Developing the breast cancer risk prediction system using hybrid machine learning algorithms.

Area of Science:

• Oncology
• Medical Informatics
• Machine Learning

Background:

• Breast cancer (BC) is a leading cause of cancer mortality in women globally.
• Machine learning (ML) models aid clinical decision-making for BC prediction.
• Early detection and risk factor prevention are crucial for population-wide BC health.

Purpose of the Study:

• To develop a predictive model using a genetic algorithm (GA) and ML algorithms for BC prediction and early warning.
• To identify key risk factors for breast cancer.
• To create a clinical decision support system for early BC detection.

Main Methods:

• Analysis of 3168 healthy individuals and 1742 BC patient records.
• Utilized a hybrid genetic algorithm (GA) for feature selection and optimization.
• Trained and evaluated multiple ML algorithms (e.g., decision tree) for BC prediction using selected features.

Main Results:

• Identified key predictors: age, dairy consumption, family history, biopsy, X-ray, hormone therapy, alcohol, weight, children, education.
• The decision tree model demonstrated superior performance with 99.3% accuracy, 99.5% specificity, and 98.26% sensitivity.
• A clinical decision support system was developed based on the best-performing model.

Conclusions:

• The developed system accurately identifies individuals at elevated risk for breast cancer.
• This predictive system serves as a vital clinical screening tool for early BC prevention.
• The tool supports the development of effective preventive health strategies for breast cancer.