Development of a Macrophage-Related Risk Model for Metastatic Melanoma.

As a metastasis-prone malignancy, the metastatic form and location of melanoma seriously affect its prognosis. Although effective surgical methods and targeted drugs are available to enable the treatment of carcinoma in situ, for metastatic tumors, the diagnosis, prognosis assessment and development of immunotherapy are still pending. This study aims to integrate multiple bioinformatics approaches to identify immune-related molecular targets viable for the treatment and prognostic assessment of metastatic melanoma, thus providing new strategies for its use as an immunotherapy. Immunoinfiltration analysis revealed that M1-type macrophages have significant infiltration differences in melanoma development and metastasis. In total, 349 genes differentially expressed in M1-type macrophages and M2-type macrophages were extracted from the MSigDB database. Then we derived an intersection of these genes and 1111 melanoma metastasis-related genes from the GEO database, and 31 intersected genes identified as melanoma macrophage immunomarkers (MMIMs) were obtained. Based on MMIMs, a risk model was constructed using the Lasso algorithm and regression analysis, which contained 10 genes (NMI, SNTB2, SLC1A4, PDE4B, CLEC2B, IFI27, COL1A2, MAF, LAMP3 and CCDC69). Patients with high+ risk scores calculated via the model have low levels of infiltration by CD8 T cells and macrophages, which implies a poor prognosis for patients with metastatic cancer. DCA decision and nomogram curves verify the high sensitivity and specificity of this model for metastatic cancer patients. In addition, 28 miRNAs, 90 transcription factors and 29 potential drugs were predicted by targeting the 10 MMIMs derived from this model. Overall, we developed and validated immune-related prognostic models, which accurately reflected the prognostic and immune infiltration characteristics of patients with melanoma metastasis. The 10 MMIMs may also be prospective targets for immunotherapy.