Identification of Prognostic Genes and a Subtype Sensitive to Cetuximab in Colorectal Cancer

Abstract

Background: Colorectal cancer (CRC) is one of the most lethal and prevalent malignancies worldwide. Cetuximab treatment appears to prolong the survival of CRC patients, although some patients do not respond or they develop drug resistance. We aimed to identify favorable subtypes and genes in CRC that are sensitive to cetuximab to better manage CRC patients. Methods: Data were retrieved from the online public databases. Differential expression analysis, weighted gene co-expression network analysis, and univariate COX analyses were utilized to identify prognostic genes that are sensitive to cetuximab in CRC. All samples were divided into several subgroups using ConsensusClusterPlus based on prognostic genes. Survival, immune features, and mutation features among the subgroups were analyzed. In addition, the optimal prognostic genes were further screened to construct a risk score model and nomogram combined with the clinical information. Results: A total of 26 prognostic sensitive genes were screened. All of these genes were allocated into cluster 1 and cluster 2. Cluster 2 exhibited better survival, immune status, and a lower tumor mutation burden. Five optimal genes (ADAMTS13, HAMP, MMP10, SLC6A1, and TMEM220) were filtrated, and a risk score model was established. The risk score model based on these five genes showed promising predictive value at 1, 3, and 5 years, with an area under the receiver operator characteristic curve of more than 7. The nomogram based on the risk score and clinical features also possessed good performance in predicting the survival status after 1, 3, and 5 years. Conclusions: The identified genes and the subtypes exhibit potentially useful prognostic abilities and may be indicators of CRC sensitivity to cetuximab.