Identification of anoikis-related genes to develop a risk model and predict the prognosis and tumor microenvironment in rectal adenocarcinoma

Background Rectal adenocarcinoma (READ) is a common malignant tumor. This study aims to establish a risk model based on anoikis-related genes (ARGs) to predict prognosis and the tumor microenvironment in READ. Methods Transcriptomic data and clinical data downloaded from the TCGA and GEO databases were used for differential analysis and Cox regression analysis. An ARGs-based prognostic risk model was constructed for READ. The survival curves and ROC curves were plotted to determine the predictive ability of the model for READ patients. The model was externally validated in the GSE87211 dataset. A nomogram, immune analysis, drug sensitivity analysis, and functional enrichment analysis were also performed to comprehensively validate the model. Results The risk model included 6 prognostic genes (ALDH1A1, BRCA1, GSN, KRT17, SCD, and SNCG). Kaplan-Meier curves for the TCGA training cohort (P < 0.0001), testing cohort (P = 0.018), and GSE87211 dataset (P = 0.036) showed better prognoses in the low-risk group. The AUC for 1-year, 3-year, and 5-year overall survival in the TCGA training cohort, testing cohort, and GSE87211 dataset were (0.962, 0.923, 0.956), (0.887, 0.838, 0.833), and (0.73, 0.817, 0.743), respectively. The nomogram showed that the risk score served as an independent predictor of overall survival. Drug sensitivity analysis revealed differences in the IC50 values of OSI-027, PLX-4720, UMI-77, and Sapitinib between the high-risk and low-risk groups. Immune microenvironment analysis suggested distinct differences in immune cells between the two risk groups. Enrichment analysis revealed that these prognostic ARGs were primarily enriched in pathways and biological processes related to tumorigenesis. Conclusion The risk model of ARGs can effectively predict READ prognosis and provide potential therapeutic targets.