A four gene-based risk score system associated with chemoradiotherapy response and tumor recurrence in rectal cancer by co-expression network analysis

Abstract

Aim: Resistance to neoadjuvant chemoradiotherapy (NCRT) and tumor recurrence presents a major clinical problem in locally advanced rectal cancer (LARC) patients. This study aimed to explore a genetic risk score related to NCRT response and tumor recurrence in rectal cancer after NCRT. Materials and Methods: Weighted gene co-expression network analysis was employed to identify hub genes associated with NCRT response from the GSE93375 dataset. Prognostic hub genes were determined using Cox regression analysis and associated with disease-free survival (DFS). A risk score system was constructed and the prognostic significance of the risk score was validated in our patient cohort. A predictive nomogram for DFS was developed and validated internally. Results: The Tan module had the highest correlations with NCRT response. Ten hub genes (COL15A1, THBS2, ITGB1, MMP2, CD34, SPARC, NOTCH3, PDGFRB, DCN, and SERPINH1) were associated with NCRT response. Immunostaining expression of four genes (NOTCH3, SPARC, DCN, and ITGB1) was found to be significantly associated with both NCRT response and DFS in our patient cohort and was selected to build a prognostic risk score for DFS as follows: risk score= (0.6188×ExpNOTCH3) + (0.6511×ExpSPARC) + (−0.2976×ExpDCN) + (1.0035×ExpITGB1). Using this risk score, patients could be separated into high-and low-risk groups for tumor recurrence. A nomogram that incorporated the risk score, ypTNM stage, and tumor regression grade (TRG) was constructed and utilized to predict DFS in LARC patients. Conclusion: The four-gene expression-based risk score system presented here could be potentially used for predicting tumor recurrence in LARC patients after NCRT.