Gut microbiome/metabolome predicts response to immune checkpoint blockers (ICB) in patients with recurrent metastatic head and neck squamous cell cancer (RM HNSCC).

6055 Background: Gut microbiome has emerged as an important predictor of response to ICB therapy in various cancers, most notably in melanoma. Due to functional redundancy of microbiota, there has been lack of consistency in gut microbial signature associated with ICB response. Microbial metabolites in addition to taxonomy may play a superior role in predicting response to ICB. Gut microbiome/metabolome signatures of ICB response is unknown in RM HNSCC patients. Methods: Two cohorts of patients were included (stool and plasma samples were collected)- cohort 1, newly diagnosed RM HNSCC patients starting first line ICB, for whom samples were collected before ICB initiation (baseline) and post-treatment at 3, 6, and 12 months. Cohort 2 with durable response (disease control lasting ≥ 6 months), with single sample collected at study entry. 16s rRNA sequencing was performed on stool samples while plasma metabolites were quantitated using Q Exactive plus Orbitrap mass spectrometer. Response was defined as partial response (PR) or complete response (CR) as per RECIST 1.1 Results: The 16-S sequence was carried on 31 samples (cohort 1- 16 baseline, 7 post treatment; cohort 2- 8 durable responders). Targeted metabolomics was completed on 92 plasma samples [cohort 1-27 baseline and 50 post-treatment, cohort 2- 15]. Responders had a significantly higher Shannon’s diversity index, lower Fermicutes to Bacteroidetes ratio, and were enriched with genus Bacteroides and Lachnospiracea incerate sedis at baseline and post treatment, compared to non-responders (p < 0.05). At species level, baseline and post treatment microbiome of responders was enriched with Eubacterium oxidoreducens and Bacteroides uniformis. Ruminococcus was preferentially enriched in durable responders. Targeted analysis of plasma metabolites (associated with gut microbial metabolism) showed that responders had a significantly lower baseline adenosine, Inosine and xanthine level as compared to non responders. Further, Inosine levels decreased with response, while levels increased in non-responders (p < 0.05), suggestive of consumption by re-activated T cells Further, ICB responders had significantly lower Kynurenine to tryptophan ratio compared to non responders Conclusions: This is the first study evaluating association of gut microbiome and metabolome on response to first line ICB, in RM HNSCC patients. We found higher diversity and specific gut microbial signatures associated with ICB response. Interestingly, we found that inosine and kynurenine/tryptophan pathways, both which play a crucial role in host as well as gut microbial metabolism were differentially expressed in ICB responders. Our results if validated in larger cohort, lays groundwork for gut microbiome and importantly microbial metabolite modulation to improve response to ICB in RM HNSCC.