Purchase this article with an account.
Ibrahim Seven, Ali Vahdati, William J Dupps; Patient-Specific Computational Analysis of Laser in situ Keratomileusis: A clinical validation study. Invest. Ophthalmol. Vis. Sci. 2016;57(12):2376.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To develop a microstructural based, 3D, patient specific, finite element model (FEM) of LASIK and compare the outcomes of the simulations to actual post-LASIK follow-up tomographies.
Patient-specific tomographic data from 20 eyes of 12 patients undergoing sphero-cylindrical LASIK treatment were meshed including epithelium, flap, wound, and residual stromal bed using a custom meshing software. Case-specific treatment settings were simulated. Simulated keratometry values (SimK) including the average curvature of the steep (K1), flat (K2) meridians, and central 3mm circular region (Kmean) were obtained from the anterior surfaces of both the actual tomographies and their generated FEMs. Mean difference (MD) and mean absolute difference (MAD) between these values were calculated to assess prediction error. Pearson correlation values (R) between individual errors in predicted Kmean and age, corneal hysteresis (CH), and corneal resistance factor (CRF) were calculated.
MD and MAD between simulated and actual post-LASIK cases were -0.13 ± 0.36 D and 0.28 ± 0.25 D respectively. The differences between simulated and actual SimK values were not statistically significant (p values for Kmean=0.1, K1=0.4, K2=0.5). R values between the individual differences in Kmean and CH, CRF, and age were 0.63, 0.53, and -0.5, respectively (p = 0.004, = 0.01, and = 0.02, respectively).
The investigated computational modeling approach demonstrated low prediction errors and may be useful for clinical guidance in planning LASIK. Clinical biomechanical metrics and surrogates explain some of the variance in prediction error and may be useful to tailor generic material properties into patient specific material properties to increase prediction accuracy.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only