The Role of Artificial Intelligence in PET Imaging

Artificial intelligence (AI) is transforming positron emission tomography (PET) across the clinical workflow. This article reviews current clinical applications, including automated scheduling and protocoling, Food and Drug Administration-cleared deep learning reconstruction, and commercial lesion detection. Emerging research includes physics-informed reconstruction, diffusion models, and large language models for automated reporting. Implementation challenges include algorithmic bias, data drift, human-AI interaction, and evolving regulatory frameworks. Future directions encompass transformer architectures, total-body PET integration, and multimodal fusion. Understanding both the capabilities and limitations of these technologies can help radiologists and nuclear medicine physicians evaluate their clinical utility and integrate AI responsibly into patient care.