Abstract

The integration of artificial intelligence (AI) into diagnostic imaging has emerged as a transformative force in modern healthcare. This systematic review synthesizes the current state of deep learning applications for early disease detection across multiple imaging modalities, including radiology, pathology, and dermatology. We analyzed 127 peer-reviewed studies published between 2020 and 2026, evaluating model performance, clinical validation, and translational challenges. Our findings indicate that convolutional neural networks (CNNs) and vision transformers (ViTs) have achieved diagnostic accuracy comparable to or exceeding that of expert clinicians in specific tasks, particularly in mammography, chest radiography, and skin lesion classification. However, significant barriers remain, including data heterogeneity, lack of external validation, and limited interpretability. We propose a framework for standardized reporting and clinical implementation to bridge the gap between AI research and real-world healthcare delivery.