Study Shows Best Depth for Viewing Retinal Nerve Fiber Bundles

Investigators explore the potential of OCT imaging in monitoring glaucoma.

The best depth to view retinal nerve fiber bundles (RNFB) is about 20 µm below the inner limiting membrane (ILM) in most of the retina, and depth range is highly correlated with retinal nerve fiber layer thickness (RNFL), researchers found in a study published in Ophthalmic & Physiological Optics. They sought to learn the configuration of RNFBs in healthy eyes to obtain reliable imaging and better identify defects through en face optical coherence tomography (OCT).

The researchers imaged 1 healthy eye from each of 10 adults (5 females) who were aged 57 to 75 years. Each subject had visual acuity of a maximum of 0.20 logMAR measured at 6 m with appropriate refractive correction, clear optical media and normal visual field, and a maximum of 21 mm Hg intraocular pressure (IOP) in 1 eye and maximum difference of 4 mm Hg between eyes.

The researchers created a pseudo-montage of 7 B-scan images of the central +/- 25° with OCT at each depth for each eye and extracted 50 single pixel (3.87 µm) thick slab images with depth-resolved attenuation coefficients from 0 to 193.5 µm below the ILM and took traditional OCT imaging of the optic nerve head (ONH), including a circle B-scan that provided the mean circumpapillary RNFL. As a background, the researchers used a rectangular image of the raphe area of the composite image 35 µm below the ILM of each participant and a value 30% below the median pixel value within this rectangle. The median of the 99th percentile of all 50 slabs normalized each slab by division.

Graders rated the subjective visibility of RNFBs of 10 retinal regions to locate both the range below the ILM at which RNFBs are visible in each region and the depth of greatest visibility. To limit any bias, graders ignored overlapping areas from different scans. Best visibility analysis was limited to cases where the proportion of equal visibility of RNFBs across multiple depths was below 30%.

The researchers found that the retinal region impacted the optimum depth for best RNFB (χ2 (9) = 58.8, P <.0001). The mean for best RNFBs visibility was 20.3 +/- 1.9 µm (range 17.4 +/-1.3 µm for raphe region to 22.8 +/- 2.0 µm for the nasal macula). The researchers excluded the ONH region from the analysis since graders could not identify a single depth with best visibility of bundles 40% of the time, compared to not being able to find it less than 30% of the time for the rest of the retinal regions.

There was good agreement (intraclass correlation coefficients (ICC)(2,1) = 0.89, 95% CI: .87–.91) and repeatability (grader 1, ICC(3,1) = 0.96, 95% CI: 0.95–0.97; grader 2, ICC(3,1) = 0.99, 95% CI: 0.98–0.99) overall between graders’ ratings of RNFB visibility, but certain regions, like the ONH, had more variability (ONH-raphe: 35.76 µm, P <.0001).

Thickness of visible RNFBs (axial distance between anterior and posterior limits of visibility) was correlated with RNFL thickness (Spearman’s ρ =.9, 95% CI: .85–.93, P <.0001) and inversely correlated with consistency in ratings between clinicians (Spearman’s ρ: .66, 95% CI: .58–.72, P <.0001).

Limitations of the study include nearly exclusive inclusion of solely one ethic group (White) for the sample eyes. The imaging process also possibly incorrectly assumed constant background throughout the whole retina.

Reference

Cheloni R and Denniss J. Depth-resolved variations in visibility of retinal nerve fibre bundles across the retina in enface OCT images of healthy eyes. Ophthalmic Physiol Opt. Published online November 4, 2020. doi: 10.1111/opo.12756