A Federated Learning Approach to Breast Cancer Prediction in a Collaborative Learning Framework.

Area of Science:

• Oncology
• Artificial Intelligence
• Medical Informatics

Background:

• Breast cancer is a significant global health issue requiring advanced diagnostic tools.
• Timely detection and management are crucial for improving patient outcomes.
• Existing diagnostic methods can be enhanced by leveraging AI and collaborative data approaches.

Purpose of the Study:

• To develop a highly accurate breast cancer detection model using Deep Neural Networks (DNN).
• To implement a Federated Learning (FL) framework for secure, collaborative model training across multiple institutions.
• To improve the accuracy and reliability of breast cancer screening.

Main Methods:

• Utilized an iterative methodology for collaborative learning with DNNs.
• Employed Federated Learning (FL) to aggregate knowledge from diverse healthcare data sources while preserving patient privacy.
• Applied optimum feature selection and data augmentation techniques to enhance model performance.

Main Results:

• Achieved a maximum accuracy of 97.54% in breast cancer detection.
• Reported a precision of 96.5% and a recall of 98.0%.
• Obtained an F1-Score of 97%, indicating robust model performance.

Conclusions:

• The developed DNN model with FL represents a significant advancement in breast cancer diagnostics.
• Federated learning offers a secure and effective way to utilize multi-institutional data for improved AI models.
• This approach holds potential for transforming early breast cancer detection and treatment, benefiting patients worldwide.