Nomogram based on autophagy related genes for predicting the survival in melanoma

Abstract

Abstract Background Autophagy, a highly conserved lysosomal degradation pathway, is associated with the prognosis of melanoma. However, prognostic prediction models based on autophagy related genes (ARGs) have never been recognized in melanoma. In the present study, we aimed to establish a novel nomogram to predict the prognosis of melanoma based on ARGs signature and clinical parameters. Methods Data from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) databases were extracted to identify the differentially expressed ARGs. Univariate, least absolute shrinkage and selection operator (LASSO) and multivariate analysis were used to select the prognostic ARGs. ARGs signature, age and stage were then enrolled to establish a nomogram to predict the survival probabilities of melanoma. The nomogram was evaluated by concordance index (C-index), receiver operating characteristic (ROC) curve and calibration curve. Decision curve analysis (DCA) was performed to assess the clinical benefits of the nomogram and TNM stage model. The nomogram was validated in GEO cohorts. Results Five prognostic ARGs were selected to construct ARGs signature model and validated in the GEO cohort. Kaplan-Meier survival analysis suggested that patients in high-risk group had significantly worse overall survival than those in low-risk group in TCGA cohort (P = 5.859 × 10–9) and GEO cohort (P = 3.075 × 10–9). We then established and validated a novel promising prognostic nomogram through combining ARGs signature and clinical parameters. The C-index of the nomogram was 0.717 in TCGA training cohort and 0.738 in GEO validation cohort. TCGA/GEO-based ROC curve and decision curve analysis (DCA) demonstrated that the nomogram was better than traditional TNM staging system for melanoma prognosis. Conclusion We firstly developed and validated an ARGs signature based-nomogram for individualized prognosis prediction in melanoma patients, which could assist with decision making for clinicians.