Advances of Deep Learning in Healthcare from Diagnosis to Decision Support

Abstract

Deep learning(DL)has rapidly emerged as a core enabling technology in modern healthcare,offering transformative potential in disease diagnosis,treatment planning,prognosis prediction,and clinical decision support.Leveraging powerful neural architectures such as convolutional neural networks,recurrent networks,and transformers,DL algorithms have demonstrated state-of-the-art performance across diverse medical domains,including imaging,biomedical signal processing,and genomic analysis.This review presents a comprehensive overview of DL techniques in healthcare,spanning foundational models,application-specific methodologies,benchmark datasets,and performance metrics.Key challenges such as data scarcity,model interpretability,ethical concerns,and deployment barriers are critically examined.In addition,we explore future research directions including multimodal learning,federated frameworks,and trustworthy AI practices.This survey aims to provide researchers and practitioners with a cohesive understanding of the current landscape and future potential of deep learning in advancing intelligent,equitable,and reliable healthcare systems.