Optical Coherence Tomography Guided Robotic Needle Insertion for Deep Anterior Lamellar Keratoplasty

Optical Coherence Tomography Guided Robotic Needle Insertion for Deep Anterior Lamellar Keratoplasty 170 177 IEEE Transactions on Biomedical Engineering (TBME)

Deep anterior lamellar keratoplasty (DALK) is a promising partial-thickness cornea transplantation technique that spares eligible patients the immunosuppression and eye rupture risks associated with full-thickness transplantation, the current standard of care.  The popular “big bubble” DALK technique involves advancing a needle from the cornea’s paracentral zone to its apex, arriving at approximately 90% depth of total cornea thickness to facilitate the subsequent tissue dissection.  Unfortunately, DALK suffers from a high failure rate due to shallow needle insertions, which produce incomplete tissue dissections, or due to endothelial perforations, which require conversion to the full-thickness transplantation.  We jointly addressed the visualization and manipulation challenges that reduce DALK success rates by demonstrating an optical coherence tomography (OCT)-guided, robotically-stabilized workstation for performing needle insertions.  The OCT subsystem provides live volumetric views of the cornea, a synthetic B-scan resampled along the detected needle shaft, and segmentation of the cornea’s epithelial and endothelial surfaces for real-time needle depth measurements.  The robot subsystem jointly shares control of the needle through an ergonomic handpiece and either attenuates the surgeon’s hand tremor during cooperative insertions or, using the needle tracking and segmentation results, completes the insertion automatically.  We conducted a study of needle insertions in ex vivo human corneas in which we compared the performance of cornea fellows with/without OCT visualization and with/without robotic stabilization to the performance of our fully automatic algorithm.  We found that OCT visualization versus the standard stereo microscope was a key enabler for cornea fellows in terms of perforation-free needle depth.  Notably, our automatic robotic needle insertion method met or exceeded the performance of the cornea fellows with regards to needle depth and perforation rate.  We believe our workstation concept can improve the success of DALK needle insertions and thereby improve outcomes for patients.