Abstract
Purpose: :
Corneal pachymetry mapping often plays an important role in the evaluation and screening of patients for refractive surgery. Visante anterior segment-optical coherence tomography (AS-OCT) has been shown to be a reliable method for measuring central corneal thickness (CCT). However, software surface detection algorithms may at times inaccurately define corneal boundaries. We sought to determine the magnitude and effect of surface tracing errors created by computer algorithm-identified boundaries on pachymetry mappings in patients undergoing evaluation for refractive surgery.
Methods: :
Retrospective review of a consecutive case series. 25 patients undergoing evaluation for refractive surgery by one surgeon in our institution from January 2008 to September 2008 had Visante AS-OCT as part of the screening process. Scans were performed on a total of 49 eyes (24 right, 25 left). AS-OCT images were examined to correlate automatic with actual corneal boundaries. Automatic surface boundaries not aligned with the contour of the cornea were manually edited. Surface tracings in 6 pachymetry maps in the right eye, and 5 in the left eye, did not warrant editing.
Results: :
Corneal thickness was assessed for central (0-2 mm), pericentral (2-5 mm), transitional (5-7 mm) and peripheral (7-10 mm) zones up to a 10 mm diameter. The average central corneal thickness was 524 microns (range, 458-624). Manual editing of the surface boundaries in the central zone was associated with < 1% change in average CCT. In the pericentral area, editing changed the minimum corneal thickness in one eye from 605 to 674 microns (13.0%). The average corneal thickness in the pericentral zone was overall increased, however the maximum increase was < 3%. In the transitional zone, the mean change in minimum, average, and maximum corneal thickness was 2.7%, 1.5%, 5.0%, respectively. The largest change in minimum, average, and maximum corneal thickness for this zone was 13.5%, 5.7%, and 15.0%, respectively. In the peripheral zone, the mean change in minimum, average, and maximum corneal thickness was 3.4%, 2.3%, 1.6%, respectively. The largest change in minimum, average, and maximum corneal thickness for this zone was 14.5%, 5.4%, and 8.5%, respectively. The largest decrease (-5.5%) in corneal thickness after editing adjustments was seen in the peripheral zone.
Conclusions: :
Manual adjustments for corneal surfacing tracing errors made by software surface detection algorithms correct errors that tend to underestimate corneal thickness, especially towards the periphery. Although generally small, these errors may near 15%.
Keywords: imaging/image analysis: clinical • refractive surgery