Demystifying diagnosis: an efficient deep learning technique with explainable AI to improve breast cancer detection

Area of Science:

• Oncology
• Artificial Intelligence
• Biomedical Informatics

Background:

• Breast cancer (BC) is a leading cause of death globally, particularly in middle-income countries.
• Early detection significantly improves patient prognosis and survival rates.
• Traditional artificial intelligence (AI) models often lack transparency in their decision-making processes.

Purpose of the Study:

• To develop an interpretable AI model for accurate breast cancer detection.
• To enhance the understanding of AI-driven breast cancer diagnosis through explainable AI (XAI).
• To improve confidence in AI-based tools for healthcare practitioners and researchers.

Main Methods:

• A hybrid deep learning model combining bi-directional long short-term memory (BiLSTM) and convolutional neural networks (CNN) was developed.
• Dataset balancing was performed prior to model training.
• Explainable artificial intelligence (XAI) techniques were integrated to interpret model decisions.

Main Results:

• The BiLSTM-CNN model achieved high performance metrics.
• Accuracy: 0.993
• Precision: 0.99, Recall: 0.99, F1-score: 0.99

Conclusions:

• The developed hybrid BiLSTM-CNN model effectively identifies breast cancer using patient data.
• Integration of XAI enhances model interpretability and trustworthiness.
• This approach shows significant promise for improving early breast cancer detection and management.