Experts highlight core concepts that could help standardize AI curriculum
As artificial intelligence continues to cement itself into radiology workflows, a new analysis offers a detailed look at how radiology programs are adapting curriculum to reflect the impending changes.
Published in the Journal of the American College of Radiology, the new paper includes data from five different published studies on the topic of AI curriculum in residency. While the paper did not signal preference for any one style of curricula, it did highlight several common areas addressed during the sessions, along with a common theme that was evident in every study—that is, that AI curriculum is resoundingly well received by those who complete it.
This was consistent in every study included in the analysis. However, something that was not consistent was the way in which these educational courses took place—something that the authors of the new paper suggest is a need that should be addressed in the near future.
“Advancements in AI in radiology have prompted interest in formalizing AI residency education with many calling for a standardized education around the subject. Although the need for an AI curriculum is well recognized, descriptions of specific AI curricula for radiology trainees have been limited,” corresponding author Paul H. Yi, MD, with the University of Maryland Medical Intelligent Imaging (UM2ii) Center, and colleagues noted.
The team was able to identify a few similarities between the papers. In all five studies, AI courses were taught by radiologists, mostly in academic radiology sessions. Course designs included didactic sessions (100%), assigned readings (80%), hands-on learning (60%) and journal clubs (60%), and one involved individualized learning plans relative to AI.
From their analysis, the authors identified several core concepts that should be considered when structuring AI curriculum. These include fundamental computer science and artificial intelligence concepts (mathematical theory, machine learning, deep learning, neural networks, etc.), AI model training and evaluation, and clinical implementation.
These concepts could be used as a starting place for AI curriculum and would provide a foundation for eventually scaling such courses, the authors suggested.
“Not only will this benefit radiology residents who participate in such curricula, but it will also create a standard for the entire field moving forward as we prepare ourselves and future radiologists for the AI era.”
The study abstract can be viewed here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.