Machine learning can classify participants into levels of expertise with 90 percent accuracy in a virtual reality neurosurgical tumor resection simulation, according to a study published online in JAMA Network Open.

Alexander Winkler-Schwartz, M.D., from McGill University in Montreal, and colleagues recruited 50 participants from a single university to identify surgical and operative factors selected by a machine learning algorithm to accurately classify participants by level of expertise in a virtual reality surgical procedure. Individuals were classified as experts, seniors, juniors, and medical students, all of whom participated in 250 simulated tumor resections. Group membership was determined using different algorithms that selected performance metrics associated with instrument movement and force, resection of tissues, and bleeding generated from the raw simulator data output.

The researchers found that accuracy was 90 percent for the K-nearest neighbor algorithm, 84 percent for the naive Bayes algorithm, 78 percent for the discriminant analysis algorithm, and 76 percent for the support vector machine algorithm. The algorithms used six, nine, eight, and eight performance metrics, respectively. Overall, two neurosurgeons, one fellow or senior resident, one junior resident, and one medical student were misclassified.

"Our study demonstrates the ability of machine learning algorithms to classify surgical expertise with greater granularity and precision than has been previously demonstrated," the authors write.

Several authors disclosed financial ties to the medical device industry.

-----------------------------------------------------------------------
Editor's Note:

If you have any suggestion to the content, 

please email: Julia.Zhang@ubmsinoexpo.com