Network-based molecular subtyping of acral melanoma

Abstract Acral melanoma is more biologically aggressive with a worse prognosis compared with other melanoma subtypes. However, the molecular basis underlying the biological and clinical behavior of this cancer is still unclear. Here, using the combination of multi-omics data analysis and network-based disease gene prediction algorithm, we first demonstrate the existence of two acral melanoma subtypes which greatly differed in clinical performance, cellular and molecular mechanisms, and discovered a biomarker panel (EREG, VSIG4, FCGR3A, RAB20) that accurately distinguished these two subtypes with the AUC of 0.946, which has been verified by clinical samples. Subtype I has thinner Breslow with a better prognosis. On the contrary, subtype II is a high-risk subtype that is easier to invade the dermis. We further analyzes the intrinsic biological mechanism of the two subtypes from the cellular level, and reveals the important role of macrophages subgroups in the molecular typing of acral melanoma. Feature genes of subtype I are enriched in FCN1+ macrophages that promote inflammatory and immune responses. In contrast, feature genes of subtype II are enriched in SPP1+ macrophages which ha the greatest impact on tumor cells. The identification of the two subtypes opens up important biological and clinical perspectives for acral melanoma.