Background: Variability is introduced in three-dimensional (3-D) ultrasound assessments of intraocular tumours when multiple observers examine the same tumour or when multiple scans and measurements of the same tumour are carried out. We evaluated the magnitude and sources of variation associated with 3-D ultrasound measurements of choroidal tumour volume.
Methods: Three examiners performed 3-D ultrasonography of five choroidal tumours. Each examiner performed three scans of each tumour and then measured each scan three times to determine the volume. We determined the variation attributable to differences between examiners, scans and measurements by evaluating the contribution of each to the overall variation in tumour volume. A similar process was performed for tumour height for comparison purposes.
Results: The reliability of tumour volume measurements was found to be very high. The amount of overall variation, as defined by the coefficient of variation, was similar for tumour volume and tumour height and ranged from 4% to 17%. A total of 97% of the variation in tumour volume measurements and 84% of the variation in tumour height measurements was due to differences between tumours. Differences between examiners accounted for the most of the remaining variation (2% for volume and 7% for height).
Interpretation: Three-dimensional ultrasonography of intraocular tumours was highly reliable. Differences between examiners were the main source of variation.