Association of antibiotic therapy and treatment efficacy in urothelial cell carcinoma patients receiving immune checkpoint inhibitors.

Abstract

4578 Background: The content of gut microbiome has been linked to the effectiveness of immune checkpoint inhibitory (ICI) therapy in various cancer types. Antibiotic therapy (ABT) induces significant alterations in gut microbiota; however, only a limited number of single-institutional studies have examined the impact of ABT on ICI treatment. Utilizing a national database, we sought to provide ‘real world outcomes’ by investigating the association between ABT and ICI on overall survival (OS) and real-world progression free survival (rwPFS) in advanced urothelial carcinoma (aUC). Methods: We used the electronic health record (EHR)-derived de-identified FlatIron Health (FIH) database after Institutional Review Board approval was obtained, and included a waiver of informed consent to select 1491 subjects with aUC who received ICI therapy. Patient characteristics were described and compared between those with known ABT use and without known ABT use (in the 3 months pre/post ICI initiation), using ANOVA and Chi-square tests. rwPFS and OS were calculated from date of initiation of ICI to date of real-world progression/date of death. 3-month landmark Kaplan Meier methods and log-rank tests were used to estimate and compare rwPFS and OS between patients with-ABT and without-ABT, as well as patients with ABT use pre-ICI initiation, post-ICI initiation, and no known ABT use. Cox proportional models were used to investigate the association between rwPFS, OS, and antibiotic use, adjusting for patient characteristics. Results: In patients who received ICI, we included 407 patients who received ABT and 1084 who did not. Those who received ABT were younger (p < 0.001), more likely Caucasian (p = 0.046) and insured by Medicare (p < 0.001). rwPFS (median: 9.0 vs 10.0 months; p = 0.41) and OS (median: 14.0 vs 16.0 months; p = 0.073) were not significantly different between the with or without ABT subgroups. We included 101, 182, and 715 patients who received ABT prior to ICI, after ICI initiation, or did not receive ABT. When assessing timing of ABT, there was a significant difference in OS amongst pre-ICI initiation ABT, post-ICI initiation ABT and no known ABT groups (median: 17.0 vs 10.9 vs 16.0; p = 0.005). However, rwPFS (median: 10.1 vs 7.5 vs 10.0 months; p = 0.22) was not statistically different. Multivariable analysis identified no antibiotic use as a significant predictor of improved OS (HR 0.733, CI 0.598-0.898; p = 0.003). Other factors associated with OS were male gender (HR 1.158, Cl 1.005-1.335; p = 0.043) and receipt of extirpative surgery (HR 0.785, Cl 0.684-0.900; p < 0.001). Conclusions: This study identifies a potential negative association of ABT and OS, particularly after starting ICI. These results support the relevance of the gut microbiome on the efficacy of ICI treatment, may guide future efforts to improve the efficacy of ICI, and emphasize the importance of antibiotic stewardship.