Brigham & Women’s Hospital has designed a data science pathway that can prepare radiology residents to lead the next era of artificial intelligence development and implementation.
Radiology has always been the technological trailblazer in healthcare. One academic medical institution is looking to keep that momentum going by training fourth-year radiology residents to be the leaders in artificial intelligence and machine learning (AI-ML).
Currently, there are few organized, detailed AI-ML curricula, and there is also a dearth of training opportunities. Radiology residents and faculty at Brigham & Women’s Hospital designed a data science pathway to address this need.
“Across the nation, there are a number of radiology residency programs that are trying to figure out how to integrate AI into their training,” said Walter F. Wiggins, M.D., Ph.D. “We thought that perhaps our experience would help other programs figure out ways to integrate this type of training into either their elective pathways or their more general residency curriculum.”
The published both the design of and their experience with the pathway in the Nov. 5 Radiology: Artificial Intelligence.
Ultimately, Wiggins’ team said, data science has the potential to reinvigorate radiology as a specialty. Consequently, radiologists-in-training need to be familiar with it so they can shepherd the development of machine learning models and successfully integrate them into clinical care. A data science pathway can give them an immersive experience through a flexible schedule of educational, experiential, and research activities, they said.
At Brigham & Women’s, three residents contributed to designing the pathway, proposing three-, six-, and eight-month pathway plans rooted in their individual learning goals while safe-guarding time for advanced clinical electives. As a result, the pathway combined formal instruction and practical problem-solving collaborations with data scientists, exposing residents to all aspects of AI-ML application development, including data curation, model design, quality control, and clinical testing.
Individual AI-ML Projects from the DSP. Each trainee contributed to design, data curation and model development of individual projects including hemorrhage detection on CT (A), abdominal body composition (B), and lumbar spine segmentation and stenosis assessment (C). Courtesy: Radiological Society of North America
Throughout the experience, Wiggins’ team said, residents worked with data scientists to get a better understanding of how they approach image analysis problems with ML tools. They also offered input to show data scientists how radiologists approach radiology problems in the clinical setting.
“An important component of a curriculum like this is to learn the language the data scientists speak and teach them a little bit about the language that we, as radiologists, speak so that we can have better, more effective collaboration,” said Wiggins, who recently accepted a position as the clinical director of AI at Duke Radiology in Durham, N.C. “Going through that process over several different projects was where I think I gained the best experience throughout all of this.”
Overall, residents gained a significant amount of first-hand experience, as well, that will prepare them for AI use and AI leadership in the future, the team said. Not only did they learn the programming languages commonly used in AI-ML, they also learned the strategic and financial considerations of AI-ML implementation by participating in meetings with non-clinical stakeholders. In fact, one resident played a key role in orchestrating the clinical testing of algorithms before they were staged for widespread clinical deployment.
By contributing to the various stages of model and tool development, residents produced 12 abstracts during the pilot period that were accepted for presentation at national meetings, including the Society for Imaging Informatics in Medicine, the Society of Abdominal Radiology, and the International Stroke Conference. Six additional abstracts are in the works, the team said.
Based on resident feedback, Brigham & Women’s established a formal AI-ML curriculum for future residents that will include near-peer mentoring for junior residents, as well as help with proposal-writing, goal-setting, and networking. This next program iteration will also include tutorial, hands-on, live, and video-taped didactic sessions, as well as background literature references and links to online resources.
Ultimately, Wiggins said, by training its residents, radiology can continue to lead the way in integrating AI into more widespread clinical use.
“Radiologists have always had to manage, analyze, and process data in order to be able to do their work," he said. "We already have the underlying skill sets and infrastructure that we can tap into to allow residents with an interest in AI and ML to really develop and become leaders in applying these skills clinically."
For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-newsletter here.