Impact of steroid dose and timing on efficacy of combination PD-1/CTLA-4 blockade
Nina B. Curkovic, Rebecca Irlmeier, Xue Bai, Can Cui, Fei Ye, Hannah Burnette, Aleigha Lawless, Juliane Czapla, Ryan J. Sullivan, Douglas B. Johnson
Abstract
With the increasing use of immune checkpoint inhibitors (ICIs) in combination regimens and in earlier stages of advanced melanoma, the effective management of immune-related adverse events (irAEs) is key to balancing immunotherapy efficacy and toxicity. Conflicting evidence exists on possible detrimental effects of immunosuppression with corticosteroids for irAEs on ICI effectiveness. We conducted a multicenter, retrospective cohort study of immunotherapy-naïve advanced melanoma patients undergoing treatment with ipilimumab and nivolumab and a small cohort treated with nivolumab/relatlimab. We utilized univariate tests to assess response, PFS, and OS based on presence of irAE, receipt of steroids for irAEs, peak dose, and time-to-steroid, as well as multivariable analysis for response, OS, and PFS in patients receiving steroids for irAEs. Among 226 total ipilimumab/nivolumab patients, those without irAEs had poorer PFS and OS compared to irAE groups regardless of steroid administration. In subgroup analysis of patients receiving steroids for an irAE, increased time-to-steroid was significantly associated with improved response (aOR, 1.026 p = 0.0005), PFS (aHR, 0.986 p = 0.001), and OS (aHR, 0.983 p = 0.0008). Higher peak steroid dose was significantly associated with poorer PFS (aHR, 1.002 p = 0.005), and OS (aHR, 1.002 p = 0.003). Use of additional immunosuppressants was associated with poorer OS (aHR, 1.941 p = 0.018). Cumulative dose was not significantly associated with outcomes. Among 42 additional patients treated with nivolumab/relatlimab, irAEs were significantly associated with improved PFS/OS, which appeared to be slightly mitigated by steroid administration; dosing relationships were limited by small numbers.