0.98) 0.88 (0.75 - 1.03) 0.92 KEYNOTE-361 Pembrolizumab + Chemotherapy 0.67 (0.59 - 0.77) 0.78 (0.65 - 0.93) 0.02 IMvigor130 Atezolizumab + Chemotherapy 0.61 (0.55 - 0.68) 0.82 (0.70 - 0.96) 0.0006 Pooled ICB + Gemcitabine-Cisplatin 0.65 (0.56 - 0.74) 0.71 (0.58 - 0.86) 0.78 Pooled ICB + Gemcitabine-Carboplatin 0.63 (0.57 - 0.70) 0.84 (0.73 - 0.97) 0.00004