A multiomics analysis-assisted deep learning model identifies a macrophage-oriented module as a potential therapeutic target in colorectal cancer

• DOI: 10.1016/j.xcrm.2024.101399
• Publication Date: 2024-02-01T15:44:18Z
• Authors: Xuanwen Bao; Qiong Li; Dong Chen; Xiaomeng Dai; Chuan Liu; Weihong Tian; Hangyu Zhang; Yuzhi Jin; Yin Wang; Jinlin Cheng; Chunyu Lai; Chanqi Ye; Shan Xin; Xin Li; Ge Su; Yongfeng Ding; Yangyang Xiong; Jindong Xie; Vincent Tano; Yanfang Wang; Wenguang Fu; Shuiguang Deng; Weijia Fang; Jianpeng Sheng; Jian Ruan; Peng Zhao

Colorectal cancer (CRC) is a common malignancy involving multiple cellular components. The CRC tumor microenvironment (TME) has been characterized well at single-cell resolution. However, spatial interaction map of the TME still elusive. Here, we integrate multiomics analyses and establish to improve prognosis, prediction, therapeutic development for CRC. We construct immune module (CCIM) that comprises FOLR2+ macrophages, exhausted CD8+ T cells, tolerant CD4+ regulatory cells. Multiplex immunohistochemistry performed depict CCIM. Based on this, utilize advanced deep learning technology predict chemotherapy response. CCIM-Net constructed, which demonstrates good predictive performance response in both training testing cohorts. Lastly, targeting macrophage therapeutics used disrupt immunosuppressive CCIM enhance vivo.