Abstract
Abstract: :
Optic disc assessment is important for detecting and monitoring glaucoma. The scanning laser ophthalmoscope provides an objective and repeatable way of assessing the optic disc parameters. However placement of the contour line on the inner scleral rim by the examiner can be difficult and inaccuracies may lead to misleading interpretation of the data. Purpose: To assess whether the use of colour photos of the optic disc can aid the definition of the disc contour when assessing patients with the Heidelberg Retinal Tomograph (HRT). Methods: Study design: Intra and inter observer study, and inter method study (with and without the aid of a photographic image of the optic disc). Images were obtained from 39 eyes with the HRT II and the Haag Streit Eye Cap system with mydriatic fundus camera by one observer (GW). Four observers (one glaucoma specialist, two experienced ophthalmology trainees, and one ophthalmic photographer) independently drew the disc margins on the HRT topographic images. The disc margins were drawn on a separate occasion using a printed colour disc photo to guide placement of the disc contour. For three observers both these assessments were repeated on a different occasion. The disc area was compared for the two methods as a good indicator of differences in the disc contour. Agreement between the two methods was assessed by Bland–Altman plots, testing for bias and 95% limit of agreement, using the F test of variance to test for significant differences in the latter. Results: The inter method variability for the disc area showed a mean difference of –0.1511 mm2 for HRT –HRT with photos, with a 95% limit of agreement of –0.4117 to 0.1096 (p<0.0001), indicating that overall the disc area was measured as larger when colour photos were used. This difference was highly significant for three observers, and non–significant for the fourth (p=0.0807). Intraobserver variability was not improved by the use of photos. Interobserver variability was calculated using a Bland–Altman plot for each observer calculated by subtracting the individual average from the average of all the observers. There was increased variability at disc sizes over 3.5mm2 with the HRT method whereas with the photo assisted method there is a more even spread across the range of disc sizes. Conclusions: Using colour photos to assist in drawing the disc contour for HRT analysis resulted in an overall larger contour being defined than when relying on the HRT reflectance image but does not appear to improve reliability.
Keywords: imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • optic disc