Clinical Impact of Artificial Intelligence on the MRI Physicist’s Domain
Description
MRI is increasingly integral across both diagnostic imaging and radiation therapy, yet quality assurance (QA) practices remain fragmented and often limited to phantom-based methods. While phantoms provide essential system checks, they cannot fully capture patient-specific variability such as geometric distortions, field inhomogeneities, or anatomical dependencies that directly influence diagnostic accuracy and treatment planning. This symposium will explore the next generation of MRI QA, emphasizing a hybrid framework that integrates phantom-based methods with patient-specific approaches. Speakers will address current limitations in routine MRI QA, discuss the role of patient-derived data in complementing phantom measurements, and highlight emerging strategies for automated and AI-driven QA. Topics will include automated distortion analysis, B0 mapping, synthetic CT generation, and AI-based correction methods that link scanner performance with clinical relevance. By considering both diagnostic and therapeutic perspectives, this session will provide a unified view of MRI QA that is scalable, clinically actionable, and adaptable across practice settings. The audience will gain insight into how automation and artificial intelligence can strengthen traditional QA practices, reduce manual burden, and ultimately improve the reliability of MRI for both diagnosis and treatment. This session aims to inspire a cross-disciplinary dialogue on advancing MRI QA from phantom-only validation toward a patient-inclusive, automated, and AI-enabled paradigm—building bridges between diagnostic imaging and therapy applications for a more consistent and clinically impactful future.