Deep Learning Model for Invasive Ductal Carcinoma Detection in Histopathology Images

Breast cancer, with Invasive Ductal Carcinoma (IDC) as its most prevalent subtype, presents substantial challenges in early detection and diagnosis. This study introduces a novel computer-aided diagnosis (CAD) system designed to detect IDC in histopathology images with high precision. Leveraging a convolutional neural network (CNN), the system incorporates advanced methodologies such as sliding window-based heatmaps and a unique oversampling technique that extracts patches from homogenous regions in stitched whole-slide images. This approach enhances minority class representation while maintaining biological integrity, addressing class imbalance effectively. The system achieved a balanced accuracy of 89.06% and an F1-score of 86.68%, outperforming existing models in the literature. The implementation of sliding heatmaps further improves interpretability, enabling pathologists to visualize model predictions seamlessly. This research demonstrates the potential of combining innovative deep learning techniques with domainspecific insights to enhance IDC detection, offering a reliable tool for clinical practice and contributing to improved patient outcomes.