Purchase this article with an account.
Ashley Sun, Emmanouil Tsamis, Xinhui Li, Katherine Tsang, Lama Al-Aswad, Dana Blumberg, George Cioffi, Jeffrey M Liebmann, Carlos Gustavo de Moraes, Donald C Hood; Detecting progression of early glaucoma using alternative methods with optical coherence tomography. Invest. Ophthalmol. Vis. Sci. 2019;60(9):5545.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To compare alternative methods for detecting progression of early glaucoma using swept source optical coherence tomography (ssOCT) widefield scans.
54 eyes from 54 individuals had (Topcon Atlantis) ssOCT scans obtained more than one year from their baseline visits (18.3±3.6 mos), as well as two scans within a baseline period (0.84± 0.61 mos). 38 eyes were diagnosed as glaucomatous or glaucoma suspects by a referring physician and 16 were healthy controls. For the metric method, average global (G), temporal (T), superior vulnerability zone (SVZ),1 and inferior vulnerability zone (IVZ)1 circumpapillary retinal nerve fiber layer (cpRNFL) thickness values were obtained. Quantile regression was utilized on the metrics obtained from the 2 baseline scans, to define the upper (95%) and lower (5%) limits of short-term variability. The metrics from the most recent scan were compared to those limits to define which eyes significantly worsened. For the region of interest (ROI) approach,2,3 abnormal regions of thinning on cpRNFL plots were manually marked for local defects on the last scan, aided by information from the B-scan and RNFL probability map (Fig. 1). This defined the ROIs for the baseline scan (Fig. 2) as well. As for the metrics, quantile regression was used on the average cpRNFL thickness (tROI) of the ROIs obtained from the 2 baseline scans, and eyes that significantly worsened from the most recent scan were identified.
For the ROI method, local defects were marked in 40 eyes (all 38 glaucomatous or suspects and 2 healthy controls). The tROI and the G thickness showed the most eyes that significantly worsened, 7 eyes each. None of these was a healthy control. Further, only 3 of the 7 eyes significantly worsened on both G and tROI. 3 of the 4 eyes identified by the tROI, but not G, had clear local damage in the cpRNFL, and all 4 eyes identified by G, but not tROI, had clear diffuse damage, as indicated by decreased cpRNFL thickness in at least 3 quadrants.
To detect progression of early glaucoma, both local (ROI), as well as global, cpRNFL thickness measures should be considered. 1. Hood DC. PRER, 2019; 2. Hood DC et al. JAMA Ophthalmol. 2015; 133:1438; 3. Wu Z et al. TVST. 2018;7:19.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only