Atezolizumab and Platinum Plus Pemetrexed With or Without Bevacizumab for Metastatic Nonsquamous Non–Small Cell Lung Cancer

Importance: The combination of an antibody to programmed cell death-1 (PD-1) or to its ligand (PD-L1) with chemotherapy is the standard first-line treatment for metastatic non-small cell lung cancer (NSCLC). Bevacizumab is expected to enhance the efficacy not only of chemotherapy but also of PD-1/PD-L1 antibodies through blockade of vascular endothelial growth factor-mediated immunosuppression, but further data are needed to support this. Objective: To evaluate the efficacy and safety of bevacizumab administered with platinum combination therapy and atezolizumab in patients with advanced nonsquamous NSCLC. Design, Setting, and Participants: An open-label phase 3 randomized clinical trial was conducted at 37 hospitals in Japan. Patients with advanced nonsquamous NSCLC without genetic driver alterations or those with genetic driver alterations who had received treatment with at least 1 approved tyrosine kinase inhibitor were enrolled between January 20, 2019, and August 12, 2020. Interventions: Patients were randomly assigned to receive either atezolizumab plus carboplatin with pemetrexed (APP) or atezolizumab, carboplatin plus pemetrexed, and bevacizumab (APPB). After 4 cycles of induction therapy, maintenance therapy with atezolizumab plus pemetrexed or with atezolizumab, pemetrexed, and bevacizumab was administered until evidence of disease progression, development of unacceptable toxic effects, or the elapse of 2 years from the initiation of protocol treatment. Main Outcomes and Measures: The primary end point was progression-free survival (PFS) as assessed by blinded independent central review (BICR) in the intention-to-treat (ITT) population. Results: A total of 412 patients were enrolled (273 men [66%]; median age, 67.0 [range, 24-89] years) and randomly assigned, with 205 in the APPB group and 206 in the APP group of the ITT population after exclusion of 1 patient for good clinical practice violation. The median BICR-assessed PFS was 9.6 months with APPB vs 7.7 months with APP (stratified hazard ratio [HR], 0.86; 95% CI, 0.70-1.07; 1-sided stratified log-rank test; P = .92). According to prespecified subgroup analysis of BICR-assessed PFS, an improved PFS with APPB vs APP was apparent specifically in driver oncogene-positive patients (median, 9.7 vs 5.8 months; stratified HR, 0.67; 95% CI, 0.46-0.98). Toxic effects related to bevacizumab were increased in the APPB group. Conclusions and Relevance: The findings of this trial did not show superiority of APPB over APP for patients with nonsquamous NSCLC; however, this regimen showed a similar tolerability and improved survival relative to APP in patients with driver oncogenes. Trial Registration: Japan Registry of Clinical Trials Identifier: jRCT2080224500.