Automated Retinal Image Analysis for Evaluation of Focal Hyperpigmentary Changes in Intermediate Age-Related Macular Degeneration

HYPERREFLECTIVE FOCI NOT SEEN AS HYPERPIGMENTARY ABNORMALITIES ON COLOR FUNDUS PHOTOGRAPHS IN AGE-RELATED MACULAR DEGENERATION

PURPOSE

To investigate the prognostic value of quantifying optical coherence tomography (OCT)-defined hyperreflective foci (HRF) that do not correspond to hyperpigmentary abnormalities (HPAs) on color fundus photographs (CFPs)-HRF (OCT+/CFP-) -when considered in addition to HPA extent, for predicting late age-related macular degeneration development. This study sought to understand the impact of HRF (OCT+/CFP-) extent on visual sensitivity.

METHODS

Two hundred eighty eyes from 140 participants with bilateral large drusen underwent imaging and microperimetry at baseline, and then 6-monthly for 3-years. The extent of HPAs on CFPs and HRF (OCT+/CFP-) on OCT was quantified at baseline. Predictive models for progression to late age-related macular degeneration, accounting for drusen volume and age, were developed using HPA extent, with and without HRF (OCT+/CFP-) extent. The association between HPA and HRF (OCT+/CFP-) extent with sector-based visual sensitivity was also evaluated.

RESULTS

Incorporating HRF (OCT+/CFP-) extent did not improve the predictive performance for late age-related macular degeneration development ( P ≥ 0.32). Increasing HPA and HRF (OCT+/CFP-) extent in each sector were independently and significantly associated with reduced sector-based visual sensitivity ( P ≤ 0.004).

CONCLUSION

The addition of HRF (OCT+/CFP-) extent to HPA extent did not improve the prediction of late age-related macular degeneration development. HRF (OCT+/CFP-) extent was also independently associated with local reductions in visual sensitivity, after accounting for HPAs.

Multi-label classification of retinal disease via a novel vision transformer model

Introduction

The precise identification of retinal disorders is of utmost importance in the prevention of both temporary and permanent visual impairment. Prior research has yielded encouraging results in the classification of retinal images pertaining to a specific retinal condition. In clinical practice, it is not uncommon for a single patient to present with multiple retinal disorders concurrently. Hence, the task of classifying retinal images into multiple labels remains a significant obstacle for existing methodologies, but its successful accomplishment would yield valuable insights into a diverse array of situations simultaneously.

Methods

This study presents a novel vision transformer architecture called retinal ViT, which incorporates the self-attention mechanism into the field of medical image analysis. To note that this study supposed to prove that the transformer-based models can achieve competitive performance comparing with the CNN-based models, hence the convolutional modules have been eliminated from the proposed model. The suggested model concludes with a multi-label classifier that utilizes a feed-forward network architecture. This classifier consists of two layers and employs a sigmoid activation function.

Results and discussion

The experimental findings provide evidence of the improved performance exhibited by the suggested model when compared to state-of-the-art approaches such as ResNet, VGG, DenseNet, and MobileNet, on the publicly available dataset ODIR-2019, and the proposed approach has outperformed the state-of-the-art algorithms in terms of Kappa, F1 score, AUC, and AVG.

Longitudinal Changes of Retinal Structure in Molecularly Confirmed C1QTNF5 Patients With Late-Onset Retinal Degeneration

Purpose

The purpose of this study was to present our findings on the natural history of late-onset retinal degeneration (LORD) in patients with molecularly confirmed C1QTNF5 heterozygous pathogenic variants and assess suitability of retinal structure parameters for disease monitoring.

Methods

Sixteen patients with C1QTNF5-LORD were retrospectively identified from Moorfields Eye Hospital, UK. Fundus autofluorescence (FAF), optical coherence tomography (OCT) scans, and best-corrected visual acuity (BCVA) were collected. Area of atrophy (AA) was manually drawn in FAF images. Ellipsoid zone (EZ) width and foveal retinal thickness of the whole retina and outer retina were extracted from OCT scans. Age-related changes were tested with linear-mixed models.

Results

Patients had median age of 62.3 years (interquartile range [IQR] = 58.8-65.4 years) at baseline, and median follow-up of 5.1 years (IQR = 2.6-7.6 years). AA, EZ width, and retinal thickness parameters remained unchanged until age 50 years, but showed significant change with age thereafter (all P < 0.0001). AA and EZ width progressed rapidly (dynamic range normalized rates = 4.3-4.5%/year) from age 53.9 and 50.8 years (estimated inflection points), respectively. Retinal thickness parameters showed slower progression rates (range = 1.6-2.5%/year) from age 60 to 62.3. BCVA (median = 0.3 LogMAR, IQR = 0.0-1.0 at baseline) showed a rapid decline (3.3%) from age 70 years. Findings from patients with earlier disease showed FAF atrophy manifests in the temporal retina initially, and then progresses nasally.

Conclusions

Patients with LORD remained asymptomatic until age 50 years, before suffering rapid outer retinal degeneration. EZ width and AA showed rapid progression and high interocular correlation, representing promising outcome metrics. Clinical measures also capturing the temporal retina may be preferable, enabling earlier detection and better disease monitoring.

Translational Relevance

Area of atrophy in FAF images and OCT-measured EZ width represent promising outcome metrics for disease monitoring in patients with C1QTNF5-LORD.

Accurate drusen segmentation in optical coherence tomography via order-constrained regression of retinal layer heights

Drusen are an important biomarker for age-related macular degeneration (AMD). Their accurate segmentation based on optical coherence tomography (OCT) is therefore relevant to the detection, staging, and treatment of disease. Since manual OCT segmentation is resource-consuming and has low reproducibility, automatic techniques are required. In this work, we introduce a novel deep learning based architecture that directly predicts the position of layers in OCT and guarantees their correct order, achieving state-of-the-art results for retinal layer segmentation. In particular, the average absolute distance between our model's prediction and the ground truth layer segmentation in an AMD dataset is 0.63, 0.85, and 0.44 pixel for Bruch's membrane (BM), retinal pigment epithelium (RPE) and ellipsoid zone (EZ), respectively. Based on layer positions, we further quantify drusen load with excellent accuracy, achieving 0.994 and 0.988 Pearson correlation between drusen volumes estimated by our method and two human readers, and increasing the Dice score to 0.71 ± 0.16 (from 0.60 ± 0.23) and 0.62 ± 0.23 (from 0.53 ± 0.25), respectively, compared to a previous state-of-the-art method. Given its reproducible, accurate, and scalable results, our method can be used for the large-scale analysis of OCT data.

Hyperreflective Foci in Age-Related Macular Degeneration are Associated with Disease Severity and Functional Impairment

PURPOSE

To analyze presence of hyperreflective foci (HRF) across different age-related macular degeneration (AMD) severities and examine its correlation with other structural and functional AMD features.

DESIGN

Longitudinal, single-center, case-control study.

PARTICIPANTS

One hundred and fifty-eight participants aged > 50 years old with varying AMD severities (including no AMD).

METHODS

Color fundus imaging was used to assess AMD severity and hyperpigmentation (PGM) presence. Subretinal drusenoid deposits (SDD) and HRF were detected on OCT volumes. The correlations of HRF with additional AMD features were evaluated using linear and logistic mixed-effects models. One study eye per participant underwent dark adaptation (DA) testing to measure rod intercept time (RIT) for structure function associations. Eyes were followed longitudinally and changes in AMD severity and RIT were measured relative to HRF presence.

MAIN OUTCOME MEASURES

The primary outcome was presence of HRF, which was compared with presence of other AMD features and DA impairment.

RESULTS

One hundred and fifty-eight participants (median baseline age of 73.1 [interquartile range (IQR) = 66-79] years) contributing 1277 eye visits were included. Hyperreflective foci (HRF) were detected more frequently in higher AMD severities. Hyperreflective-foci presence was significantly associated with PGM presence (odds ratio 832.9, P < 0.001) and SDD presence (odds ratio 9.42, P = 0.017). Eyes with HRF demonstrated significantly longer DA (median 27.1 [IQR = 16-40] minutes) than those without HRF (13.5 [10-22] minutes) but less than eyes with SDD only (40 [28-40] minutes). Highest RIT values were found in eyes with both HRF and SDD (40.0 [40-40] minutes). Age and HRF explained a similar proportion of RIT variability as age and SDD. Eyes that developed HRF demonstrated baseline RITs closer to eyes with HRF at baseline, compared with eyes that never developed HRF (29.1 [16-40], 38.5 [22-40] versus 13.1 [10-22] minutes; Kruskal-Wallis P < 0.001).

CONCLUSIONS

The progressively increased presence of HRF in higher AMD severities, and its correlation with previously associated AMD biomarkers, suggests HRF is an important OCT feature adding to the understanding of disease progression. Hyperreflective foci presence was associated with delays in DA, indicating HRF is a marker for visual cycle impairment.

FINANCIAL DISCLOSURE

Proprietary or commercial disclosure may be found after the references.

The Development and Clinical Application of Innovative Optical Ophthalmic Imaging Techniques

The field of ophthalmic imaging has grown substantially over the last years. Massive improvements in image processing and computer hardware have allowed the emergence of multiple imaging techniques of the eye that can transform patient care. The purpose of this review is to describe the most recent advances in eye imaging and explain how new technologies and imaging methods can be utilized in a clinical setting. The introduction of optical coherence tomography (OCT) was a revolution in eye imaging and has since become the standard of care for a plethora of conditions. Its most recent iterations, OCT angiography, and visible light OCT, as well as imaging modalities, such as fluorescent lifetime imaging ophthalmoscopy, would allow a more thorough evaluation of patients and provide additional information on disease processes. Toward that goal, the application of adaptive optics (AO) and full-field scanning to a variety of eye imaging techniques has further allowed the histologic study of single cells in the retina and anterior segment. Toward the goal of remote eye care and more accessible eye imaging, methods such as handheld OCT devices and imaging through smartphones, have emerged. Finally, incorporating artificial intelligence (AI) in eye images has the potential to become a new milestone for eye imaging while also contributing in social aspects of eye care.

From Data to Deployment: The Collaborative Community on Ophthalmic Imaging Roadmap for Artificial Intelligence in Age-Related Macular Degeneration

OBJECTIVE

Health care systems worldwide are challenged to provide adequate care for the 200 million individuals with age-related macular degeneration (AMD). Artificial intelligence (AI) has the potential to make a significant, positive impact on the diagnosis and management of patients with AMD; however, the development of effective AI devices for clinical care faces numerous considerations and challenges, a fact evidenced by a current absence of Food and Drug Administration (FDA)-approved AI devices for AMD.

PURPOSE

To delineate the state of AI for AMD, including current data, standards, achievements, and challenges.

METHODS

Members of the Collaborative Community on Ophthalmic Imaging Working Group for AI in AMD attended an inaugural meeting on September 7, 2020, to discuss the topic. Subsequently, they undertook a comprehensive review of the medical literature relevant to the topic. Members engaged in meetings and discussion through December 2021 to synthesize the information and arrive at a consensus.

RESULTS

Existing infrastructure for robust AI development for AMD includes several large, labeled data sets of color fundus photography and OCT images; however, image data often do not contain the metadata necessary for the development of reliable, valid, and generalizable models. Data sharing for AMD model development is made difficult by restrictions on data privacy and security, although potential solutions are under investigation. Computing resources may be adequate for current applications, but knowledge of machine learning development may be scarce in many clinical ophthalmology settings. Despite these challenges, researchers have produced promising AI models for AMD for screening, diagnosis, prediction, and monitoring. Future goals include defining benchmarks to facilitate regulatory authorization and subsequent clinical setting generalization.

CONCLUSIONS

Delivering an FDA-authorized, AI-based device for clinical care in AMD involves numerous considerations, including the identification of an appropriate clinical application; acquisition and development of a large, high-quality data set; development of the AI architecture; training and validation of the model; and functional interactions between the model output and clinical end user. The research efforts undertaken to date represent starting points for the medical devices that eventually will benefit providers, health care systems, and patients.

Multimodal Retinal Image Analysis via Deep Learning for the Diagnosis of Intermediate Dry Age-Related Macular Degeneration: A Feasibility Study

Results

The CNN trained using OCT alone showed a diagnostic accuracy of 94%, whilst the OCT-A trained CNN resulted in an accuracy of 91%. When multiple modalities were combined, the CNN accuracy increased to 96% in the AMD cohort.

Conclusions

Here we demonstrate that superior diagnostic accuracy can be achieved when deep learning is combined with multimodal image analysis.

Prognostic Value of Retinal Layers in Comparison with Other Risk Factors for Conversion of Intermediate Age-related Macular Degeneration

PURPOSE

To analyze longitudinal thickness changes of retinal layers in comparison with established risk factors in eyes with age-related macular degeneration (AMD) with regard to their prognostic value for conversion into advanced AMD stages.

DESIGN

Prospective, longitudinal natural history study.

PARTICIPANTS

Ninety-one eyes of 91 patients with AMD (73.3±7.3 years; 62 female patients [50.4%]) of the Molecular Diagnostic of Age-related Macular Degeneration (MODIAMD) study without exudative or nonexudative late-stage AMD in the study eye at baseline.

METHODS

At each annual follow-up visit, all subjects underwent ophthalmic examination with assessment of best-corrected visual acuity (BCVA) and retinal imaging, including spectral-domain OCT (SD-OCT), over a study period of 6 years.

PURPOSE

To analyze longitudinal thickness changes of retinal layers in comparison with established risk factors in eyes with age-related macular degeneration (AMD) with regard to their prognostic value for conversion into advanced AMD stages.

MAIN OUTCOME MEASURES

Qualitative structural AMD features and SD-OCT-based quantitative thickness changes of different retinal layers, such as the retinal pigment epithelium-drusen complex (RPEDC), were assessed by multimodal imaging. Their prognostic relevance regarding disease conversion was determined using Cox regression (cloglog link function).

RESULTS

In the multivariable analysis, the presence of focal hyperpigmentation, almost reaching statistical significance, showed the strongest effect regarding the development of nonexudative late-stage AMD (hazard ratio [HR], 5.88; 95% confidence interval [CI], 0.69-50.2; P = 0.052) followed by the presence of refractile drusen (HR, 4.82; 95% CI, 1.33-17.44; P = 0.0164). A thickening of the RPEDC was the only assessed retinal layer that exhibited a significant effect on the development of nonexudative advanced AMD (HR, 1.03; 95% CI, 1.0-1.07; P = 0.0393), whereas no association was observable for the other retinal layers. Neither qualitative nor quantitative markers were significant predictors for the development of exudative late-stage AMD (P > 0.05).

CONCLUSIONS

The results indicate that the development of both exudative and nonexudative AMD is associated with distinct prognostic features. However, compared with the assessment of qualitative AMD features, the quantification of retinal layers on average across the central retina had less prognostic impact. Further studies are needed to identify and validate robust biomarkers in early AMD stages.

Artificial intelligence in retina

Major advances in diagnostic technologies are offering unprecedented insight into the condition of the retina and beyond ocular disease. Digital images providing millions of morphological datasets can fast and non-invasively be analyzed in a comprehensive manner using artificial intelligence (AI). Methods based on machine learning (ML) and particularly deep learning (DL) are able to identify, localize and quantify pathological features in almost every macular and retinal disease. Convolutional neural networks thereby mimic the path of the human brain for object recognition through learning of pathological features from training sets, supervised ML, or even extrapolation from patterns recognized independently, unsupervised ML. The methods of AI-based retinal analyses are diverse and differ widely in their applicability, interpretability and reliability in different datasets and diseases. Fully automated AI-based systems have recently been approved for screening of diabetic retinopathy (DR). The overall potential of ML/DL includes screening, diagnostic grading as well as guidance of therapy with automated detection of disease activity, recurrences, quantification of therapeutic effects and identification of relevant targets for novel therapeutic approaches. Prediction and prognostic conclusions further expand the potential benefit of AI in retina which will enable personalized health care as well as large scale management and will empower the ophthalmologist to provide high quality diagnosis/therapy and successfully deal with the complexity of 21st century ophthalmology.