Reliability of ultrasound biomicroscopic measurements of the anterior segment

Abstract

We determined the precision of the measurement obtained by ultrasound biomicroscopy (UBM) of the anterior segment of the eye. In 10 subjects (10 eyes) we evaluated the reproducibility of two measurements on the same UBM image and on two images obtained at different times, calculating the coefficient of correlation (r). In 10 other subjects (10 eyes) we assessed the variability of 5 measurements taken during 5 successive examinations, calculating the mean coefficient of variation (CV) for each parameter.

On the same UBM image, the within-observer agreement was high for all parameters (r values ranging between 0.95 and 0.99), whereas the between-observer agreement was less satisfactory for the angle-opening distances and for iris thickness 1 (r ranging from 0.74 to 0.79). On images of the same ocular section obtained at two different times the within- and between-observer agreement was unacceptable for 5 parameters (angle-opening distance 250 and 500 μm, iris thicknesses 1 and 2, iris-lens angle) and for 6 parameters (same measurements plus scleral-ciliary process angle), respectively, with r values ranging between 0.21 and 0.76.

Analysis of the CV of 5 measurements showed good reproducibility (CV < 10%) for the determination of anterior chamber depth, scleral thickness, trabecular-ciliary process distance, iris thicknesses 2 and 3, iris-zonule distance, scleral-iris angle and scleral-ciliary process angle. Less satisfactory CV values, ranging from 10.5 to 19.8%, were obtained for trabecular-iris angle in degrees, angle-opening distance 250 and 500 μm, iris-ciliary process distance, iris thickness 1, iris-lens contact distance and iris-lens angle.

In conclusion, the precision of UBM is insufficient for many parameters. The reproducibility of the technique may be increased by more strictly standardized examination conditions and by better training of the observers.