Novel Pathomics Model Predicts Breast Cancer Prognosis using Machine Learning Algorithms

Strategic Insights -

Researchers at the First People's Hospital in Guizhou, China, have developed a novel pathomics model that reliably predicts breast cancer prognosis using machine learning algorithms. The study, published in the journal PLOS ONE, analyzed the expression of checkpoint kinase 1 (CHEK1) in breast cancer tissue and found that high CHEK1 expression levels were associated with poorer prognosis and lower survival rates, particularly in patients with high pathomics scores (PS). The team also found that patients with high PS responded better to anti-PD-1 and anti-CTLA4 treatments.

Key Takeaways:

  • The novel pathomics model developed by researchers at the First People's Hospital in Guizhou, China, successfully predicts breast cancer prognosis using machine learning algorithms.
  • High CHEK1 expression levels were associated with poorer prognosis and lower survival rates in breast cancer patients.
  • The model was able to predict CHEK1 expression levels using a subset of 8 recursive feature elimination (RFE)-screened features extracted from PyRadiomics, mRMRe, and Gradient Boosting Machine (GBM) algorithms.
  • Patients with high PS (pathomics scores) responded better to anti-PD-1 and anti-CTLA4 treatments.
  • The study suggests that the pathomics model may provide potential clinical utility for prognosis and treatment guidance in breast cancer patients.

Statistics:

  • The study used a 633 x 10 sub-image set for training and a 158 x 10 set for validation.
  • A total of 1,488 features were extracted from the images, and 8 recursive feature elimination (RFE)-screened features were used to generate the model.
  • The study found that high PS was associated with CHEK1 overexpression, significantly correlating with survival outcomes within 96 months post-diagnosis.

Sources:

  • Percival, M. First People's Hospital Reports Findings in Breast Cancer (Predicting breast cancer prognosis based on a novel pathomics model through CHEK1 expression analysis using machine learning algorithms). Journal of Engineering, 2025; p 908.
  • Gao, D., Chen, C., Yue, H., Wang, H., Qu, R., Hu, X., & Luo, L. Predicting breast cancer prognosis based on a novel pathomics model through CHEK1 expression analysis using machine learning algorithms. PLOS ONE, 2025;20(5).