Comparison of Cone Mosaic Metrics From Images Acquired With the SPECTRALIS High Magnification Module and Adaptive Optics Scanning Light Ophthalmoscopy

Abstract

Purpose To compare cone mosaic metrics derived from adaptive optics scanning light ophthalmoscopy (AOSLO) images with those derived from Heidelberg Engineering SPECTRALIS High Magnification Module (HMM) images. Methods Participants with contiguous cone mosaics had HMM imaging performed at locations superior and temporal to the fovea. These images were registered and averaged offline and then aligned to split-detection AOSLO images; 200 × 200-µm regions of interest were extracted from both modalities. Cones were semi-automatically identified by two graders to provide estimates of cone density and spacing. Results Thirty participants with contiguous cone mosaics were imaged (10 males, 20 females; age range, 11–67 years). Image quality varied, and 80% of our participants had analyzable HMM images. The intergrader intraclass correlation coefficients for cone metrics were good for both modalities (0.688–0.757 for HMM; 0.805–0.836 for AOSLO). Cone density estimates from HMM images were lower by 2661 cones/mm 2 (24.1%) on average compared to AOSLO-derived estimates. Accordingly, HMM estimates of cone spacing were increased on average compared to AOSLO. Conclusions The cone mosaic can be visualized in vivo using the SPECTRALIS HMM, although image quality is variable and imaging is not successful in every individual. Metrics extracted from HMM images can differ from those from AOSLO, although excellent agreement is possible in individuals with excellent optical quality and precise co-registration between modalities. Translational Relevance Emerging non-adaptive optics-based photoreceptor imaging is more clinically accessible than adaptive optics techniques and has potential to expand high-resolution imaging in a clinical environment.

Bull’s Eye Maculopathy in Near-Infrared Reflectance as An Early Sign of Hydroxychloroquine Toxicity

“…In our series, we cannot comment on whether an NIR bull's eye lesion arises from changes within or outside the interdigitation zone, as, in all cases, the SD-OCT was unremarkable [10,12]. However, in these cases, further testing may be of interest, such as using adaptive optics or the new cone mosaic metrics derived from Heidelberg Engineering Spectralis High Magnification Module images [12,13,16].…”

Progression of Rare Inherited Retinal Dystrophies May Be Monitored by Adaptive Optics Imaging

“…Figure 1 depicts the photoreceptor mosaic in a healthy eye. The image is taken paracentrally (2 superiorly) due to the limited ability of foveal image acquisition by Rtx™ [ 17 ]. By changing the focus point, the quality of an acquired image provides the assessment of photoreceptor parameters in parafoveal cones.…”

“…The image dimensions are 4 × 4 , which corresponds to approximately 1.2 mm × 1.2 mm of the retina. The observation of foveal cones in Rtx™ is limited, as mentioned above [ 17 ], so images of the extrafoveal retina are typically acquired. Researchers working with this device often bypass this limitation by choosing a parafoveal region for the analysis.…”

“…AO enables the visualisation of rods and cones, the two types of retinal photoreceptors. The parameters that adaptive optics can measure include: cone density, cone spacing, Voronoi analysis of hexagonal cells, reflectivity, regularity, metrics for the preferred orientation of cones, and local spatial anisotropy [ 6 , 16 , 17 , 26 ].…”

“…DM is expressed in 1/mm ; it is inversely correlated with SM, which measures the neighbour distance of each cone. REG (expressed in %) is important for providing the exclusion of inaccuracies caused by cell identification errors [ 17 ]. Our study focused on the use of AO in IRDs, specifically CD, CRD, and STGD, over a 6-year observation period.…”