Autonomous OCT-Guided Robotic System for Vertical Needle Insertion and Big-Bubble Dissection in Corneal...

Area of Science:

• Ophthalmology
• Robotics
• Medical Imaging

Background:

• Deep anterior lamellar keratoplasty (DALK) requires high precision to separate corneal layers without damaging Descemet's membrane (DM).
• Conventional manual techniques carry risks of incomplete dissection or DM perforation, impacting surgical outcomes.
• Autonomous surgical systems offer potential for improved consistency and safety in complex procedures.

Purpose of the Study:

• To introduce AUTO-DALK, an autonomous robotic system utilizing real-time optical coherence tomography (OCT) sensing for precise needle insertion in DALK.
• To evaluate the efficacy of a sensor-guided vertical needle trajectory for uniform pneumodissection compared to traditional methods.

Main Methods:

• Development of AUTO-DALK, integrating an OCT fiber optic sensor with an anatomical segmentation model for real-time depth perception and closed-loop needle control.
• Evaluation of the system through ex vivo and in vivo rabbit models, comparing autonomous robotic insertion against freehand, OCT-guided manual, and teleoperated robotic methods.

Main Results:

• AUTO-DALK demonstrated superior performance in needle placement accuracy and dissection precision.
• The autonomous system achieved significantly faster completion times and improved 'big bubble' formation, crucial for successful DALK.
• AUTO-DALK's vertical insertion strategy proved more effective than conventional horizontal approaches in preventing DM perforation.

Conclusions:

• AUTO-DALK represents a feasible and effective OCT-guided autonomous robotic solution for partial-thickness keratoplasty.
• The system has the potential to enhance safety, consistency, and efficiency in DALK procedures.
• Autonomous robotic execution guided by real-time OCT sensing shows promise for advancing ophthalmic microsurgery.