HistoBreast CAD System: Un modelo computacional para la clasificación automática de imágenes histopatológicas de mama.

Abstract

Abstract: Breast cancer is one of the leading causes of mortality among women worldwide, and histopathological analysis remains the gold standard for its diagnosis. However, a large part of the diagnostic process relies on the expertise of pathologists and is subject to interobserver variability and high operational workload. In this context, the present work aims to develop and evaluate a computer-aided diagnosis system for breast cancer histopathology based on artificial intelligence techniques. A CAD system is proposed that integrates segmentation and classification models applied to whole-slide histopathological images. The system is based on a unified architecture that combines segmentation and classification into a single inference pipeline, unlike other methodologies that treat these tasks as independent problems. The methodology includes data preprocessing, automatic segmentation of regions of interest, multiclass tissue classification, and comparative validation against specialist interpretation. Additionally, a web-based application is implemented to visualize the system s predictions through heatmaps overlaid on the original images. The results obtained demonstrate that the employed models achieve consistent performance and that the developed solution can support the diagnostic process. Overall, this work confirms the feasibility of integrating artificial intelligence into CAD systems for digital pathology.