Purchase this article with an account.
M. Balasubramanian, C. Bowd, D. J. Kriegman, R. N. Weinreb, P. A. Sample, L. M. Zangwill; Detecting Glaucomatous Changes Using Only One Optic Nerve Head Scan Per Exam. Invest. Ophthalmol. Vis. Sci. 2010;51(13):2723.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate a new retinal reflectance based method to detect glaucomatous changes using only 1 HRT optic disk scan per follow-up exam.
Each HRT scan L is modeled as L = I R cos(k) using a surface reflectance model; I is the incident illumination component; R is a reflectance map that describes surface reflectivity of the retina; and k is the incident angle between the laser light beam and retinal surface orientation. The reflectance R is less dependent on the incident angle k and likely has lower intra/inter-exam variability than the HRT image L. So, we used reflectance maps to detect glaucomatous changes. For each HRT exam, 1 of 3 scans was chosen and its reflectance map R computed using a Homomorphic filter (Oppenheim A, et al, IEEE Proc. 1968). At each superpixel (4×4 neighboring locations), change in reflectance from baseline to follow-up was estimated using a two-sample t-test. To identify locations with significant reduction in retinal height (red-pixels), p-value cutoffs were estimated at a false positive rate of 5% using the false discovery rate control method (FDR). For each follow-up, the ratio of total no. of red-pixels within the optic disk to disk size (red-pixels ratio) was calculated. Progression was defined as the presence of one or more follow-ups with red-pixels ratio > false positive rate. The results were compared to TCA with FDR (Balasubramanian M, et al, IOVS 2009; 50: ARVO E-Abstract 2573). 167 participants (246 eyes) with ≥4 HRT-II exams from the UCSD Diagnostic Innovations in Glaucoma Study (DIGS) were included; 36 eyes progressed by stereophotographs or showed likely progression on SAP Guided Progression Analysis (progressors). All other eyes were considered to be non-progressing. Specificities were estimated in 21 longitudinal normal eyes (normals).
The diagnostic accuracies of the reflectance based method / TCA with FDR were: 64% / 69% sensitivity in progressors; 74% / 53% specificity in non-progressors; and 100% / 90% specificity in normals, respectively.
The reflectance based method using 1 scan per exam provides high specificity and moderate sensitivity comparable to TCA with FDR that uses 3 scans per exam. Because reflectance maps are less dependent on variability in retinal surface orientation, these results suggest that it is possible to achieve comparable diagnostic power from just 1 scan per exam. The proposed method can be extended to detect glaucomatous changes from stereophotographs and spectral domain optical coherence tomography exams typically with 1 scan per exam.
This PDF is available to Subscribers Only