Identification of a Novel Tumor Inflammation Signature for Risk Stratification, Prognosis Prediction, and Immune Status in Colorectal Cancer

Abstract

Background. Inflammation and immune cell dysfunction have been widely known as an essential role in the tumorigenesis of colorectal cancer (CRC). Yet, the role of tumor inflammation signature (TIS) associated with CRC prognosis, immune infiltration, and drug resistance remained unknown. Method. The transcriptome sequencing data, as well as clinical data of CRC from the public dataset, were acquired for further investigation. Inflammation-related gene expression patterns were obtained and analyzed. Bioinformatics methods were used to build a prognostic TIS, and its prediction accuracy was verified by using ROC curve analyses. The independent prognostic factors in CRC were identified through multivariable Cox regression analysis. In addition, the specific features of the immunological landscape between low- and high-risk CRC cohorts were analyzed. Results. We firstly screened the differentially expressed inflammation-related genes in CRC and constructed a prognostic TIS. We further classified CRC patients into high or low TIS score groups based on the optimal cutoff of prognostic TIS, and patients with high-risk scores had shorter overall survival (OS) than those in the low-risk cohort. The diagnostic accuracy of TIS was evaluated and confirmed with ROC analysis. The result of the univariate and multivariate analysis found that TIS was directly and independently linked to OS of CRC. Otherwise, an optimal nomogram model based on TIS exhibited a better prognostic accuracy in OS. Finally, the immunological status and immune cell infiltration were observed different in the two-risk cohorts. Conclusion. In summary, the risk model of the TIS in CRC was found to be useful for predicting patient prognosis and guiding individual treatment. This risk signature could also serve as potential biomarkers and immunotherapeutic targets and indicate immunotherapy response for patients with CRC.