Abstract
Purpose :
To generate microstructurally-based 3D finite element models (FEM) of PRK for myopia and compare the outcomes of the simulations to actual post-PRK follow-up tomographies.
Methods :
Patient-specific tomographic data from 21 eyes of 15 patients who underwent sphero-cylindrical PRK in a retrospective consecutive case series were imported into custom meshing software and case-specific treatment settings were simulated in each FEM. Average tangential curvatures over the central 3mm circular region (Kmean) were obtained from the anterior surfaces of both the actual tomographies and their generated FEMs at pre- and post-treatment intervals. Prediction error in Kmean was calculated by absolute and change-based metrics (based on the difference in actual and simulated postoperative values and the difference in actual and simulated PRK-induced change, respectively). Prediction error ≤ 0.50D was set as the validation criterion.
Results :
Prediction error in absolute postoperative Kmean and induced change in Kmean were -0.19 ± 0.39 D and -0.13 ± 0.37 D, respectively. Prediction error was not statistically significant (p>0.1). 81% of simulated cases met the validation criterion. Simulated PRK-induced change in Kmean accounted for 92% of the variance in actual spherical equivalent refractive change (Figure1).
Conclusions :
The investigated computational modeling approach demonstrated low prediction errors and may be useful for clinical guidance in planning PRK. Mean prediction errors were within the test-retest repeatability of current clinical tomography systems.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.