Abstract Background: In recent years, attention has shifted to modification of tumor response to immunotherapy via the host microbiome. The mechanisms of these associations, causative or consequential, remain incompletely understood. We seek to explore this further with a longitudinal study of lung cancer patient microbiomes and immunotherapy-related adverse effects (AEs) preceding and during immunotherapy. Methods: Patients with lung cancer (LC) treated with immunotherapy (anti-PD-1/L1 agents including pembrolizumab, nivolumab, atezolizumab, and durvalumab) with or without chemotherapy at the University of Iowa from November 2018–April 2019 were consented for this ongoing study. Fecal samples and nasal and buccal swabs were obtained prior to therapy. Patients were treated and monitored per respective disease protocol; responses were recorded per RECIST 1.1 and AEs per CTCAE 5.0. Samples underwent DNA extraction followed by 16S rRNA metagenomic analysis and taxonomic profiling using Divisive Amplicon Denoising Algorithm (DADA)-2 pipeline. Microbiomes from all body sites were compared and correlated to treatment response and AEs. In addition, baseline LC gut microbiomes were compared to fecal samples provided by healthy controls (HC) in the same geographic location. This project is registered on clinicaltrials.gov (NCT03688347). Results: Gut microbiota significantly differed compared to oral and buccal microbiota in all patients. Gut microbiota from LC patients was compared to HC samples from the same geographic area. LC patients exhibited drastically different baseline composition, including dramatic increases in Firmicutes, Actinobacteria, and Verrucomicrobia, and significant decreases in Bacteroidetes, Proteobacteria, and Cyanobacteria. We noticed a clear inversion of Firmicutes/Bacteroidetes ratio between HC and LC patients, differences also reflected at the genus level. NSCLC patients experiencing immunotherapy-related adverse events were found to have at baseline markedly fewer Bifidobacterium and Desulfovibrio (p < 0.05 and FDR <0.05) when compared to those who did not experience AEs, regardless of grouping parameters (e.g., grade 0 vs. 1-4; 0-1 vs. 2-4; and 0 vs. 1-2 vs. 3-4). In addition, patients who responded to combined immunotherapy and chemotherapy (best response >= PR) exhibited enriched baseline Clostridiales (phylum Firmicutes, p=0.018) but reduced Rikenellaceae (phylum Bacteroidetes, p=0.016). Conclusion: Our study found promising trends in microbiome constitution. Compared to HC, we found significant differences in baseline LC gut microbiome at phylum and genus levels. More importantly, there are notable differences comparing LC patients who suffered AEs to those with none and between immunotherapy responders vs. nonresponders. Our project marks an important first step in a long-term study that could shed new light on the microbiome’s influence on immunotherapy treatment outcomes. Citation Format: Justin J. Chau, Meeta Yadav, Muhammad Furqan, Keri Mercer, Evan Eastman, Shailesh Shahi, Qun Dai, Carlos Chan, George Weiner, Taher Abu Hejleh, Ashutosh Mangalam, Jun Zhang. The microbiome in lung cancer under immunotherapy: Significant compositional differences associated with treatment response and AEs [abstract]. In: Proceedings of the AACR Special Conference on the Microbiome, Viruses, and Cancer; 2020 Feb 21-24; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2020;80(8 Suppl):Abstract nr B24.
Read full abstract