AI can be a powerful substitute teacher for the next generation of medical students

AI can be a powerful substitute teacher for the next generation of medical students

With an increasingly acute shortage of surgeons, artificial intelligence could help fill this gap by coaching medical students to practice surgical techniques.

A new tool based on videos of experienced surgeons at work offers students personalized advice in real time as they practice suturing. Initial experiments suggest that AI can be a powerful substitute teacher for more experienced students.

"We are at a crucial time. The shortage of providers is getting worse and we need to find new ways to provide more and better opportunities for practice. Right now, an attending surgeon, who is already short on time, has to come in and watch students practice, evaluate them and give them detailed feedback - that is simply not scalable," said lead author Mathias Unberath, an expert in AI-assisted medicine who focuses on how humans interact with AI. “The next best thing might be our explainable AI that shows students how their work differs from the work of experienced surgeons.”

The groundbreaking technology developed at Johns Hopkins University was recently presented and recognized at the International Conference on Medical Image Computing and Computer Assisted Intervention.

Currently, many medical students watch videos of experts performing surgeries and try to imitate what they see. There are even existing AI models that evaluate students, but Unberath says they fall short because they don't tell students what they're doing right or wrong.

“These models can tell you whether you have high or low ability, but they have difficulty telling you why,” he said. “If we want to enable meaningful self-training, we need to help learners understand what they need to focus on and why.”

The team's model involves so-called "explainable AI," an AI approach that - in this example - assesses how well a student closes a wound and then tells them exactly how they can improve.

The team trained their model by tracking the hand movements of experienced surgeons as they closed incisions. When students attempt the same task, the AI ​​immediately texts them to let them know how they did compared to an expert and how they can refine their technique.

Learners want someone to objectively tell them how they did it. We can calculate their performance before and after the procedure and see if they are approaching expert practice.”

Catalina Gomez, lead author, PhD student in computer science at Johns Hopkins University

The team conducted a first-of-its-kind study to find out whether students learn better through AI or by watching videos. They randomly assigned twelve medical students with experience in suturing using one of the two methods.

All participants practiced closing an incision with sutures. Some received instant AI feedback, while others tried to compare their work to a surgeon in a video. Then everyone tried to sew again.

Compared to students who watched videos, some AI-trained students learned much faster with more experience.

“For some people, AI feedback has a big effect,” said Unberath. “Beginners still had difficulty with the task, but students with solid surgical knowledge who are at the point where they can implement the advice had a big impact.”

Next, the team plans to refine the model to make it easier to use. They hope to eventually create a version that students can use at home.

“We want to provide computer vision and AI technology that allows someone to practice from the comfort of their own home with a sewing kit and a smartphone,” Unberath said. "This will help us expand education in the medical field. It's really about how we can use this technology to solve problems."

Authors include Lalithkumar Seenivasan, Xinrui Zou; Jeewoo Yoon; Sirui Chu; Ariel Leon; Patrick Kramer; Yu Chun Ku; Jose L. Porras; and Masaru Ishii, all from Johns Hopkins, and Alejandro Martin-Gomez from the University of Arkansas.

Sources: