Immune gene expression in primary melanomas to predict lower risk of recurrence and death.
3014 Background: Improved biomarkers are needed to define recurrence risk in patients with completely resected skin melanomas. Standard prognostic indicators used in staging, including depth, ulceration, and mitotic rate, while useful, often fail to accurately predict recurrence for individual patients. The immune system may prevent recurrence in this population, but no evidence-based immune biomarkers are in clinical use. Biomarker development has been hindered by clinical standards necessitating that the entire specimen be formalin fixed and paraffin embedded (FFPE) for morphology evaluation, a process damaging to RNA. Methods: To define a biomarker for melanoma recurrence, mRNA copy number of immune-related genes from FFPE melanoma was measured using NanoString, a hybridization assay suited for analysis of partially degraded RNA. Genes predictive of non-recurrence were defined using receiver operating characteristic (ROC) curves in a training cohort and then validated in an independent patient cohort. Results: A panel of 21 genes predictive of non-recurrence were defined using ROC curves in a training cohort (N=44). This result was validated in an independent patient cohort (N=37, AUC=0.794). Protein levels of the most differentially expressed gene, CD2, also associated with non-recurrence (p<0.001). The immune gene panel and CD2 staining associated with prolonged survival (p<0.001 and p=0.019, respectively). Conclusions: mRNA copy number of immune-related genes in primary FFPE melanomas predicts non-recurrence and prolonged survival. This data highlights the impact of immunosurveillance in primary human melanoma and the identified gene panel may be a useful tool for patient stratification for adjuvant immunotherapy studies.