Seeds planted for a needs-based, radiology-specific AI curriculum

There’s no shortage of educational resources for teaching radiologists at all learning levels the principles and particulars of medical AI, but the need for radiology-specific materials is pronounced, according to a study published June 23 in Academic Radiology [1].

The research was conducted at Children’s Hospital of Philadelphia and accordingly targeted for a pediatric audience, although its key findings are germane across radiology.  

Corresponding author Maria Camila Velez-Florez, MD, and colleagues convened a small focus group comprising radiology residents, radiology fellows, MDs applying for radiology residency and an attending radiologist.

The researchers chose the direct participant-input format, one centered in semistructured, moderator-guided interviews, as a first step toward developing an AI curriculum based on the expressed needs of radiology trainees at their institution.

Project leaders and study co-authors included a pediatric radiology fellow, two AI research fellows and two pediatric radiologists experienced in care, teaching and research.

‘Incomplete Understanding of Radiology AI Hinders Its Clinical Applicability’

Querying participants to uncover their perceptions of artificial intelligence in general, perceived competence in interpretation of medical AI literature and expectations from radiology AI educational programs, the team found that most participants:

  1. prefer a case-based approach to teaching AI, similar to the standard techniques used for radiology resident training;
  2. believe incomplete understanding of AI hinders its clinical applicability; and
  3. recognize a marked need for improved training in the interpretation and application of AI literature.

They also found a broad diversity of opinions, perceptions and perspectives on AI even though the cohort contained just seven participants.

Any successful solution to build an AI curriculum “must account for the wide range of these interests and needs,” the authors comment in their discussion. “The patterns of experience which emerge from assessing trainee perspectives should serve as a guide for AI education and future directions.”

‘It’s Too Complicated, And Later We Don’t Know How to Apply It in Our Clinical Practice’

Sample remarks offered by participants during the interviews and included in the study report:

  • “When you try to take this information and try to understand it in the context of neuroradiology, it starts to be very difficult. Then you go to the methodology and it gets very difficult to understand and interpret. It’s not so easy for me to read a paper on AI because it’s too complicated, and later we don't know how to apply that in our clinical practice.”

 

  • “I have learned most of my AI through residents who chose to launch off and learn from their colleagues but needed a mentorship. So I have collaborated with others on AI projects. That’s where I mostly learn my background knowledge in AI.”

 

  • “It would be helpful to know how much I need to learn from the IT (information technology) part of it. The question is do I need to learn coding as a radiologist and to what degree. I don't know how deep we need to go into [it].”

 

  • “We need a manual or guide for residents and attendings, with real examples just explaining on the figures what it is. Showing the real cases and, based on that fundamental knowledge, we can go from that and work on real projects.”

‘Improved Learning Materials for Radiologists Should Be a Priority in Trainee Education’

Velez-Florez and co-authors conclude that, although numerous resources for learning the general aspects of AI exist, “the current lack of radiology-specific materials argues that improved learning materials for radiologists should be a priority in trainee education.”

In particular, they add, “education through case-based scenarios could serve as a foundation for engaging a variety of trainees from different backgrounds and interest levels, while ensuring the necessary baseline level of competency in AI.”

 

More Coverage of Medical Imaging AI:

4 increasingly common uses for imaging AI in the ED

Ultrasound-embeddable AI sharp at diagnosing clogged carotid arteries

AI mammography, prostate imaging algorithms cleared for market

2 imaging AI vendors teaming up to present 1 point of sale

AI reads of neck ultrasounds could displace thyroid biopsies

 

Reference:

  1. Maria Camila Velez-Florez, Susan Sotardi, et al., “Artificial Intelligence Curriculum Needs Assessment for a Pediatric Radiology Fellowship Program: What, How, and Why?” Academic Radiology, June 23, 2022. DOI: https://doi.org/10.1016/j.acra.2022.04.026
Dave Pearson

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.

Around the web

The patient, who was being cared for in the ICU, was not accompanied or monitored by nursing staff during his exam, despite being sedated.

The nuclear imaging isotope shortage of molybdenum-99 may be over now that the sidelined reactor is restarting. ASNC's president says PET and new SPECT technologies helped cardiac imaging labs better weather the storm.

CMS has more than doubled the CCTA payment rate from $175 to $357.13. The move, expected to have a significant impact on the utilization of cardiac CT, received immediate praise from imaging specialists.