Teaching neurosurgery with AI

Algorithms can detect performance errors and suggest ways to improve

Shawn Hayward | 12 Aug 2019

Becoming a neurosurgeon is an incredibly demanding and difficult education. Even the smallest hand movement during surgery can be the difference between success and failure, with lifelong implications for the patient.

A research team at The Neuro has shown AI guided virtual reality simulators can be powerful tools to help surgeons reach their full potential. This could accelerate their progress to becoming outstanding surgeons by allowing them to practice on simulators before stepping into the operating room.

Researchers from the Neurosurgical Simulation and Artificial Intelligence Learning Centre made 50 participants in different stages of neurosurgical training perform 250 complex tumour resections using a VR simulator. Using data collected from these operations, algorithms could develop performance measures such as instrument position and force applied, as well as outcomes such as amount of tumour removed and blood loss, which could predict the level of expertise of each participant with 90 per-cent accuracy. The top performing algorithm could classify participants using just six performance measures.

This study shows that the fusion of AI and VR neurosurgical simulators can accurately and efficiently assess the performance of surgeon trainees. AI-assisted mentoring systems can be developed that focus on improving patient safety by guiding trainees through complex surgical procedures.

Click here for the fully study text, published in The Journal of the American Medical Association.

Link to the article: https://www.mcgill.ca/neuro/article/research/teaching-neurosurgery-ai