State recognition of decompressive laminectomy with multiple information in robot-assisted surgery

Area of Science:

• Robotics in Surgery
• Surgical State Recognition
• Neurosurgery Technology

Background:

• Decompressive laminectomy is a common procedure for lumbar spinal stenosis.
• Robot-assisted telesurgery offers potential benefits but poses challenges due to limited surgeon perception.
• State recognition is crucial for enhancing safety and decision-making in remote robotic surgery.

Purpose of the Study:

• To develop and validate a novel state recognition system for robot-assisted telesurgery.
• To improve surgeon awareness and decision support during critical surgical maneuvers.
• To enhance safety in robot-assisted decompressive laminectomy procedures.

Main Methods:

• An image-based state recognition method using a U-Net derived network for the fenestration (grinding) phase.
• A grayscale redistribution and dynamic receptive field approach to control grinding bit trajectory.
• An audio and force-based state recognition method for internal fixation (drilling) using signal feature extraction and LSTM prediction.

Main Results:

• The proposed system accurately recognizes surgical states during fenestration and internal fixation.
• The image-based method prevents the grinding bit from damaging spinal nerves.
• The audio and force-based method prevents the drilling bit from damaging the vertebral pedicle.

Conclusions:

• The developed state recognition system reliably enhances safety in robot-assisted telesurgery.
• Combining learning and traditional methods allows the robot to provide surgeon-like insights.
• This technology aids surgeons in performing safer remote-controlled procedures.