Posters

Comparison of astigmatic prediction accuracy in pseudophakic eyes for different diagnostic devices

Poster Details

First Author: P.Hoffmann GERMANY

Co Author(s):    M. Abraham   N. Hirnschall   O. Findl        

Abstract Details

Purpose:

To compare corneal measurements obtained by a swept source fourier domain OCT (CASIA SS-1000), an autokeratometer (Haag-Streit Lenstar), a hybrid topographer (Tomey TMS-5), a Placido topographer (Tomey TMS-5 in Placido mode) and a Scheimpflug tomographer (Oculus Pentacam) to manifest subjective refraction.

Setting:

Private eye clinic in Germany

Methods:

104 pseudophacic patients with non-toric IOLs were measured at least 6 months after surgery. Corneal astigmatism as measured on the anterior corneal surface as well as total corneal astigmatism including posterior surface data was compared to manifest refractive cylinder (cross-cylinder strategy) by computing difference vectors and correlation analysis of power vectors.

Results:

The OCT (0.43 ± 0.25 D) and the hybrid topographer (0.44 ± 0.25 D) yielded the smallest difference vector to the subjective cylinder and by far the lowest percentage of outliers > 0.75 D (≈ 10%). The rotating Scheimpflug camera showed the largest (0.70 ± 0.41 D) difference vector. The best predictive precision (0.37 ± 0.22) could be achieved by vector averaging Lenstar keratometry and OCT.

Conclusions:

Autokeratometry yielded the least measuring noise but OCT as well as hybrid topography had better predictive precision due to posterior curvature data. Scheimpflug tomography suffered from high measuring noise. Combination of keratometry and OCT data yielded the best precision for planning of toric IOL implantation. To get a reliable target cylinder for TIOL calculation, accuracy of the measuring device is crucial. Keratometry and Placido topography lack the information of the posterior corneal curvature while Scheimpflug devices suffer from higher measuring noise. In this paper, a combination of OCT with autokeratometry yielded the best predictive quality. FINANCIAL INTEREST: NONE