Background: MCM is a rare melanoma subtype (only 1% of melanomas in the US). MCM has a lower tumor mutational burden than cutaneous melanoma (CM). While some patients (pts) with MCM respond to immune checkpoint inhibitor (ICI) therapy, predictive markers of response have not been established. We analyzed a cohort of MCM from pts treated with ICI to identify gene expression signatures associated with tumor response and clinical outcome. Methods: Fifty-eight MCM specimens were collected from 3 institutions. RNA was extracted from FFPE tissue slides and analyzed by NanoString 770 Immune Profiling Panel. Gene expression profiles were linked to anatomical location and disease outcome after ICI therapy: response as defined by RECIST v1.1 and median overall survival (mOS). Results: Fifty-one pts were treated with ICI - anti-CTLA-4 (n = 9), anti-PD1 (n = 38), or both (n = 5) ) and had tumor response evaluation. Three were without response data, 2 were without disease recurrence after surgery, 2 did not receive ICI. Among 51 pts with response data, seven were without long-term follow-up (1CR, 2PR, 3SD). The overall response rate (ORR) was 40.3%, similar to the prior study (Shoushtari et al, Cancer 2016). A signature involving differential expression of 87 immunoregulatory genes correlated with tumor response: ORR: 75% (12/16) signature high vs. 33.3% (7/21) signature low (p = 0.02, high vs. low) vs. 14.3% (2/14) signature average (p < 0.01; high vs. average). mOS for the whole population was 12.4 months. Pts with increased gene signature expression had superior mOS: signature-high: Not reached, signature-low: 8.2 months, (HR: 0.2; 95%CI: 0.07-0.55, p < 0.01). Transcript pathway analysis of the gene signature showed association with chemokine receptors, interleukin-10 signaling, and Treg development. Conclusions: We have identified a gene expression signature that involves chemokine receptors, IL-10 signaling, and Treg development gene sets, that appears to predict for tumor response and mOS in pts with advanced MCM treated with ICI. Further validation of these gene signatures is underway.