Abstract
Purpose :
Various studies have reported that the Hood-Kardon linear model (HKLM) [1] describes the structure-function relationship between visual field total deviation (VF-TD) loss and both retinal nerve fiber layer and local retinal ganglion cell plus inner plexiform layer (RGC+) thickness. The purpose here was to understand systematic deviations from the model reported for macular RGC+ thickness vs VF, when 24-2 loss was mild. [2,3]
Methods :
Swept-source optical coherence tomography (ssOCT), 9x12mm cube scans (DRI-OCT, Topcon Inc) and 24-2 and 10-2 VFs were obtained for 50 patients (54.1±14.2 yrs [mean±SD]) and 37 healthy controls (age = 53.8±8.4 yrs [mean±SD]). All patients had suspicious or glaucomatous discs with a 24-2 MD better than -6 dB. RGC and IPL layers were segmented using an automated algorithm and manual correction.[2] Layer thicknesses were determined at points corresponding anatomically to 10-2 points and analyzed within rings of eccentricity.[2] The HKLM was fitted to the points at each ring.[2]
Results :
Table 1 contains the results by eccentricity. A majority of points fell above the HKLM of RGC+ vs TD for both groups, especially near fixation (Fig. 1).[3] The HKLM predicts this deviation for controls, as well as suspects without “real” VF loss (Fig. 1A). In fact, for the patients, many points above the model had TD values within the normal range (pink rectangle); others fell within the normal range when PD was used. There were also a number of points below the HKLM prediction.[2] In some of these eyes, the test spot fell near the edge of a local defect, which may have underestimated VF loss due to a variation in individual RGC distribution and/or the test light overlapping healthy and affected regions. Using RGC thickness instead of RGC+ did not reduce systematic deviation, but significantly reduced overall variation along the y-axis (Table 1B).
Conclusions :
For macular RGC measures, deviations from the HKLM occur and will depend upon the sample of patients. In fact, points should fall above the predicted curve if eyes (“normal”/suspects) without glaucomatous damage are included. On the other hand, eyes with local damage near fixation may fall below the curve if the test spot is not perfectly aligned with the region of local damage. 1. Hood, Kardon, PRER, 2007 2. Raza et al, AO, 2011 3. Rao et al, IOVS, 2015
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.