VIRTUAL REALITY SIMULATOR FOR VITREORETINAL SURGERY USING INTEGRATED OCT DATA
In an attempt to follow the fast-evolving advances in retinal surgery, we feel there is an unmet need for advanced surgical simulator training in vitreoretinal interface diseases. The integration of retinal images from the 3D optical coherence tomography (OCT) volume scans, which could serve as a disease template, is a further step to bring surgical simulation closer to reality. This report for the first time describes the integration of OCT images into a virtual reality surgical simulator.
For the purpose of the study we used random, previously acquired patients’ images from the database of the King Khaled Eye Specialist Hospital in Riyadh, Saudi Arabia. Eyes with epiretinal membrane (n=3) and macular hole (n=2) were imaged using the RTVue instrument (Optovue, Freemont, CA, USA). The image raw data were downloaded, exported with no patient identifier, and used for 3D rendering in MicroVisTouch. The study adhered to the tenets of the Declaration of Helsinki in clinical research.
We then designed a computer-based simulator for epiretinal membrane (ERM) and internal limiting membrane (ILM) peeling procedures with the idea to develop a high performance haptics-, physics-, and graphics-enabled simulator. The study was conducted on the MicroVisTouch surgical simulator. High fidelity haptic feedback is rendered to provide the sense of real surgery sensations. Opening and closing of the forceps is governed by a pressure sensor attached to the Geomagic Touch haptic device. The deformable retina is modeled from volumetric data generated from the OCT scan (Optovue) using position-based dynamics with improved performance over traditional mass-spring models.
A customized peeling algorithm with full controllability was developed to simulate the peeling of the ERM and ILM. Figure 2 shows an example of a successful integration of an OCT image from a real patient with vitreoretinal interface abnormality which will serve as a template for ERM removal practice. The surface membranes can be distinguished from retinal tissue on the integrated OCT scan. A combination of vertex and fragment shaders is adopted to provide improved realism on graphical/visual effect. The forceps are used for surgical simulation and the user is instructed to apply enough pressure to the sensor in order for the membrane not to slip out of the forceps.
Several metrics can be used to assess surgical performance. The circularity score measures how perfect the circle of removed ERM or ILM is with respect to an ideal peel. The accuracy score indicates the number of times the retina is touched by the forceps. The fluency score indicates the number of times the forceps are closed to grab the membrane. These all can be used as part of surgical skills evaluation both at initial performance and subsequent sessions.
In summary, the concept introduced here can be utilized to practice image-based vitreoretinal surgery using OCT scans from real patients. Contrary to other surgical simulators which use a standardized pre-operative starting condition, the integration of retinal imaging into a surgical simulator will allow students to be challenged with numerous different conditions. We believe this innovation will contribute to enhancing preclinical surgical experience in vitreoretinal surgery.