During endoscopic endonasal neurosurgery, neurosurgeons grind the sphenoid bone using a high-speed surgical drill to access skull-base tumours. However, this procedure exposes them to the risk of developing symptoms of musculoskeletal disorders, thereby negatively impacting their overall performance.

Robotics could enhance ergonomics and reduce fatigue; however, current systems lack adequate grinding tools and feedback to mitigate the risk of soft tissue and nerve damage due to bone grinding overshoot.

To prevent bone grinding overshoot, we are currently developing a machine-learning-based bone layer segmentation model, capable of discerning the bone’s different layers while grinding in a constrained environment, such as in the context of endoscopic endonasal neurosurgery.