Cross-sectional computed tomography assessment of exophthalmos in comparison to clinical measurement via Hertel exophthalmometry.

To determine protrusion assessment via Hertel exophthalmometry in comparison to measurement on Computed Tomography (CT). Retrospective blinded comparison of exophthalmos measurements on axial CT with Hertel exophthalmometry measurements in 113 patients. Descriptive statistics, Pearson's correlation, Kruskal-Wallis and Mann-Whitney-U test were employed for analysis. Mean difference of proptosis between both eyes was 2.4 (SD ± 2.0) mm in CT and 2.2 (SD ± 2.0) mm in Hertel measurements. Proptosis of 0-2 mm was present in 69 (61.1%), and > 2 mm in 42 (38.9%) patients in Hertel measurements (CT 64 (56.6%), and 49 (43.4%) patients). Pearson's coefficient showed a correlation of 0.793 between both methods (p < 0.001). Accuracy of Hertel measurement depended significantly from the examiners' experience (< 5 (group 1), 5-15 (2) and > 25 (3) years, p = 0.042, Kruskal-Wallis analysis; p = 0.086 group 1 vs. 2, p = 0.014 group 1 vs. 3, p = 0.688 group 2 vs. 3, Mann-Whitney-U-test), reflected by levels of Pearson's coefficient (correlation of both methods 0.691 (group 1), 0.837 (2) and 0.831 (3), respectively, p = 0.01). Generally, Hertel exophthalmometry correlates well with CT measurements. Subgroup analysis confirmed a superior quality of Hertel measurements in favour of experienced examiners. Teaching of accurate Hertel exophthalmometry should be improved. Assessment of exophthalmos using standardized criteria should be implemented for imaging reports.