Author: Mauro Carrara, Olivera Ciraj Bjelac, John E. Damilakis, Andre L. Dekker, Serafina Di Gioia, Renato Padovani, Egor Titovich, Qingrong Jackie Wu 👨🔬
Affiliation: University of Crete, Duke University Medical Center, Maastro Clinic, Dosimetry and Medical Radiation Physics Section, Division of Human Health, International Atomic Energy Agency, International Centre for Theoretical Physics 🌍
Purpose: The purpose of this work is to present the International Atomic Energy Agency (IAEA) activity in providing medical physicists (MPs) with knowledge, skills, and competencies to support the safe and efficient application of Artificial Intelligence (AI)-based systems in radiation medicine, including commissioning and routine quality control. The efforts are aligned with the IAEA publication TCS-83 Artificial Intelligence in Medical Physics (TCS-83) [1], recommending various continuing professional development (CPD) activities addressing different levels of clinical implementation and use of AI.
Methods: Knowledge of the main theoretical principles for the clinical implementation of AI in radiation medicine is considered essential for MPs involved in any AI application across medical physics sub-specialties. Therefore, content for a one-week course focused on theoretical AI principles from a user’s perspective was designed. The course was first delivered as a joint ICTP-IAEA Workshop in 2023 in Trieste, Italy, to early- and mid-career MPs from various backgrounds, and their feedback was collected.
Results: The course features 18 hours of net lectures and 6 hours of practical sessions. Key topics covered in the course, along with the individual lectures delivered, are listed in Table 1. Practical sessions, which are essential for a better understanding of the presented topics, might be adapted according to the background and needs of participating MPs. The joint ICTP-IAEA Workshop was attended by 59 MPs working in hospitals in 50 IAEA Member States. Based on the feedback from the participants, who rated the workshop with an overall grade of 4.56 out of 5, the course fulfils the needs outlined in TCS-83.
Conclusion: The course content enables participants to understand the principles behind AI-based technology and critically assess AI systems' output. Organisation of similar courses by professional organizations or other training providers is encouraged. The IAEA is considering creating an e-learning module.