Purchase this article with an account.
Abhijit Sinha Roy, Barbara Fant, Karoline Rocha, William Dupps, Jr.; Estimation Of Modulus Change After Corneal Crosslinking (cxl) Using Multiple Post-cxl Topographies And Inverse Finite Element. Invest. Ophthalmol. Vis. Sci. 2012;53(14):6896.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate modulus change after corneal crosslinking (CXL) and its correlation with absolute and change in diagnostic curvature indices
We have recently developed a 3-D patient-specific finite element (FE) of corneal CXL of keratoconus eyes. In this study, the model was customized into an inverse optimization approach where the change in stromal elastic modulus was estimated from change in corneal curvature. Data from CXL of 12 eyes was used. 3-D imaging was performed before and after CXL using Pentacam (Oculus Inc., Germany). Follow up data ranging from 3 months to over a year was used. Multiple exams were taken at each post-CXL time point to minimize the variability in corneal topography in KC eyes. The inverse technique used an optimization error function which included all the post-CXL exams for each eye. The change in modulus was correlated to standard optical indices such as mean axial and tangential curvature in the central 1.5 dia. zone. Also changes in cone-localized indices (mean axial and tangential curvature in a 1.5 mm dia. zone around the steepest point) were evaluated.
The maximum tangential curvature post-CXL was 57.50+4.97D and 57.13+5.03D, in vivo and predicted by inverse FE, respectively. At 3 and 6 months, modulus increased (Stiffening factor -1 in figure) by 77.5% and 96.4%, respectively. Linear regression among all optical variables including aberrations yielded high correlations with lowest R2 equal to 0.88 for SimK-flat. Cone-localized specific variables were also evaluated and yielded high correlation (R2 > 0.94).
Inverse finite model was able to predict increases in modulus after CXL. The variance of estimated stiffness changes was lower than a prior study that did not use replicate measurements (ASIA ARVO 2011, #49). However, replicate data sets improved the coprrelation between in vivo measurments and inverse model results, particulary lower order aberrations.
This PDF is available to Subscribers Only