Development and evaluation of an ovarian cancer prognostic model based on adaptive immune-related genes.

Abstract

The adaptive immune system plays a vital role in cancer prevention and control. However, research investigating the predictive value of adaptive immune-related genes (AIRGs) in ovarian cancer (OC) prognosis is limited. This study aims to explore the functional roles of AIRGs in OC. Transcriptomic, clinical-pathological, and prognostic data for OC were downloaded from public databases. Differential expression analysis, univariate, and Lasso Cox regression analyses were utilized to construct a risk signature. Kaplan-Meier survival analysis, enrichment analysis, somatic mutation analysis, immune infiltration analysis, and drug sensitivity analysis were performed to characterize differences between high-risk and low-risk groups. Independent prognostic factors were identified through multivariate Cox regression analysis to construct a nomogram. Expression of signature-related AIRGs was validated using in OC cells and tissues. A total of 109 AIRGs significantly associated with overall survival (OS) in OC were identified, of which 15 were selected to construct the risk signature: AP1S2, AP2A1, ASB2, BTLA, BTN3A3, CALM1, CD3G, CD79A, EVL, FBXO4, FBXO9, HLA-DOB, LILRA2, MALT1, and PIK3CD. This signature stratified the OC cohort into high-risk and low-risk groups, which exhibited significant differences in prognosis, gene expression, mutation profiles, immunotherapy response, and drug sensitivity. Specifically, the low-risk group showed better prognosis, higher tumor mutational burden, greater response to immunotherapy, increased M1 macrophage and T follicular helper (Tfh) cell infiltration, and higher sensitivity to cisplatin and gemcitabine. The nomogram, integrating the AIRG-derived risk signature with age and clinical stage, demonstrated superior performance in predicting OC prognosis compared to other factors. Moreover, the differential expression of signature-related AIRGs were further confirmed in OC cells and tissue as compared to the normal cells or tissues. Our findings highlight the significant association between AIRGs and the prognosis of OC. The prognostic model developed using AIRGs demonstrates strong predictive capabilities.