Detection of photoreceptor disruption by adaptive optics fundus imaging and Fourier-domain optical coherence tomography in eyes with occult macular dystrophy

Adaptive optics imaging in inherited retinal diseases: A scoping review of the clinical literature

Adaptive optics (AO) imaging enables direct, objective assessments of retinal cells. Applications of AO show great promise in advancing our understanding of the etiology of inherited retinal disease (IRDs) and discovering new imaging biomarkers. This scoping review systematically identifies and summarizes clinical studies evaluating AO imaging in IRDs. Ovid MEDLINE and EMBASE were searched on February 6, 2023. Studies describing AO imaging in monogenic IRDs were included. Study screening and data extraction were performed by 2 reviewers independently. This review presents (1) a broad overview of the dominant areas of research; (2) a summary of IRD characteristics revealed by AO imaging; and (3) a discussion of methodological considerations relating to AO imaging in IRDs. From 140 studies with AO outcomes, including 2 following subretinal gene therapy treatments, 75% included fewer than 10 participants with AO imaging data. Of 100 studies that included participants' genetic diagnoses, the most common IRD genes with AO outcomes are CNGA3, CNGB3, CHM, USH2A, and ABCA4. Confocal reflectance AO scanning laser ophthalmoscopy was the most reported imaging modality, followed by flood-illuminated AO and split-detector AO. The most common outcome was cone density, reported quantitatively in 56% of studies. Future research areas include guidelines to reduce variability in the reporting of AO methodology and a focus on functional AO techniques to guide the development of therapeutic interventions.

Spatial Functional Characteristics of East Asian Patients With Occult Macular Dystrophy (Miyake Disease); EAOMD Report No. 2

PURPOSE

To describe the functional phenotypic features of East Asian patients with RP1L1-associated occult macular dystrophy (ie, Miyake disease).

DESIGN

An international multicenter retrospective cohort study.

METHODS

Twenty-eight participants (53 eyes) with Miyake disease were enrolled at 3 centers (in Japan, China, and South Korea). Ophthalmologic examinations including spectral-domain optical coherence tomography (SDOCT) and multifocal electroretinogram (mfERG) were performed. Patients were classified into 3 functional groups based on mfERG: Group 1, paracentral dysfunction with relatively preserved central/peripheral function; Group 2, homogeneous central dysfunction with preserved peripheral function; and Group 3, widespread dysfunction over the recorded area. Three functional phenotypes were compared in clinical parameters and SDOCT morphologic classification (severe phenotype, blurred/flat ellipsoid zone and absence of the interdigitation zone; mild phenotype, preserved ellipsoid zone).

RESULTS

There were 8 eyes in Group 1, 40 eyes in Group 2, and 5 eyes in Group 3. The patients in Group 1 showed significantly later onset (P = .005) and shorter disease duration (P = .002), compared with those in Group 2. All 8 eyes in Group 1 showed the mild morphologic phenotype, while 43 of 45 eyes in Groups 2 and 3 presented the severe phenotype, which identified a significant association between the functional grouping and the morphologic classification (P < .001).

CONCLUSIONS

A spectrum of functional phenotypes of Miyake disease was first documented with identification of 3 functional subtypes. Patients with paracentral dysfunction had the mildest phenotype, and those with homogeneous central or widespread dysfunction showed overlapping clinical findings with severe photoreceptor changes, suggesting various extents of visual impairment.

Automated identification of cone photoreceptors in adaptive optics optical coherence tomography images using transfer learning

Automated measurements of the human cone mosaic requires the identification of individual cone photoreceptors. The current gold standard, manual labeling, is a tedious process and can not be done in a clinically useful timeframe. As such, we present an automated algorithm for identifying cone photoreceptors in adaptive optics optical coherence tomography (AO-OCT) images. Our approach fine-tunes a pre-trained convolutional neural network originally trained on AO scanning laser ophthalmoscope (AO-SLO) images, to work on previously unseen data from a different imaging modality. On average, the automated method correctly identified 94% of manually labeled cones when compared to manual raters, from twenty different AO-OCT images acquired from five normal subjects. Voronoi analysis confirmed the general hexagonal-packing structure of the cone mosaic as well as the general cone density variability across portions of the retina. The consistency of our measurements demonstrates the high reliability and practical utility of having an automated solution to this problem.

Assessment of inner and outer retinal layer metrics on the Cirrus HD-OCT Platform in normal eyes

Purpose

Ellipsoid zone (EZ) and outer retinal integrity are strongly linked to visual prognosis, but quantitative normative data is lacking. This study evaluates the EZ, outer retina, and inner retina in eyes without macular disease across a wide age spectrum.

Methods

An IRB-approved study was performed for eyes without macular pathology undergoing Spectral Domain Optical Coherence Tomography (SD-OCT) scans on the Cirrus HD-OCT system (Carl Zeiss Meditec, Oberkochen, Germany). Scans were analyzed using a previously described automated EZ mapping tool with line-by-line manual verification. Segmentation included internal limiting membrane (ILM), outer nuclear layer/Henle fiber layer complex (ONL/HFL), EZ, and the retinal pigment epithelium (RPE). The output included metrics for the inner retina (ILM-OPL/HFL), outer retina (ONL/HFL-RPE), EZ-RPE area and volume, and en face EZ mapping. EZ-RPE attenuation on en face mapping was defined as EZ-RPE thickness < 20 um, and total attenuation was 0 um. Imaging parameters were assessed for the group and compared to age, sex, visual acuity and spherical equivalent.

Results

167 eyes from 167 subjects were included. Mean age was 49.7 years (range 10–84 years). The mean foveal retinal thickness was 200.58 ± 19.22 um. Mean inner retinal thickness was 21.47 ± 13.60 um. Mean outer retinal thickness was 179.11 ± 18.52 um. Mean EZ-RPE thickness was 50.58 ± 6.01um. The mean EZ-RPE volume was 1.20 ± 0.10 mm³. Mean EZ attenuation percentage per macular map area was 0.87% ± 1.13% and mean percentage total attenuation was 0.12% ± 0.14%. Total and inner retinal thickness metrics decreased with age. Mean outer retinal thickness increased with age. EZ-RPE parameters were unchanged with age. However, EZ attenuation was negatively correlated with age.

Conclusion

This study provides important information for inner and outer retinal parameters. Future research on quantitative EZ integrity can utilize this data for comparison.

Deep learning based detection of cone photoreceptors with multimodal adaptive optics scanning light ophthalmoscope images of achromatopsia

Fast and reliable quantification of cone photoreceptors is a bottleneck in the clinical utilization of adaptive optics scanning light ophthalmoscope (AOSLO) systems for the study, diagnosis, and prognosis of retinal diseases. To-date, manual grading has been the sole reliable source of AOSLO quantification, as no automatic method has been reliably utilized for cone detection in real-world low-quality images of diseased retina. We present a novel deep learning based approach that combines information from both the confocal and non-confocal split detector AOSLO modalities to detect cones in subjects with achromatopsia. Our dual-mode deep learning based approach outperforms the state-of-the-art automated techniques and is on a par with human grading.

Open source software for automatic detection of cone photoreceptors in adaptive optics ophthalmoscopy using convolutional neural networks

Imaging with an adaptive optics scanning light ophthalmoscope (AOSLO) enables direct visualization of the cone photoreceptor mosaic in the living human retina. Quantitative analysis of AOSLO images typically requires manual grading, which is time consuming, and subjective; thus, automated algorithms are highly desirable. Previously developed automated methods are often reliant on ad hoc rules that may not be transferable between different imaging modalities or retinal locations. In this work, we present a convolutional neural network (CNN) based method for cone detection that learns features of interest directly from training data. This cone-identifying algorithm was trained and validated on separate data sets of confocal and split detector AOSLO images with results showing performance that closely mimics the gold standard manual process. Further, without any need for algorithmic modifications for a specific AOSLO imaging system, our fully-automated multi-modality CNN-based cone detection method resulted in comparable results to previous automatic cone segmentation methods which utilized ad hoc rules for different applications. We have made free open-source software for the proposed method and the corresponding training and testing datasets available online.

Automated measurements of human cone photoreceptor density in healthy and degenerative retina by region-based segmentation

Purpose

The purpose of this study was to develop an algorithm based on region-based segmentation for automated calculations of human cone photoreceptor density of en face images obtained by an adaptive optics scanning laser ophthalmoscope (AOSLO).

Subjects and methods

Cone mosaics of 15 eyes of 15 healthy subjects were photographed by a custom-built AOSLO. The cone density was calculated at 0.5, 1.0, and 1.5 mm temporal from the fovea using a region-based segmentation method (RSM) developed in our laboratory. The cone density was also determined by a manual identification method (MIM) and a conventional spatial filtering method (SFM). The cone densities of three eyes of three patients with retinal degeneration were calculated by the three methods and compared to the results from normal eyes.

Results

The cone densities in healthy retinas determined by the RSM at 0.5, 1.0, and 1.5 mm temporal from the fovea were 28,436, 21,233, and 13,620 cells/mm², respectively. These densities were in good agreement with a histological study and with in vivo AOSLO studies. The cone densities determined by RSM were different from those determined by MIM with a difference of 5% in healthy eyes. In eyes with retinal degeneration, with the appropriate threshold-level settings or spatial frequency bandwidth, the cone density measured by MIM was significantly closer to that measured by RSM than by SFM.

Conclusion

These results suggest that our method is more stable than conventional methods in cases of non-periodical photoreceptor structures such as the affected retinal area. Our method can be used in the longitudinal follow-up of retinal degenerative diseases and to determine the effect of therapy.

Multimodal imaging of foveal cavitation in retinal dystrophies

PURPOSE

Inherited retinal dystrophies and cone dysfunction syndromes may show a sharp hyporeflective interruption in the outermost retinal layers on optical coherence tomography (OCT), known as foveal cavitation (FC). The aim of the study was to describe the morpho-functional features of FC in patients affected by retinal dystrophies by means of multimodal imaging.

METHODS

A consecutive series of patients affected by FC were prospectively recruited for the study. Patients underwent short-wavelength (SW) and near-infraRed (NIR) fundus autofluorescence (FAF), spectral domain OCT (SD-OCT), microperimetry (MP), and multifocal electroretinogram (mfERG). Mean size of FC on OCT was correlated with best-corrected visual acuity (BCVA).

RESULTS

Overall, 15 patients (30 eyes) were enrolled. Mean age was 38.2 ± 14.5 years (range: 10-60), with nine females (60 %). Mean BCVA was 0.5 ± 0.4 LogMAR. SD-OCT revealed focal loss of outer retinal layers and disruption of inner segment ellipsoid zone. Vertical height of FC (mean 27.77 ± 18.77 μm) was indirectly related to BCVA; complete forms of FC, with total loss of outer OCT bands, showed a poorer visual outcome. The FC size on NIR-FAF turned out to be larger with respect to SD-OCT and SW-FAF.

CONCLUSION

Our data indicate that the presence of FC worsens functional outcome in patients affected by retinal disorders; complete and higher lesions are associated with a worse morpho-functional prognosis in these eyes.

Automatic detection of cone photoreceptors in split detector adaptive optics scanning light ophthalmoscope images

Quantitative analysis of the cone photoreceptor mosaic in the living retina is potentially useful for early diagnosis and prognosis of many ocular diseases. Non-confocal split detector based adaptive optics scanning light ophthalmoscope (AOSLO) imaging reveals the cone photoreceptor inner segment mosaics often not visualized on confocal AOSLO imaging. Despite recent advances in automated cone segmentation algorithms for confocal AOSLO imagery, quantitative analysis of split detector AOSLO images is currently a time-consuming manual process. In this paper, we present the fully automatic adaptive filtering and local detection (AFLD) method for detecting cones in split detector AOSLO images. We validated our algorithm on 80 images from 10 subjects, showing an overall mean Dice's coefficient of 0.95 (standard deviation 0.03), when comparing our AFLD algorithm to an expert grader. This is comparable to the inter-observer Dice's coefficient of 0.94 (standard deviation 0.04). To the best of our knowledge, this is the first validated, fully-automated segmentation method which has been applied to split detector AOSLO images.

Ellipsoid zone on optical coherence tomography: a review

Emergence of the high-resolution optical coherence tomography has allowed better delineation of retinal layers, and many of the anatomical correlations of these layers have now been agreed upon. However, some anatomical correlates still remain contentious, such as the second hyper-reflective band, which is now termed ellipsoid zone. Despite the lack of consensus of the actual origin of the ellipsoid zone, there has been much interest in evaluating its integrity and intensity in different disease processes. This review paper aims to provide an overview of the ellipsoid zone and its clinical and research applications.

Multimodal imaging and multifocal electroretinography demonstrate autosomal recessive Stargardt disease may present like occult macular dystrophy

PURPOSE

To describe multimodal imaging and electrophysiologic characteristics of an unusual subset of patients with genetically confirmed autosomal recessive Stargardt disease (STGD1) who exhibited a central form of cone dysfunction resembling occult macular dystrophy that preceded the development of lipofuscin flecks, atrophy of retinal pigment epithelium (RPE), or full-field electroretinography abnormalities.

METHODS

Retrospective, observational descriptive case series.

RESULTS

Five patients with compound heterozygous ABCA4 mutations presented with bilateral visual acuity reduction, normal-appearing fundi, and blocked choroidal fluorescence on fluorescein angiography. One sibling each of two probands with identical genotypes was also included for analysis. Full-field electroretinography testing was normal in all patients, but multifocal electroretinography demonstrated centripetally depressed amplitudes exceeding areas of fundus autofluorescence, infrared imaging, and spectral domain optical coherence tomography abnormalities. Spectral domain optical coherence tomography initially revealed disruption of the inner segment ellipsoid band accompanying an ovoid hypofluorescent foveolar lesion. Progression to later stages was accompanied by the loss of the foveal photoreceptor outer segments, creating foveal cavitation with preservation of the RPE. Fundus autofluorescence and infrared imaging demonstrated corresponding bull's eye lesions. Over time, the foveal potential space on spectral domain optical coherence tomography collapsed, and three patients developed RPE atrophy and visible lipofuscin flecks. The flecks were detectable by fundus autofluorescence and infrared imaging earlier than by biomicroscopy. From these findings, a staging system for this subset of Stargardt disease presenting with central cone dysfunction was developed and presented herein.

CONCLUSION

Autosomal recessive Stargardt disease may present as a central cone dysfunction syndrome before the development of lipofuscin flecks, atrophy of RPE, or full-field electroretinography abnormalities. If emerging therapies for Stargardt disease succeed, early recognition and treatment of patients with preserved foveal photoreceptor and RPE cell bodies may yield a more favorable visual prognosis.

Photoreceptor inner segment ellipsoid band integrity on spectral domain optical coherence tomography

Spectral domain optical coherence tomography cross-sectional imaging of the macula has conventionally been resolved into four bands. However, some doubts were raised regarding authentication of the existence of these bands. Recently, a number of studies have suggested that the second band appeared to originate from the inner segment ellipsoids of the foveal cone photoreceptors, and therefore the previously called inner segment-outer segment junction is now referred to as inner segment ellipsoidband. Photoreceptor dysfunction may be a significant predictor of visual acuity in a spectrum of surgical and medical retinal diseases. This review aims to provide an overview and summarizes the role of the photoreceptor inner segment ellipsoid band in the management and prognostication of various vitreoretinal diseases.

Adaptive optics retinal imaging--clinical opportunities and challenges

The array of therapeutic options available to clinicians for treating retinal disease is expanding. With these advances comes the need for better understanding of the etiology of these diseases on a cellular level as well as improved non-invasive tools for identifying the best candidates for given therapies and monitoring the efficacy of those therapies. While spectral domain optical coherence tomography offers a widely available tool for clinicians to assay the living retina, it suffers from poor lateral resolution due to the eye's monochromatic aberrations. Ophthalmic adaptive optics (AO) is a technique to compensate for the eye's aberrations and provide nearly diffraction-limited resolution. The result is the ability to visualize the living retina with cellular resolution. While AO is unquestionably a powerful research tool, many clinicians remain undecided on the clinical potential of AO imaging - putting many at a crossroads with respect to adoption of this technology. This review will briefly summarize the current state of AO retinal imaging, discuss current as well as future clinical applications of AO retinal imaging, and finally provide some discussion of research needs to facilitate more widespread clinical use.

Automatic cone photoreceptor segmentation using graph theory and dynamic programming

Geometrical analysis of the photoreceptor mosaic can reveal subclinical ocular pathologies. In this paper, we describe a fully automatic algorithm to identify and segment photoreceptors in adaptive optics ophthalmoscope images of the photoreceptor mosaic. This method is an extension of our previously described closed contour segmentation framework based on graph theory and dynamic programming (GTDP). We validated the performance of the proposed algorithm by comparing it to the state-of-the-art technique on a large data set consisting of over 200,000 cones and posted the results online. We found that the GTDP method achieved a higher detection rate, decreasing the cone miss rate by over a factor of five.

Analysis of macular cone photoreceptors in a case of occult macular dystrophy

Purpose

To investigate changes in cone photoreceptors with adaptive optics (AO) fundus imaging and spectral domain optical coherence tomography (SD-OCT) in a case of occult macular dystrophy (OMD).

Patient and methods

Both eyes of a 42-year-old woman diagnosed with OMD were examined. We used an AO fundus camera to obtain images of cone photoreceptors in the macula of the OMD subject and five healthy control subjects. Correlations between the AO images and the SD-OCT images were examined. Cone photoreceptors in eight areas in the macula of OMD and healthy control subjects were analyzed and compared.

Results

SD-OCT showed a loss of the cone outer-segment tips line outside of the fovea in both eyes of the subject with OMD. The left eye with decreased visual acuity showed a discontinuous photoreceptor inner-segment and outer-segment line and cone outer-segment tips line at the fovea in SD-OCT and loss of cone mosaics as a dark spot in the AO image. In panoramic AO images and cone-density maps, less cone density was observed in a ring-like region outside the fovea than in the peripheral retina. In most of the areas examined, the cone densities were lower in the OMD eyes than in the healthy control eyes.

Conclusions

Cone densities in the macula of the OMD patient were greatly decreased. AO images were found to be useful to evaluate morphologic changes in cone photoreceptors in patients with OMD.

Adaptive optics dioptric scanning ophthalmoscope with a wider field of view similar to those of normal ophthalmoscopes

We introduce a newly developed adaptive optics dioptric scanning ophthalmoscope (AO-DSO) in which all powered optical parts were refractive lenses instead of concave or convex mirrors. By designing dioptric optics, we were able to achieve a compact instrument with a 10 deg field of view (FOV10) and 1.5 deg field of view (FOV1.5) high-resolution imaging. Although the resolution of FOV10 was sacrificed because of the variation of aberrations of the eye over the 10 deg field, our system works with AO in the case of FOV1.5 and can be used as a scanning laser ophthalmoscope with good optical slicing in the case of FOV10. To test the ability of the AO-DSO, we performed imaging on a normal subject and on a patient with occult macular dystrophy.

The inner segment/outer segment border seen on optical coherence tomography is less intense in patients with diminished cone function

PURPOSE; The integrity of the inner segment ellipsoid (ISe) band, previously called the inner segment/outer segment (IS/OS) border, seen on optical coherence tomography (OCT) scans is of clinical significance. To better understand the influence of cones on the appearance of this band, the intensity of its signal in patients with diminished cone function was examined.

METHODS

Horizontal line scans through the fovea of 30 healthy controls, 10 patients with achromatopsia (A), and six with cone dystrophy (CD) were obtained with frequency domain (fd) OCT. The fdOCT borders were segmented with a computer-aided manual technique. The ISe was divided into regions 60.1 μm wide and 19.5 μm deep. The relative ISe intensity of each region was defined as its intensity divided by the intensity of a local region, which extended in depth from the choroid to the retinal ganglion cell/retinal nerve fiber layer.

RESULTS

Except for the central fovea, all patients had a clear ISe band across the region studied, ± 3 mm from the foveal center. However, the relative ISe intensity was significantly lower (P < 0.0001) in patients (A: 1.14 ± 0.14; CD: 1.27 ± 0.14), than in controls (1.61 ± 0.16). There were no differences in the relative intensity of the other retinal layers.

CONCLUSIONS

Although present, the intensity of this ISe band is lower in patients with diminished cone function than it is in healthy controls. This is consistent with the hypothesis that both rod and cone receptors must be absent or damaged for the ISe band to be missing.

Wavefront sensorless adaptive optics ophthalmoscopy in the human eye

Wavefront sensor noise and fidelity place a fundamental limit on achievable image quality in current adaptive optics ophthalmoscopes. Additionally, the wavefront sensor 'beacon' can interfere with visual experiments. We demonstrate real-time (25 Hz), wavefront sensorless adaptive optics imaging in the living human eye with image quality rivaling that of wavefront sensor based control in the same system. A stochastic parallel gradient descent algorithm directly optimized the mean intensity in retinal image frames acquired with a confocal adaptive optics scanning laser ophthalmoscope (AOSLO). When imaging through natural, undilated pupils, both control methods resulted in comparable mean image intensities. However, when imaging through dilated pupils, image intensity was generally higher following wavefront sensor-based control. Despite the typically reduced intensity, image contrast was higher, on average, with sensorless control. Wavefront sensorless control is a viable option for imaging the living human eye and future refinements of this technique may result in even greater optical gains.