PURPOSE: To evaluate intraobserver and interobserver reproducibility of measurements in images obtained by the 50 MHz Humphrey Ultrasound Biomicroscope. METHODS: The first stored image of central cornea, central anterior chamber, and angle structures obtained at 50 examinations were measured twice by three observers. Nine different parameters were measured in angle images and two in images of central cornea and anterior chamber. Intraobserver reproducibility was assessed by calculating the coefficient of variation and interobserver reproducibility by a two-ways ANOVA. RESULTS: Intraobserver reproducibility was high for all measurements of central cornea thickness and anterior chamber depth with a coefficient of variation (%) < or = 3.8%. In angle images the intraobserver reproducibility of measurements was often lower (coefficient of variation (%) 1.3-52.6%). The lower reproducibility was most noticeable in measurements involving the less defined structures. Interobserver reproducibility was poor. CONCLUSIONS: Comparison of measurements are best done of well defined structures and by only one observer.
PURPOSE: To evaluate intraobserver and interobserver reproducibility of measurements in images obtained by the 50 MHz Humphrey Ultrasound Biomicroscope. METHODS: The first stored image of central cornea, central anterior chamber, and angle structures obtained at 50 examinations were measured twice by three observers. Nine different parameters were measured in angle images and two in images of central cornea and anterior chamber. Intraobserver reproducibility was assessed by calculating the coefficient of variation and interobserver reproducibility by a two-ways ANOVA. RESULTS: Intraobserver reproducibility was high for all measurements of central cornea thickness and anterior chamber depth with a coefficient of variation (%) < or = 3.8%. In angle images the intraobserver reproducibility of measurements was often lower (coefficient of variation (%) 1.3-52.6%). The lower reproducibility was most noticeable in measurements involving the less defined structures. Interobserver reproducibility was poor. CONCLUSIONS: Comparison of measurements are best done of well defined structures and by only one observer.
Authors: Haitao Li; Christopher Kai Shun Leung; Carol Yim Lui Cheung; Lee Wong; Chi Pui Pang; Robert Neal Weinreb; Dennis Shun Chiu Lam Journal: Br J Ophthalmol Date: 2007-05-02 Impact factor: 4.638