Radiology residents appreciate, benefit by in-house AI training; attendings hungry too but may lack nonclinical time
Radiology residents who completed an intensive, single-day workshop in artificial intelligence came away reporting significantly improved understanding of the technology.
Those who opted for a seven-month, seven-lecture series presenting the same material registered similarly positive feedback.
Attending radiologists were invited to participate as well, but their survey responses were not analyzed because of high attrition rates between course signup and completion.
The experimental AI curriculum, called AI-RADS, was created and conducted at Dartmouth Health and its Geisel School of Medicine in New Hampshire.
The project is described in a study published Jan. 4 in Frontiers in Medical Technology [1].
Internal medicine resident Alexander Lindqwister, MD, radiologist Jessica Sin, MD, PhD, and colleagues designed the course to cover primary ground—e.g., “foundational algorithms”—before taking on secondary topics like data representation and abstraction.
Attending radiologists consider AI education important
In post-program surveying, the researchers found residents edified by the curriculum in both short- and long-form formats.
The seven-month version earned higher scores for overall satisfaction—average 9.8 out of a possible perfect 10—although the one-day workshop was well received too, averaging 4.3 out of a possible 5.
Moreover, both versions achieved meaningful increases in perceived understanding of AI.
Lindqwister and co-authors surmise the high attrition rate among attending radiologists may reflect “a combination of the high clinical burdens placed on attending radiologists, limits in academic time compared to trainees or other factors.”
Interestingly, attendings rated the importance of AI education higher (although not statistically significantly so) as compared to residents.”
Results, feedback ‘suggest a potential new way forward’ in AI training
The authors remark that AI’s growing role in medical imaging demands commensurate familiarity with the technology by radiologists.
“Indeed, it is not unrealistic to anticipate core concepts in machine learning to become a fundamental aspect of radiologist training analogous to magnetic resonance physics,” they write. “There is both significant want and pragmatic need for radiologists to understand these techniques.”
The promising results of both the longitudinal AI-RADS curriculum as well as the condensed single-day version are suggestive of a potential new way forward in engaging trainees with this material.”
The journal has posted a detailed description of the project in full for free, and the American College of Radiology is offering iterations of individual lectures from the work.
Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.