Immune profiling of metastatic uveal melanoma and response to immune checkpoint inhibitors.

Abstract

9565 Background: Response to immune checkpoint inhibitors (ICI) in uveal melanoma (UM) is low. We aimed to elucidate tumor markers correlated with improved survival in ICI treated UM patients. Methods: Tumor samples of UM patients were tested at Caris Life Sciences (Phoenix, AZ) with NextGen Sequencing on DNA (592 genes assay or whole exome sequencing) and RNA (whole transcriptome sequencing). Somatic mutations were totaled to calculate tumor mutational burden (TMB) and cutoff for high vs low was 10 mt/MB. PDL1 was tested with immunohistochemistry for tumor staining and cutoff was ≥2+, 5% for high vs low. NCOA2 gene amplification was considered a surrogate for gain of chromosome 8q (cutoff ≥6). Median RNA expression level for LAG3 was calculated for each cohort and used as cutoff for high vs low. All ICI treated patients were considered to have metastatic disease. Real-world overall survival (rwOS) was obtained from insurance claims data and calculated from tissue collection to last contact. Time on treatment (TOT) was calculated from start to finish of ICI treatment and was considered as surrogate for progression-free survival (PFS). Comparison of survival was performed by Kaplan-Meier analysis. Results: A total of 450 UM samples were analyzed. Of these, 108 were from ICI treated patients and were obtained from primary (10/108) or metastatic (98/108) sites. Most tumors were PDL1 low in the entire UM (86%, 240/279) and ICI treated (62%, 55/89) cohorts. There was no difference in TOT between PDL1 high vs low in ICI treated cohort (HR 1.46, 95% CI 0.82-2.6, median TOT 3.1 months vs 2.3 months). Similarly, 98% (257/263) of all UM samples had low TMB. ICI treated patients with high LAG3 expression had similar TOT compared to low (HR 1.3, 95% CI 0.59-2.9, median TOT 6 months vs 2 months). In the entire UM cohort, most tumors were NF1-wildtype (95%, 56/59). NF1-wildtype status was associated with a longer rwOS compared to NF1-mutated (HR 0.18, 95% CI 0.051-0.64, median rwOS of 20.8 months vs 7 months). NCOA2 amplification was associated with a worse rwOS as compared to patients without NCOA2 amplification in the entire UM (HR 0.68, 95% CI 0.50-0.91) but not in ICI treated cohort (HR 0.84, 95% CI 0.52-1.4). There was no difference in TOT in ICI treated patients by BAP1 and SF3B1 mutational status. Conclusions: UM lacks traditional markers of response to ICI. Short TOT seen in our study is consistent with PFS of 3 to 5.5 months seen in clinical trials. High LAG3 expression was associated with a clinically significant improvement in TOT. Traditional markers of poor prognosis were not implicated in survival differences in ICI treated patients. This likely represents a poor prognosis in all mUM patients regardless of traditional prognostic markers. NF1 mutation is uncommon in UM and its significance as a prognostic marker should be validated in a larger cohort. Ongoing research is needed to understand the biology of UM and approach to treatment.