Partitioned overall survival: Comprehensive analysis of survival states over 4 years in CheckMate 9ER comparing first-line nivolumab plus cabozantinib versus sunitinib in advanced renal cell carcinoma (aRCC).

Abstract

4507 Background: Immunotherapy regimens can be associated with prolonged disease control after treatment discontinuation without the need for further anticancer therapy. An integrated, comprehensive partitioned survival analysis describes how patients spend overall survival (OS) time, on/off treatment, with/without toxicity. Previous analysis of first-line (1L) nivolumab (NIVO) + ipilimumab (IPI) for aRCC in CheckMate 214 showed treatment-free survival (TFS) twice as long vs sunitinib (SUN). TFS and survival states for immune checkpoint inhibitor plus tyrosine kinase inhibitor (TKI) combination are of interest. Methods: Overall, 651 patients (pts) with aRCC were randomized to receive 1L NIVO + cabozantinib (CABO) or SUN in the CheckMate 9ER trial; treatment continued until progression or intolerance, NIVO for at most 24 months. Minimum follow-up was 4 years. We partitioned area under the Kaplan–Meier (KM) OS curve into 3 survival states, defined from randomization: time on 1L protocol therapy, TFS, and survival after subsequent systemic therapy (second line [2L]) initiation. TFS and protocol therapy were subdivided as mean times with/without reported grade 2+ treatment-related adverse events (TRAEs). Areas under and between KM curves were estimated by 48-month restricted mean times to event. Bootstrapped 95% CIs for between-group differences are reported. Results: At 4 years post randomization, KM estimates of OS were 49.2% vs 40.2% of pts assigned to NIVO+CABO vs SUN, respectively; 17.6% vs 4.7% of pts were surviving treatment-free, and 15.8% vs 8.2% were continuing on 1L protocol therapy. Over the 48-month period since randomization (Table), the mean TFS (95% CI) was 2.4 (0.8–3.9) months longer after treatment with NIVO+CABO than SUN. At least half of TFS was spent with toxicity in both treatment groups, resulting in difference in mean TFS (95% CI) without toxicity of 0.7 (−0.4 to 1.8) month. The NIVO+CABO group spent a mean (95% CI) of 8.5 (6.2–10.8) months more survival time on 1L protocol therapy, whereas the SUN group spent a mean (95% CI) of 6.5 (4.4–8.6) months more survival time after 2L therapy initiation. Conclusions: Over the 4-year period since initiation of 1L therapy for aRCC, the longer OS achieved with NIVO+CABO was accompanied by 1.5 times longer mean time surviving treatment-free before 2L therapy compared with SUN, a smaller difference than seen for NIVO + IPI vs SUN. [Table: see text]