AI screening of nuclear medicine safety breaches: patterns, causes, and opportunities for improved protocols: a systematic review.
Authors
Affiliations (6)
Affiliations (6)
- Research Laboratory of Biophysics and Medical Technologies, University of Tunis El Manar, The Higher Institute of Medical Technologies, Tunis, Tunisia.
- Radiation Effects Department, Radiation, Chemicals, Climate and Environmental Hazards Directorate, UK Health Security Agency, Harwell Campus, Chilton, Didcot, Oxfordshire, OX11 0RQ, UK.
- Nuclear Medicine Research Center, Mashhad University of Medical Sciences, Mashhad, Iran.
- Université Côte d'Azur, 06100, Nice, France. [email protected].
- College of Arts & Sciences, Gulf University for Science and Technology, Hawally, Kuwait. [email protected].
- Research Laboratory of Biophysics and Medical Technologies, University of Tunis El Manar, The Higher Institute of Medical Technologies, Tunis, Tunisia. [email protected].
Abstract
Nuclear medicine differs from other specialties of radiology by employing unsealed radionuclides. Moreover, it may heighten the risks of incidents for nuclear medicine healthcare professionals (NMHP). On the other hand, artificial intelligence (AI) methods improve their ability to assess, understand, and prevent these incidents. This systematic review examines the critical incidents affecting NMHP and reviews the potential of AI in improving, controlling, and evaluating the occupational exposure, to predict and prevent these accidents. A systematic search of PubMed, Science Direct, Scopus, and the NLM was conducted using the keywords and Mesh terms, with no language restrictions. A protocol based on PRISMA guidelines was developed. To streamline both the search strategy and the study selection process, EndNote X7.8 was employed. 49 studies were reviewed. The primary causes of incidents in nuclear medicine are due to inadequate handling of radionuclides, malfunctioning equipment, and the loss or theft of radioactive sources. Furthermore, our research highlights the potential of AI algorithms to facilitate better identification of radioactive sources, radiation dose optimization, and strengthen the decision-making processes during potentially hazardous incidents. Our systematic study intervenes to improve the role of AI in the surveillance and improvement of the occupational exposure situation for NMHP. In addition, AI tools can contribute to better decision-making in real time during nuclear medicine emergency situations. Such advancements underscore the crucial need for ongoing development and implementation of AI technologies in nuclear medicine to enhance radiation protection for NMHP.