Breast Cancer Classification Using Synthesized Deep Learning Model with Metaheuristic Optimization Algorithm

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Breast cancer remains a leading cause of mortality in women, underscoring the need for early and accurate detection.
• Traditional methods for breast cancer detection and classification have limitations.
• Convolutional Neural Networks (CNNs) show promise in enhancing medical image analysis for tumor identification.

Purpose of the Study:

• To develop a comprehensive classification technique for breast cancer detection using a synthesized CNN and an enhanced optimization algorithm.
• To assist radiologists in the rapid and accurate identification of breast cancer anomalies.
• To improve the accuracy, sensitivity, and specificity of breast cancer classification in mammographic datasets.

Main Methods:

• A novel Enhanced Ant Colony Optimization (EACO) algorithm, modified with opposition-based learning (OBL), was developed to optimize CNN hyperparameters.
• The EACO algorithm was integrated with the Residual Network-101 (ResNet101) CNN architecture, creating the EACO-ResNet101 model.
• The proposed model was evaluated on the MIAS and CBIS-DDSM mammographic datasets.

Main Results:

• The EACO-ResNet101 model achieved high performance on the CBIS-DDSM dataset, with 98.63% accuracy, 98.76% sensitivity, and 98.89% specificity.
• On the MIAS dataset, the model demonstrated 99.15% accuracy, 97.86% sensitivity, and 98.88% specificity.
• The proposed model significantly outperformed conventional methods in breast cancer classification.

Conclusions:

• The developed EACO-ResNet101 model offers a superior approach for breast cancer classification in mammography.
• This AI-driven technique has the potential to enhance diagnostic accuracy and support clinical decision-making for radiologists.
• The findings highlight the effectiveness of combining advanced optimization algorithms with deep learning for medical image analysis.