Every surgical training program has a time when a resident completes a procedure and then waits, sometimes for days, to receive any meaningful feedback. The attending surgeon might have been preoccupied. Perhaps the notes were ambiguous. Perhaps the criticism was influenced more by the preferences of a single instructor than by a standard. For years, Keenan Gibson, a resident in vascular surgery at UC Davis Health, was a part of that system. And he made the decision to construct something better somewhere between Sacramento and Accra.
The Association of Program Directors in Vascular Surgery recently awarded Gibson a grant to create an AI system based on computer vision that can watch surgical trainees perform procedures and provide structured, unbiased feedback on their technique. There is no need for costly equipment, a specialist watching over you, or institutional infrastructure because the system is made to function with just a smartphone and a downloadable app. It’s a sophisticated solution to a truly messy issue.
No one in the field freely acknowledges that the feedback loop in surgical education has always been inconsistent. Hospitals, teachers, and cities all have different standards. What is commended in one residency program may be criticized in another. According to Gibson, there isn’t a single, widely recognized technique for evaluating technical proficiency in all surgical specialties. That’s a big difference. This is a structural failure that has been developing for decades, and it raises concerns about whether some surgeons are being certified based more on their training than their actual abilities.
The origins of Gibson’s ideas are part of what makes them intriguing. He went on a medical mission trip to Ghana earlier this year, and the conversations he had there obviously changed him. There is no official vascular surgery training program in Ghana. From inside a Sacramento hospital, it is challenging to fully understand how access to specialized care is restricted. In order to provide trainees with something to practice on, Gibson began creating and 3D printing inexpensive surgical models that could be shipped abroad. However, he discovered that practice without feedback doesn’t really bridge the skills gap. It merely passes the time.
That insight motivated him to pursue AI. His model will evaluate the trainee’s technique, analyze video recordings of simulated procedures, and provide consistent, useful feedback. This kind of system could be very important in situations where skilled surgeons are just not available to regularly oversee training. Although the model’s precise accuracy across various skill levels and procedure types is still unknown—validation work is still to come—the fundamental idea is difficult to refute.
In the future, Gibson hopes to incorporate the system into the board certification procedure for vascular surgeons, which could offer an impartial technical standard that is lacking from existing evaluations. Additionally, there is early research on real-time procedural guidance, where the AI is trained using 3D-printed patient-specific anatomy before surgeons even enter the operating room. It’s too soon to tell if everything works out the way he wants it to.
What is evident is that the issue Gibson is tackling—the subtle disparity ingrained in surgical education across the globe—deserves precisely this level of ongoing focus. Improved training results in more than just better surgeons. Patients benefit from it because they never know how the surgeon above them became so competent, confident, and well-prepared.
