Abstract
Abstract: :
Purpose: Arc–scan very high frequency ultrasound (VHFU) allows visualization of the entire cornea in a single image. Because of the relatively small difference in speed of sound between the normal saline coupling medium and the cornea, and because the arc–scan maintains approximate normality of the beam axis relative to the corneal surface, refraction is negligible, facilitating measurement of posterior curvature. Our aim was to use the Artemis 2 high–frequency ultrasound system to acquire scan data on test objects of known radius and the corneas of human subjects, with comparison of ultrasound determined anterior radii with manual and automated keratometry. Methods: We developed software for measurement of anterior and posterior corneal radii of curvature for successive zones from 3 to 10 mm in diameter. After calibration with an 8 mm radius glass sphere, radii of 10 human subjects (20 eyes) were determined. In each case, five scans were acquired and analyzed in both vertical and horizontal planes, allowing determination of repeatability. Ultrasound determined anterior curvature values were then compared with manual and automated optical keratometry. Results: Standard deviations of successive ultrasound curvature measurements were 0.20 mm and 0.16 mm for anterior and posterior corneal surfaces respectively. Root mean square errors for mean ultrasound determined anterior radius of curvature versus automated and manual optical keratometry were 0.12 mm and 0.11 mm, respectively. The root mean square difference between automated and manual keratometry was 0.075 mm. Conclusions: Ultrasound determination of corneal radius of curvature is based on tracking the corneal surface in a scan plane and fitting a circle to the contour. Although it takes only a fraction of a second to scan a single plane, microsaccades may introduce uncertainties in curvature measurements. Our findings show greater variation in curvature measurements between ultrasound and optical keratometry than between automated and manual optical methods. However, because arc–scan ultrasound is largely unaffected by refraction, posterior corneal curvature values are readily determined with the same accuracy as anterior curvature. The ability to measure corneal thickness and the curvature of both corneal surfaces can allow improved assessment of corneal power. Because ultrasound also allows visualization and measurement of Bowman’s layer, it is also conceivable that an ultrasound–based ray tracing model can be developed based on the thickness, curvature and individual refractive indices of the epithelium and stroma.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • refraction