A novel risk-scoring model developed based on tertiary lymphoid structure related genes predict the prognosis and immunotherapeutic efficacy in cutaneous melanoma.

Abstract

e13574 Background: Emerging evidence has proven that tertiary lymphoid structures (TLSs) are predictive biomarkers of favorable clinical outcomes in melanoma, whereas the exact role of TLS-related genes remained unclear. In the present study, we aimed to establish a novel risk-scoring model based on TLS-related genes to predict the prognosis and immunotherapeutic efficacy in cutaneous melanoma. Methods: Patients in the TCGA-SKCM cohort (n = 468) were clustered into two subtypes based on the prognosis-related genes associated with TLSs though consensus clustering analysis, and differentially expressed genes were screened between two clusters. By univariate and the least absolute shrinkage and selection operator (LASSO) Cox regression analysis, a 13-gene signature was selected to construct a prognostic risk-scoring model. The median risk score served as the unified cutoff to divide patients into high- and low-risk groups. GSE65904 (n = 220) and GSE544467 (n = 79) cohorts were used for external validation. Univariate and multivariate cox analysis were used to select significant variables for construction of nomogram. The nomogram model was evaluated by concordance index, time-dependent receiver operating characteristics (ROC) curve, calibration curve and decision curve analysis (DCA). GSEA, CIBERSORT, Tumor Immune Dysfunction and Exclusion (TIDE) and Immunophenoscore (IPS) were utilized to explore the potential mechanism of immune-related process and immunotherapy. Results: Thirteen signature genes (CCL8, DLX3, CNTFR, CNFN, BCAN, CD33, HAPLN3, ABCB1, CLIC2, GCNT1, FGF1, A2ML1, GBP2) were developed to construct a risk-scoring model. Melanoma patients could be grouped into high- and low-risk groups, and low-risk patients showed better survival than high-risk patients. Area under the curve values of 1, 3, and 5 years were 0.76, 0.72, and 0.73 in the training set, respectively, indicating a good predictive efficacy. In addition, based on independent prognostic factors including risk score, age, stage, T and N Staging after univariate and multivariate Cox regression analyses, a nomogram was constructed to better estimate individual survival. For the prediction of 1, 3, and 5-year overall survival probability according to the nomogram, the area under the receiver operating characteristic curve was 0.81, 0.80 and 0.76, respectively. Furthermore, we explored the correlation of the prognostic risk-scoring model with tumor microenvironment (TME) and immunotherapeutic efficacy. Patients in the low risk group featured immunoactive TME, higher IPS scores and lower TIDE scores, and was regarded as the potential beneficiaries of immunotherapy. Conclusions: The 13-gene risk score proposed in the research may provide new insights into the individualized evaluation of cutaneous melanoma prognosis and Immunotherapeutic Efficacy