Conoidal fitting of corneal topography height data after excimer laser penetrating keratoplasty.

Abstract

To demonstrate a mathematical method for approximation of discrete corneal topography height data with a biconic model surface for better appreciation of the optical performance of the graft after penetrating keratoplasty. In this retrospective study we included 50 eyes of 50 patients (30 keratoconus; 20 Fuchs' dystrophy) undergoing nonmechanical excimer laser penetrating keratoplasty. Conventional keratometry, corneal topography (TMS-1), subjective refraction, and best spectacle-corrected visual acuity (BSCVA) were assessed preoperatively, 3 and 6 months postoperatively, and before/after suture removal. A biconic model surface was fitted to the topographic raw data (8.0-mm region of interest) minimizing the root mean square error and a set of parameters (meridional power, axis, conic constant, and approximation error) was determined. The refractive cylinder was correlated with keratometric power readings, the Simulated Keratometry (SimK) of the topography system, and the respective parameters of the model surface. Keratometric/SimK astigmatism increased from preoperatively (3.40/4.30 D) to 3 months (4.30/4.80 D) and decreased to 3.40/3.90 D after suture removal. Refractive cylinder/cylinder of the biconic increased slightly from 2.10/2.60 D preoperatively to 3.20/3.30 D after suture removal. The topographic cylinder SimK yielded the highest and the refractive cylinder the lowest values at each follow-up examination. Central keratometric power readings were stable before suture removal and decreased (about 1 D) due to suture removal. The conic constants in both meridional cross-sections changed from a prolate to a spherical shape in the early time course after penetrating keratoplasty and reconverted to a prolate shape after suture removal. Regarding cylinder axis, there was a significant correlation of the model surface with the refractive cylinder at all examination (P < .05) but only a mild correlation of the keratometric and SimK cylinder axis to the refractive cylinder axis at some examination stages. Approximation of corneal topography height data with a biconic model surface renders reconstruction of clinically relevant corneal topography parameters including corneal asphericity with marked data compression. The correlation of amount/axis of refractive cylinder was best represented with the model surface parameters.