AI-Powered Framework for Early Breast Cancer Detection and Classification Using Image Analysis

Abstract

Breast cancer remains one of the most prevalent and life-threatening diseases among women worldwide. Early detection plays a vital role in improving survival rates, and artificial intelligence (AI) has emerged as a transformative force in enhancing diagnostic accuracy through image analysis. This systematic review examines the latest AI-powered frameworks designed for early breast cancer detection and classification, with a focus on radiological and pathological imaging modalities, including mammography, ultrasound, MRI, and histopathology. A comprehensive literature search was conducted across leading digital libraries, including PubMed, IEEE Xplore, ACM Digital Library, Springer, Elsevier, and Google Scholar, to cover peer-reviewed studies from 2015 to December 2025. Following PRISMA guidelines, we selected works that employed machine learning, deep learning, hybrid architectures, and explainable AI models. In addition to traditional convolutional neural networks (CNNs), this review highlights cutting-edge advancements, including Vision Transformers (ViTs), self-supervised learning, federated learning for privacy-preserving diagnostics, and multimodal fusion techniques that combine image and textual data. We systematically compare AI techniques, datasets, performance metrics, and experimental settings while identifying common limitations in clinical deployment, including model interpretability, data heterogeneity, and integration with existing healthcare workflows. This review offers insights into the evolution of AI-based diagnostic systems. It outlines critical future directions, aiming to guide researchers, engineers, and clinicians in developing robust, transparent, and clinically applicable solutions for breast cancer diagnosis and treatment.