Ferroptosis, a novel form of regulating cell death, is related to various cancers. However, the role of ferroptosis-related genes (FRGs) on the occurrence and development of colon cancer (CC) needs to be further elucidated.
CC transcriptomic and clinical data were downloaded from TCGA and GEO databases. The FRGs were obtained from the FerrDb database. The consensus clustering was performed to identify the best clusters. Then, the entire cohort was randomly divided into the training and testing cohorts. Univariate Cox, LASSO regression and multivariate Cox analyses were used to construct a novel risk model in training cohort. The testing and the merged cohorts were performed to validate the model. Moreover, CIBERSORT algorithm analyze TIME between high- and low- risk groups. The immunotherapy effect was evaluated by analyzing the TIDE score and IPS between high- and low- risk groups. Lastly, RT-qPCR were performed to analyze the expression of the three prognostic genes, and the 2-years OS and DFS between the high- and low- risk groups of 43 clinical CC samples to further validate the value of the risk model.
SLC2A3, CDKN2A, and FABP4 were identified to construct a prognostic signature. Kaplan–Meier survival curves showed that OS between the high- and low-risk groups were statistically significant (pmerged<0.001, ptraining<0.001, ptesting<0.001). TIDE score and IPS were higher in the high-risk group (pTIDE<0.005, pDysfunction<0.005, pExclusion<0.001, pmAb-CTLA-4 = 3e-08, pmAb-PD-1 = 4.1e-10). The clinical samples were divided into high- and low- risk groups according to the risk score. There was a statistical difference in DFS (p=0.0108).
This study established a novel prognostic signature and provided more insight into the immunotherapy effect of CC.