Development of a clinical decision support system for breast cancer detection using ensemble deep learning

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Oncology Diagnostics

Background:

• Breast cancer poses a significant global health challenge, necessitating improved diagnostic technologies for early detection and better patient outcomes.
• Current diagnostic methods face limitations in precision and speed, highlighting the need for advanced computational approaches.

Purpose of the Study:

• To develop and evaluate a unique Deep Learning (DL) Ensemble Clinical Decision Support System (EDL-CDSS) for precise and rapid breast cancer diagnosis.
• To improve the accuracy and efficiency of breast cancer detection by combining multiple DL models.

Main Methods:

• The study proposes an EDL-CDSS integrating various DL models, including Kelm Extreme Learning Machine (KELM) and Deep Belief Network (DBN).
• This ensemble approach leverages the strengths of individual DL architectures to extract complex patterns from medical imaging data.
• The system's performance was rigorously tested on diverse datasets, comparing it against individual DL models and traditional diagnostic techniques.

Main Results:

• The EDL-CDSS demonstrated superior performance in diagnosing breast cancer compared to individual DL models and conventional methods.
• Key performance metrics evaluated included precision, sensitivity, specificity, F1-score, and overall accuracy.
• The system achieved a remarkable overall accuracy of 96.14%, significantly outperforming prior advanced methodologies.

Conclusions:

• The developed EDL-CDSS offers a promising advancement in breast cancer diagnostic technology.
• The ensemble DL approach effectively enhances diagnostic precision and speed, potentially leading to improved patient outcomes.
• This system shows significant potential for clinical application in early and accurate breast cancer detection.