The effect of the microbiome on immune checkpoint inhibitor toxicity in patients with melanoma.

9568 Background: Immune-checkpoint inhibitor (ICI) immunotherapy has increased survival in patients with melanoma. However, only half of the patients respond, and many experience immune-related adverse events (irAEs). Recent evidence suggests that modification of the gut microbiome may increase response to ICIs and decrease toxicity. Here we describe the first results of a clinical trial to determine if the microbiome can predict the response or toxicity during the first 16 weeks of ICI treatment. Methods: We enrolled patients aged 18 or older in a prospective observational cohort study at The Ohio State University Comprehensive Cancer Center Skin Cancer Clinic (OSUCCC-SCC) who were to receive treatment with pembrolizumab or nivolumab alone or in combination with other treatments (e.g. nivolumab and ipilimumab) for melanoma. Patients receiving systemic or oral corticosteroids at the start of ICI cycle 1 were excluded but were eligible if receiving adrenal physiologic replacement. Patients collected stool samples at baseline, within 2 days of an adverse event (if applicable), and at 12 weeks. The response to ICIs was evaluated by Response Evaluation Criteria in Solid Tumors (RECIST v1.1) at a 12-week computed tomography scan. Metagenomic whole-genome shotgun sequencing was performed on an Illumina NovaSeq 6000 and then classified using HUMAnN3. The effect of microbe relative abundances on potential irAEs was modeled by logistic regression with the R package glmm. Results: In total, 88 patients consented to the trial. Pre-treatment microbiome samples were collected from 49 patients. Potential irAEs were observed in 16 out of the 49 patients for whom pre-treatment microbiome samples were collected. There was no significant difference in the ages (p = 0.150, genders (p = 0.2), stages (p = 0.2) or treatments (p = 0.07) of those who developed potential irAEs. Pretreatment abundance of the family Ruminococaceae was most strongly associated with the development of a potential irAE (p = 0.03), followed by a taxon in an unclassified order within the phylum Firmicutes (p = 0.05). The family Bacteroidaceae was most strongly associated with no potential irAE (p = 0.05). Conclusions: Longitudinal and event-driven biospecimen collection in the context of treatment with immunotherapies was feasible in the OSUCCC-SCC. The abundance of the two high-taxonomic rank microbe groups was significantly associated with potential irAEs. The association with Ruminococaceae is consistent with previous studies where it was associated with response to ICIs and, in separate studies, development of an irAE was associated with a better response. The unclassified taxon is potentially a new biomarker for the prediction of toxicity and a therapeutic target to reduce treatment side effects. Future analyses will associate microbes with treatment response and test for consistent microbiome changes at the time of irAE development. Clinical trial information: NCT05102773.