Harnessing Anoikis-Related Gene Signatures for Immunotherapy Precision: A Machine Learning-Driven Predictive and...

Area of Science:

• Oncology
• Immunology
• Genetics

Background:

• Cancer immunotherapy, particularly immune checkpoint inhibitors (ICIs), shows variable patient responses.
• Anoikis-related genes (ARGs) play a role in cancer progression and may influence immune responses.

Purpose of the Study:

• To investigate the association between ARGs and immunotherapy efficacy.
• To develop a machine learning model for predicting immunotherapy response using ARGs.

Main Methods:

• Integrated single-cell RNA sequencing and multiomics data to identify ARGs.
• Developed and validated machine learning models (including SVM) for predicting response.
• Utilized CRISPR screening to identify immune-resistant ARGs and performed pan-cancer prognostic analysis.

Main Results:

• Identified a significant association between anoikis-related gene signatures (Anoikis.Sig) and immunotherapy response.
• The SVM-based model demonstrated superior predictive performance (AUC=0.782) across validation datasets.
• Key immune-resistant ARGs (e.g., BCL2L1, ITGAV, PTK2) were identified, and an 11-gene signature showed strong prognostic value, especially in hepatocellular carcinoma (HCC).

Conclusions:

• ARGs are crucial in modulating the tumor-immune microenvironment and serve as effective biomarkers for ICI efficacy.
• The developed machine learning model aids in patient stratification and personalized therapy.
• Targeting ARGs presents a promising novel strategy to improve immunotherapy outcomes.