1035P Effect of gut microbiota on immunotherapy of advanced NSCLC

Abstract

Background:The microbiota community is considered as an organ of the human body. Recent studies have found that gut microbiota may impact on the interaction between immune regulation and tumor treatment. The aim of this study is to characterize the gut microbiota in advanced NSCLC patients, and its relationship with response to immune checkpoint blockade (ICB).Methods: Stool samples from 84 advanced NSCLC patients were collected prior ICB as first or second line treatment. 16S rRNA gene sequencing and SILVA_release_132 database were used for taxonomic profiling. Diversity indices, including Chao1, Shannon and Inverse Simpson index (IDI), abundance of certain taxa, clinicopathological characteristics, dietary habits and antibiotic usage were evaluated for association with clinical benefit (CB) [complete or partial response, stable disease vs. progressive disease (PD), according to RECIST1.1]. Continuous variables were stratified into high or low using median as cut-off. For survival analysis, Cox Regression and Kaplan Meier curves with log-rank test were performed.Results: From 84 NSCLC patients, 60 presented PD-L1 positive tumors and 39 were treated with ICB as first line. A total of 8872307 sequencing reads were obtained and clustered in 357 genera, being the most frequent Bacteroides (27.9%) and Alistipes (7.1%). Patients with CB exhibited higher IDI compared with those with PD (p¼0.004). Moreover, higher IDI was associated with prolonged progression-free survival (PFS) (p¼0.014). High abundance of Butyricimonas in the samples was correlated with increased RR (p¼0.01) and longer PFS (p¼0.011) compared to low abundance. On the other hand, high frequency of Dialister was associated with reduced RR (p¼0.002) and shorter PFS (p¼0.003). Finally, the expression of PD-L1 was associated with increased response rate (RR) (p¼0.020). Conclusions:This study shows that diversity of gut microbiota is related to the ICB treatment response. In addition, high abundance of Butyricimonas and low abundance of Dialister could be considered as potential predictors of clinical benefit and PFS. Further analyses are being undertaken to find a compositional signature with predictive value. Funded by CB16/12/00350 from CIBEROnc, Arnal Planelles Foundation, AMACMA, GIDO group and PI18/00226 from ISCIII.