Along with the increasing knowledge of long noncoding RNA, the interaction between the long noncoding RNA (lncRNA) and tumor immune infiltration is increasingly valued. However, there is a lack of understanding of correlation between regulation of specific lncRNAs and tumor-infiltrating macrophages within melanoma. In this research, a macrophage associated lncRNA signature was identified by multiple machine learning algorithms and the robust and effectiveness of signature also validated in other independent datasets. The signature contained six specific lncRNAs (PART1, LINC00968, LINC00954, LINC00944, LINC00518 and C20orf197) was constructed, which could diagnose melanoma and predict the prognosis of patients. Moreover, our signature achieves higher accuracy than the previous well-established markers and regarded as an independent prognostic indicator. The pathway enrichment revealed that these lncRNAs were closely correlated with many immune processes. In addition, the signature was associated with different immune microenvironment and applied to predict response of immune checkpoint inhibitor therapy (low risk of patients well respond to anti-PD-1 therapy and high risk is insensitive to anti-CTLA-4 therapy). Therefore, our finding supplies a more accuracy and effective lncRNA signature for tumor-infiltrating macrophages targeting treatment approaches and affords a new clinical application for predicting the response of immunotherapies in melanomas.
Keywords: cutaneous; immunotherapy; lncRNA; macrophage; melanoma; prognosis.