A Machine-Learning Prognostic Model for Colorectal Cancer Using a Complement-Related Risk Signature
Jun Li1, Kangmin Yu1, Zhiyong Chen1
1Department of Vascular Surgery, The First Affiliated Hospital of Anhui Medical University, Hefei, 230022, China.
Oncology Research
|November 3, 2025
Summary
A new six-gene complement-related risk signature (CRRS) model accurately predicts colorectal cancer (CRC) patient survival. This model aids in understanding the CRC immune microenvironment and guides personalized treatment decisions.
Area of Science:
- Oncology
- Computational Biology
- Immunology
Background:
- Colorectal cancer (CRC) presents significant global mortality and heterogeneous patient outcomes.
- Understanding the CRC immune microenvironment is crucial for improving therapeutic strategies.
Purpose of the Study:
- To develop a machine-learning prognostic model using a complement-related risk signature (CRRS) for colorectal cancer.
- To analyze the relationship between the CRRS and the CRC immune microenvironment.
Main Methods:
- Transcriptomic data from TCGA and GEO CRC cohorts were analyzed.
- A random survival forest (RSF) model was trained and validated to identify prognostic CRRS genes.
- Immune infiltration, mutational burden, pathway enrichment, and drug sensitivity were compared between risk groups.
Main Results:
- The six-gene CRRS model effectively stratified CRC patients based on survival.
- Low-risk patients showed increased immune cell infiltration and predicted better response to immunotherapy/chemotherapy.
- High-risk patients exhibited complement activation and matrix remodeling pathway enrichment; FAM84A promoted CRC progression.
Conclusions:
- The CRRS is a key factor influencing the colorectal cancer immune microenvironment.
- The developed CRRS model offers precise risk prediction for individualized CRC therapy.
