A Non-Invasive Method for Tracking Retinal Motion During Pars Plana Vitrectomy
thesisposted on 18.02.2018 by TanviShri Jatla
In order to distinguish essays and pre-prints from academic theses, we have a separate category. These are often much longer text based documents than a paper.
Pars Plana Vitrectomy is performed to access the posterior portion of the eye, specifically when retinal repair is required. The tear in the retina can be instigated by many causes: shrinkage of vitreous fluid, injury, diabetes, inflammatory eye diseases, and many more. Surgeons today have numerous types of cutters to select from when performing this minimally invasive surgery. Previous experimenters have performed their analysis based on physiological parameters in ex-vivo. For this reason, a method to quantitatively detect the movement of the retina in-vivo during Pars Plana Vitrectomy is introduced. After the surgery was performed in-vivo, the video of the surgery was used to determine the motion of the retinal movement. The movement of the structure in a two dimensional image time series was determined using ImageJ and Kymograph plugin. The net movement and velocity of the retina were obtained. Since acquiring a common anatomical reference point was impossible because of magnification of video, another reference point was established: edge of camera. Finally, a manual frame-by-frame analysis was performed to validate the approach of using ImageJ and Kymograph as a tracking method. After taking into consideration the constraints and assumptions established in the scope of this project, it is evident that the distance between the probe and the retinal tear made an impact on the movement of the retinal tear. The lowest velocity was believed to have had the probe farther away in distance from the retinal tear when compared to the video with the highest velocity. Based on this analysis, the use of ImageJ is an appropriate way to measure the motion of the retina during Pars Plana Vitrectomy. Using the camera edge as a reference, the average distance away from the mean length was significantly lesser than the motion of the retinal tear (p<0.0001). Furthermore, the results for the frame-by-frame analysis showed that the average of the actual value obtained from the Plot Profile on ImageJ was close to the average of the experimental value obtained from the Kymograph. The relative error between them of 0.369% articulates that the Kymograph provides results very close to the actual results.