Related Experiment Video
Updated: Sep 18, 2025
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
9.1K
Interpretable Machine Learning for Serum-Based Metabolomics in Breast Cancer Diagnostics: Insights from
Emek Guldogan1, Fatma Hilal Yagin2, Hasan Ucuzal1
1Department of Biostatistics, and Medical Informatics, Faculty of Medicine, Inonu University, 44280 Malatya, Turkey.
Medicina (Kaunas, Lithuania)
|June 27, 2025
Summary
This study identifies novel serum metabolic biomarkers for breast cancer detection using advanced metabolomics and explainable artificial intelligence (XAI). The findings improve diagnostic accuracy and offer biological insights into disease progression.
Area of Science:
- Metabolomics
- Machine Learning
- Biomarker Discovery
Background:
- Breast cancer is a leading cause of death in women globally.
- Traditional biomarkers for breast cancer lack sensitivity and specificity, especially in early stages.
- Metabolomics and explainable artificial intelligence (XAI) offer promising avenues for improved breast cancer diagnostics.
Purpose of the Study:
- To identify and validate serum-based metabolic biomarkers for breast cancer detection.
- To enhance diagnostic accuracy using advanced metabolomic profiling and machine learning.
- To leverage SHapley Additive exPlanations (SHAP) for model interpretability and biological insight.
Main Methods:
- Serum samples from 103 breast cancer patients and 31 controls were analyzed using LC-TOFMS and GC-TOFMS.
- Multi-Objective Feature Selection (MOFS) was employed for robust biomarker discovery.
- Light Gradient Boosting Machine (LightGBM) with SHAP analysis was used for classification and metabolite importance ranking.
Main Results:
- LightGBM achieved 86.6% accuracy, 89.1% sensitivity, and 84.2% specificity in breast cancer detection.
- SHAP analysis identified 2-Aminobutyric acid, choline, and coproporphyrin as key influential metabolites.
- Metabolite dysregulation was significantly associated with breast cancer risk.
Conclusions:
- This study integrates SHAP explainability with metabolomics for enhanced breast cancer diagnostics.
- The identified biomarkers improve diagnostic accuracy and reveal metabolic dysregulations linked to breast cancer.
- Combining metabolomics with XAI-driven machine learning shows significant potential for clinical adoption.
