Microbiome, socioeconomic status (SES), fiber intake, and response to checkpoint inhibitor therapy in patients with non-small cell lung cancer (NSCLC): A pilot study.

Abstract

e14710 Background: Lung cancer affects a disproportional number of socioeconomically disadvantaged patients (pts). Responses to checkpoint inhibitor therapy (CIT) vary. Microbiome diversity may correlate with benefit from CIT. High fiber diets lead to microbiome diversity. We hypothesize that unexplained lung cancer disparities may be in part due to differences in microbiome diversity determined by unhealthy dietary patterns dictated by lower SES. Methods: Following IRB review, 25 pts with metastatic NSCLC starting CIT-based treatment were enrolled. Pre-treatment oral and fecal samples were collected. 24-hour dietary recalls were obtained. SES index was calculated based on education, employment, housing and annual income. High fiber intake was defined as > 20 g/day. Best treatment response per RECIST 1.1 was assessed. Microbial 16s rRNA were sequenced with Illumina MiSeq platform. Sequence analyses utilized QIIME 2 with DADA2 plugin. Taxonomic assignments used the Greengenes database. Microbiome composition was analyzed by calculating alpha diversity (Chao1, Shannon diversity and Simpson index), beta diversity and differential abundance of microbial compositions. Hierarchical linear model was constructed for each SES variable with and without biological covariates. Results: Pts with samples >1000 operational taxonomic unit (OTU) levels were included (N = 22) in the analysis. 50% were male, median age was 71, 36% (8/22) were of Black race and 23% (5/22) were of low SES. Median fiber intake was 11.9 grams. Males had higher alpha diversity than females (p=.038). Pts with high fiber intake had greater fecal Proteobacteria compared to low fiber intake pts (p<0.05). Bacteroides and Firmicutes fecal levels were significantly different for all pts compared to Human Microbiome Project (HMP) reported normal levels (p<0.01). Actinobacteria and Proteobacteria levels for low-fiber intake pts were also significantly elevated compared to HMP normal levels (p<0.05). There were no differences observed in microbiome when comparing SES index. However, high SES index patients had elevated fecal levels of Actinobacteria, Firmicutes and Proteobacteria and lower levels of Bacteroides compared to HMP levels. There was a trend towards higher alpha diversity of oral samples in partial responders pts compared to those with stable or progression of the disease (p=0.08). Conclusions: Pts with metastatic NSCLC exhibited different microbial composition compared to HMP normal levels. Associations between SES and microbial distribution in such pts warrant further investigation. Increased sample size is needed to better determine differences in oral microbiome in responders to immunotherapy and better evaluate the role of sex in this setting.