Delineating the role of the microbiome and tumor microenvironment interactions driving mucosal melanoma (MM) response and resistance to immune checkpoint inhibitor (ICI) treatment.

Abstract

9561 Background: MM is a rare melanoma subtype with distinct biology and low response rates to ICIs. We seek to prospectively profile a cohort of MM patients to determine the interplay of gut, mucosal and tumor microbiome and modifiable risk factors on ICI response and resistance. Methods: Patients (pts) with MM presenting at MD Anderson are being prospectively enrolled for longitudinal collection for molecular, microbiome, and lifestyle factor profiling. Planned analyses will evaluate (1) fecal specimens [whole metagenome shotgun (WMS) sequencing, microbiome profiling and 16S rRNA gene sequencing (16S)]; (2) mucosal surface swabs (WMS sequencing); (3) formalin-fixed paraffin embedded (FFPE) tumor specimens [Bruker Digital Spatial Profiling (DSP)]; and (4) fresh tumor samples [whole exome (WES) and whole genome (WGS) of normal and tumor tissue, RNA sequencing, and T cell receptor sequencing]. Results: 82 pts with sequenced fecal specimens and mucosal surface swabs were enrolled as of 12/2024. MM primary sites included 25 (30%) naso-oral (including sinonasal), 14 (17%) urogenital, 30 (37%) anorectal, 10 (12%) conjunctival and 3 (4%) other. Disease stages were IVM1a/M1b in 9 (9%) pts, and IVM1c/M1d in 27 (33%) pts. Initial analysis of gut and mucosal surface swabs’ microbiome by MM primary site displayed a wide range of intrasample heterogeneity and microbial signatures that correlate with the MM primary site. WES and WGS data analysis indicate low tumor mutational burden (TMB) of 1.34mut/Mb (median, range 0.52 – 14.72). Common mutations included SF3B1 (23%), KIT (15%) and NRAS (8%). Anorectal and urogenital tumors contributed to mutational signatures associated with DNA mismatch repair and microsatellite instability (COSMIC v3.211). Conjunctival and naso-oral tumors showed an association with UV exposure. Results in the gut microbiome analysis from 78 pts treated with ICIs showed compositional differences in the presence and proportion of bacterial taxa in responders (R), compared to non-responders (NR). Distinct bacteria such as Streptococcus, Collinsella , and Blautia were identified in R, and Butyricicoccus in NR. DSP analysis by treatment response showed higher expression of CD56 and CD20 in the immune and tumor compartments in R. Conclusions: This is an ongoing prospective study that is expected to drive insights into the tumor/microenvironment/host interactions and factors regulating immunogenicity to predict response and resistance to ICIs in a rare and understudied melanoma subtype. Interrogation of the role of the gut microbiome and its modifiable determinants will lead to the investigation of new therapeutic strategies to modulate the microbiome to improve treatment outcomes in MM.