Abstract
Background: The tumor microenvironment (TME) is increasingly recognized as a key player in colorectal cancer biology, however, its potential for improving diagnosis, prognosis, and treatment remains unclear. The major aim of this study is to explore the prognostic value of TME related gene in colorectal cancer. Method: Expression matrices and clinical data of colorectal cancer obtained from public databases were divided into TME relevant clusters according to immune characterization. A 11-gene molecular classifier was constructed based on differentially expressed genes between TME clusters and machine learning regression processes. Results: The efficacy and effectiveness of TME based prognostic signature (TPS) were examined in both the training and validation groups. The result indicated that TPS was able to serve as a superior prognosis indicator for colorectal cancer, alone or jointly with other clinical factors. Also, the study demonstrated that high risk colorectal cancer defined by TPS was considered to link with elevated immune infiltration, increased tumor mutation, and worse overall prognosis. Finally, potential therapeutic agents specialized for different risk subgroups of TPS was also identified to improve personalized treatment for colorectal cancer in the future. Conclusions: TPS might be a novel tool to improve the prognosis judgement and personalized treatment of the colorectal cancer in the future.