The effects of antibiotics on the efficacy of immune checkpoint inhibitors in patients with non–small-cell lung cancer differ based on PD-L1 expression

Background: Immune checkpoint inhibitors (ICIs) are essential for treatment of various malignancies, including non–small-cell lung cancer (NSCLC). Recently, several studies have shown that the gut microbiome plays an important role in ICI treatment of solid cancers, and antibiotic (ATB) use had a negative impact on the outcomes of ICI treatment via dysbiosis in the gut. However, whether this is applicable to NSCLC remains unclear. The impact of ATBs based on PD-L1 expression also remains unclear. Methods: We retrospectively reviewed the medical records of patients with NSCLC who received ICI monotherapy (anti–PD-1 or anti–PD-L1 antibody) at nine institutions from December 2015 to May 2018. Outcomes with use of ATBs during the 2 months before or a month after initiation of ICI treatment, including progression-free survival (PFS) and overall survival (OS), were investigated using the Kaplan-Meier method. Multivariate analysis was also conducted using a Cox proportional hazards model. Results: A total of 531 patients were included in this study, among whom 98 (18.5%) received ATBs before or after ICI treatment. ATB use was significantly associated with a shorter median OS (11.7 months in the ATB group vs. 16.1 months in the non-ATB group; p = 0.028), whereas the difference in PFS was not significant (3.5 months in both the groups; p = 0.287). We next investigated the association based on PD-L1 expression in the 265 patients for whom PD-L1 expression was determined. There was no significant difference in the median OS or PFS between patients with NSCLC and PD-L1 expression <50% receiving ATBs and those not receiving ATBs (PFS: 3.3 vs. 2.8 months, p = 0.88; OS: 9.5 vs. 17.1 months, p = 0.24). Conversely, patients with NSCLC and PD-L1 expression ≥50% receiving ATBs showed significantly shorter median PFS and OS (PFS: 4.2 vs. 9.4 months, p = 0.012; OS: 11.9 vs. 28.4 months, p = 0.011). The impact of ATBs in patients with NSCLC and PD-L1 expression ≥50% was more significant than that in the entire cohort. Conclusions: Our results indicate that the impact of ATB use on the efficacy of ICIs differed based on PD-L1 expression in patients with advanced NSCLC. A negative impact of ATB use was found in patients with NSCLC and PD-L1 expression ≥50% but not in those with PD-L1 expression <50%.