Purchase this article with an account.
Casie Goldman, Shaban Demirel, Cindy Blachly, Michael Whitworth, Steven Mansberger, Stuart Gardiner, Deborah Goren; Comparison of the strengths of ‘moving correlations’ between SD-OCT and SLP nerve fiber layer thickness and standard automated perimetry. Invest. Ophthalmol. Vis. Sci. 2013;54(15):2279.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To compare the strength of structure-function correlations between standard automated perimetry (SAP) and retinal nerve fibre layer thicknesses (RNFLT), as measured by either spectral-domain ocular coherence tomography (CorrSD-OCT/SAP) or scanning laser polarimetry (CorrSLP/SAP), at different stages of glaucoma.
SAP visual fields (SITA Std, 24-2) and RNFLT measures were collected for 400 eyes of 209 participants with early to moderate glaucoma (Mean deviation (MD) mean: -0.31±2.87; range: -16.20 to 3.27). To avoid predetermined definitions of disease stage, a rolling binning process was employed that sorted by 6 metrics to stage the disease. These included four functional metrics (MD, Mean sensitivity (MS), number of locations outside normal limits (p<5%) on the total deviation (TD) probability map, and number of locations on the pattern deviation (PD) probability map) and two structural metrics (mean RNFLT from a circle scan as measured by SD-OCT, mean RNFLT from SLP). For each staging, the first bin was defined as the first 80 observations within the sorted database. Subsequent bins were created for observations 21-100, 41-120, 61-140 etc. This process resulted in 17 overlapping bins specific to each staging. For each bin, Pearson correlation coefficients were calculated between linearized mean sensitivity (MSLin=10MS*0.1) and RNFLT and compared using Steiger’s test. This process was repeated for the 6 sorting metrics, resulting in 6 distinct stagings.
For all stagings and all bins, CorrSD-OCT/SAP was either equivalent to, or higher than CorrSLP/SAP. When using MD for staging, CorrSD-OCT/SAP values were significantly higher than CorrSLP/SAP for bins with MD ranges -0.88 to 0.2 (p<0.001) and -0.45 to 0.35 (p=0.008). When using TD and PD for staging, CorrSD-OCT/SAP values were significantly higher than CorrSLP/SAP for bins representing eyes with a few defective locations (1-4 abnormal TD locations: p=0.033. 1-3 abnormal PD locations: p=0.029).
Function (MSLin) correlates significantly more strongly with structure (RNFLT) as assessed by SD-OCT than when assessed by SLP but only during early stages of glaucomatous damage. Correlations between structure and function are similar for SD-OCT and SLP later in the disease process.
This PDF is available to Subscribers Only