Identification of Aging and Young Subtypes for Predicting Colorectal Cancer Prognosis and Immunotherapy Responses

Abstract

Colorectal cancer (CRC) is critically related to aging and severely threatens human lives. To better explore the effects of aging on CRC progression and therapy outcome, a reliable aging subtypes identification of CRC is urgently desired. Here, 28 aging-related genes associated with the CRC prognosis were selected by univariate Cox analyses. Based on these 28 genes, CRC patients were divided into the aging subtype and young subtype by non-negative matrix factorization clustering. Aging subtype and young subtype of CRC were identified with distinct molecular features and clinical prognosis. The aging subtype was characterized by upregulation of senescence-associated secretory phenotype, higher frequencies of TP53 and immune checkpoint molecules, and high sensitivity to protein kinase and angiogenesis inhibitors. Furthermore, 14 genes were selected by LASSO penalized Cox regression analyses for aging-related risk signature construction. The constructed aging risk signature exhibited good prediction and the nomogram showed robust discrimination power over the traditional CRC staging system. In conclusion, this study successfully established aging subtype and young subtype of CRC, which is helpful to identify patients with aging characteristics to evaluate prognosis and treatment outcomes. Introducing aging-based subtypes into clinical concern and patient prognostication provides new opportunities for personalized CRC treatment.

Keywords

• colorectal cancer
• aging
• unsupervised clustering
• tumor microenvironment
• prognosis