AUTHOR=Zhang Chuan , Dang Dan , Wang Yuqian , Cong Xianling 

TITLE=A Nomogram Combining a Four-Gene Biomarker and Clinical Factors for Predicting Survival of Melanoma

JOURNAL=Frontiers in Oncology

VOLUME=Volume 11 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.593587

DOI=10.3389/fonc.2021.593587

ISSN=2234-943X

ABSTRACT=Background Considering melanoma, as a highly heterogeneous tumor and the deadliest skin cancer, presently lacks an effective prognostic indicator, there is an urgent need to identify an optimized immune-associated prognostic biomarker for melanoma treatment.
Method Melanoma RNAseq data and the corresponding immune score were analyzed using Weighted Gene Co-expression Net Analysis (WGCNA) to identify the most immune-related module, which was subsequently verified by the function enrichment analysis. Univariate Cox regression model and LASSO regression were utilized to determine the qualified genes, which would be used to construct the prediction model. Time-dependent ROC (tROC) and log rank test were conducted to assess the established model in train set, validation set and test set, respectively. The significance of the model was further explored in terms of tumor immune scores, tumor-infiltrating immune cells (TII cells) and KEGG pathways. Finally, univariate Cox regression model and nomagram were performed in order to guide clinical decision-making for melanoma.
Results Module Black consisting of 618 genes was identified and validated as the most immune-related module. And then these genes were further processed by means of univariate Cox regression and LASSO regression. Four genes of CLEC7A, CLEC10A, HAPLN3 and HCP5 were eventually isolated to comprise an immune-related prognostic biomarker using multivariate Cox regression model. The forecasting performance of the biomarker was validated using tROC and log rank test in train set, validation set and test set. Consistent with that, the four genes were also significantly associated with favorable survival of melanoma, which could serve as potential therapeutic targets for melanoma. In addition, macrophage M1, NK cell, CD4+ and CD8+ T cells were all upregulated in tumor microenvironment of low risk group. GSEA findings also revealed that more immune pathways were activated in low risk population. Furthermore, the prognostic biomarker was proven to serve as an independent prognostic feature. 
Conclusion Our findings demonstrated an optimized immune-associated prognostic biomarker consisting of just four relevant genes, and revealed favorable tumor-infiltrating immune cells and beneficial activated immune pathways present in low risk score populations, which could serve as potential therapeutic targets for melanoma.