Machine learning algorithms predict breast cancer incidence risk: a data-driven retrospective study based on...

Area of Science:

• Biochemistry
• Oncology
• Machine Learning

Background:

• Current breast cancer prediction models lack sufficient inclusion of blood-based biomarkers.
• This study aimed to identify novel breast cancer risk factors using machine learning.

Purpose of the Study:

• To identify novel breast cancer risk factors by integrating clinical data and peripheral blood biochemical biomarkers.
• To enhance the understanding of breast cancer risk through advanced analytical methods.

Main Methods:

• Screening and normalization of data based on inclusion/exclusion criteria.
• Application of logistic regression with forward selection and six machine learning algorithms.
• Evaluation of model performance using area under the curve (AUC) via 5-fold cross-validation.

Main Results:

• Logistic regression identified age, gamma-glutamyl transferase (GGT), and alanine transaminase (ALT) as increasing breast cancer incidence risk factors.
• Six machine learning algorithms consistently highlighted GGT and ALT as significant predictive features.
• Model performance showed AUC values ranging from 0.779 to 0.862 and accuracy from 0.780 to 0.841.

Conclusions:

• Gamma-glutamyl transferase (GGT) and alanine transaminase (ALT) are identified as promising biochemical biomarkers for breast cancer prediction.
• Future research should focus on incorporating these biomarkers into tailored breast cancer risk prediction models.