Brief Report: Exploratory Analysis of Maintenance Therapy in Patients With Extensive-Stage SCLC Treated First Line With Atezolizumab Plus Carboplatin and Etoposide

Introduction: In the phase 1/3 IMpower133 study, atezolizumab plus carboplatin and etoposide (CP/ET) followed by maintenance atezolizumab for first-line treatment of extensive-stage SCLC (ES-SCLC) led to improvement in both overall survival (OS) and progression-free survival (PFS) versus placebo plus CP/ET followed by maintenance placebo. We explored the benefit of atezolizumab versus placebo in the subset of patients who reached the IMpower133 maintenance phase and the safety profile of maintenance therapy. Methods: Patients with untreated ES-SCLC were randomized 1:1 to four 21-day cycles of CP/ET with atezolizumab or placebo, followed by maintenance atezolizumab or placebo. The primary end points were OS and investigator-assessed PFS. A multivariate Cox model from the start of maintenance treatment was used to evaluate the treatment effect and account for lead-time bias; a generalized linear model was used to identify prognostic or predictive characteristics for reaching the maintenance phase. Results: A similar proportion of patients in each arm received at least the first dose of maintenance therapy (atezolizumab: 77%, n = 154 of 201; placebo: 81%, n = 164 of 202) and were included in the maintenance analysis population. An Eastern Cooperative Oncology Group performance status of 0 and absence of liver metastases at baseline were identified as prognostic factors for reaching the maintenance phase. The positive treatment effect with atezolizumab remained after adjusting for baseline characteristics. Median OS and PFS from the start of maintenance therapy in the atezolizumab versus placebo arm were 12.5 versus 8.4 months (hazard ratio = 0.59, 95% confidence interval: 0.43–0.80) and 2.6 versus 1.8 months (hazard ratio = 0.63 [95% confidence interval: 0.49–0.80]), respectively. Treatment-related adverse events from the start of maintenance therapy occurred in 41% (n = 64 of 155) and 25% (n = 41 of 163) of safety-evaluable patients in the atezolizumab and placebo arms, respectively, and were grade 3 or 4 in 28% (n = 43 of 155) and 23% (n = 37 of 163) of the respective populations; no patient in the atezolizumab arm and one patient in the placebo arm had a grade 5 treatment-related adverse event. Conclusions: These data in the context of other immunotherapy trials in ES-SCLC suggest that induction with atezolizumab plus CP/ET and maintenance treatment with atezolizumab are important components that contributed to the OS benefit observed in IMpower133. Safety results from randomization and from the start of maintenance therapy were similar between the treatment arms despite the continuation of atezolizumab in the maintenance phase.