2026 Proffered Presentations
S175: THE ROLE OF ARTIFICIAL INTELLIGENCE IN ENHANCING MODERN NEUROSURGICAL TRAINING AND SIMULATION
Andre Payman, MD; Marco Obersnel, MD; Chiara Angelini, MD; Hao Tang, MD; Roberto Rodriguez Rubio, MD; UCSF
Introduction: Traditional neurosurgical training, reliant on cadaveric and synthetic models, faces limitations in case reproducibility and opportunities for repetitive practice. While virtual reality (VR) simulation has offered a risk-free environment for surgical practice, skill assessment has remained largely subjective. This work explores the integration of artificial intelligence (AI) to create objective, data-driven methods for evaluating and enhancing neurosurgical skill acquisition.
Methods: This study reviews the application of various machine learning (ML) models in analyzing performance metrics from VR neurosurgical simulators, specifically focusing on subpial tumor resection and spine surgery simulations. Algorithms such as K-nearest neighbor (KNN), support vector machines (SVM), artificial neural networks (ANN), and long-short-term memory (LSTM) networks were employed to classify user skill levels. Performance data, including instrument handling, force application, tissue contact, and task completion time, were collected from participants ranging from medical students to expert neurosurgeons. The development and testing of an AI-driven feedback tool, the Virtual Operative Assistant (VOA), is also described.
Results: ML algorithms demonstrated high accuracy in distinguishing between different levels of surgical expertise. In simulated subpial tumor resection, a KNN algorithm achieved 90% accuracy in classifying participants. Similarly, in a VR spine surgery simulation, an ANN model achieved 83.3% accuracy. The AI-powered VOA provided personalized, automated feedback that led to significant improvements in trainee performance, with the VOA group outperforming control groups who received feedback from expert instructors or no feedback at all. AI-derived metrics also successfully generated learning curves, providing detailed insights into the skill acquisition process over time.
Conclusion: The integration of AI into neurosurgical simulation provides a robust and objective framework for skill assessment and training. AI-driven platforms can accurately classify surgical expertise, track learning progression, and deliver personalized feedback more effectively than traditional methods. These technologies hold the potential to revolutionize neurosurgical education by standardizing training curricula and accelerating the development of technical proficiency in a risk-free environment. Broader adoption of these AI-enhanced models is essential for advancing the field of neurosurgical education.