Comparison of A-scan device accuracy

Abstract

A-scan biometry is recognized as a useful aid in predicting intraocular lens power. Measurements are reported to be accurate to better than .+-. 0.1 mm. Three biometry devices were compared in examinations of 159 persons, each examination being repeated several times. Axial length averaged 23.77 mm when mesured by the immersion technique; applanation and modified applanation techniques yielded 0.1 mm and 0.3 mm shorter distances, respectively. Measured values of axial lengths did not have the same probability of being measured, even if they were close together. Results were not distributed in smooth Gaussian curves; on the contrary, clusters of values, on the pattern of our choice of ultrasound wavelength, were seen even when biometry was performed with electronic gates in the devices. Retest reliability decreased when short distances in the anterior segment of the eye were measured. Measurements of lens thickness were less readily reproducible in cataractous lens than in healthy young eyes; anterior chamber depth, on the other hand, was measured more reliably in cataract patients. This last finding may have resulted in part from uncertainties about ultrasound velocity in the cataractous lens and in part from accommodatation. In cataract patients, axial length was measured most reproducibly by the immersion technique; it was measured less accurately in young healthy eyes with a modified applanation device.