Purchase this article with an account.
I.M. Tavares, C. Bowd, F.A. Medeiros, L.M. Zangwill, E.M. Hoffmann, P.A. Sample, R.N. Weinreb; Quality Assessment of HRT II and GDx VCC Images in the Diagnostic Innovations in Glaucoma Study (DIGS) . Invest. Ophthalmol. Vis. Sci. 2006;47(13):3646.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To assess HRT II (software version 2.01) and GDx VCC (software version 5.0.1) image quality and to report the rate and reasons for image rejection in the Diagnostic Innovations in Glaucoma Study (DIGS).
243 eyes from 243 patients with HRT II (Heidelberg Engineering, Dossenheim, Germany) and GDx VCC (Carl Zeiss Meditec, Dublin, CA) images taken within 3 months were selected from the DIGS. For both instruments, a mean image (composed of 3 individual images) was reviewed by each of two experienced graders masked to the assessment from the other grader and any other clinical information. Images were judged acceptable or unacceptable by subjective and objective assessment of quality. Subjective assessment discrepancies between the two graders were adjudicated by a third experienced grader. Subjective criteria were centering, clarity and illumination of images and presence of vessel doubling and floaters within the calculation area for both devices. Moreover, evidence of eye movement during HRT image acquisition was evaluated by viewing "movies" of the imaging sequence. Objective criteria for image rejection were standard deviation > 50 µm and sensitivity > 90 for HRT II, and residual retardance > 12 nm, typical scan score < 40, and quality score < 7 unless low score was due to fixation error (centering), for GDx VCC. All objective criteria were based on manufacturers suggestion.
For HRT, 36/243 (15%) images were rejected, and the main reasons for image rejection were sensitivity > 90 (20/36), standard deviation > 50 µm (15/36) and floaters (6/36). For GDx VCC, 54/243 (22%) images were rejected, and the main reasons for image rejection were typical scan score < 40 (19/54), quality score < 7 (11/54), vessel doubling (20/54) and illumination (13/54). Many images had more than one reason for rejection. A significantly larger number of GDx VCC than HRT images were rejected (Fisher’s Exact test, P = 0.047).
Significantly more GDx VCC images than HRT II images were rejected due to poor quality. A large proportion of GDx VCC images were rejected based on the objectively determined presence of atypical scans.
This PDF is available to Subscribers Only