Advances and Perspectives of AI in Nuclear Medicine: From Image Analysis to Theranostics

Abstract

Recent advancements in artificial intelligence (AI), particularly deep learning, have led to significant progress in both diagnostic and therapeutic applications in nuclear medicine, including image quality enhancement, diagnostic support, dose calculation, and treatment response prediction. Image processing techniques using convolutional neural networks and generative adversarial networks have contributed to reduced radiation exposure and shortened acquisition times, while tumor segmentation-based approaches are expected to improve the efficiency of personalized medicine. However, challenges remain, such as the scarcity of annotated nuclear medicine image datasets, variations in imaging protocols across institutions, and the necessity of incorporating explainable AI to ensure transparency and interpretability. This article outlines recent advances in AI technologies in the field of nuclear medicine, discusses their clinical applicability, and addresses current challenges and future perspectives.