Zhihong Jewel Hu

The rapid evolution of diagnostic and therapeutic technologies continues to reshape ophthalmology and its interface with systemic diseases

Age-related macular degeneration (AMD) is the leading cause of central vision loss in aging populations. Geographic atrophy (GA) is the advanced, non-neovascular form of AMD. Predicting the longitudinal progression of GA remains a critical challenge in ophthalmic clinical practice and clinical trial design. Forecasting the trajectory of GA is complicated by highly variable growth rates and the inherent scarcity of long-term, high-quality imaging data. To address these challenges, we introduce the Sliding Window Attention U-Net (SWAU-Net), a hybrid architecture that integrates Transformer-based temporal modeling of GA growth with precise spatial modeling of GA location with a U-Net convolutional neural network (CNN). To ensure generalization in the low-data regime, SWAU-Net embeds explicit temporal and geometric consistency priors via a weight-shared Sliding Window Attention core and feature-level regularization that preserves sparse, high-frequency lesion boundaries across frames. Experimental results demonstrate that these structural constraints prevent the model from overfitting to imaging noise, achieving a Growth Mask Dice Similarity Coefficient (DSC) of 0.66 (representing the spatial overlap between the predicted and ground truth lesion expansion regions), a significant improvement over unregularized Transformer and standard recurrent baseline models. Our framework provides a robust tool for predicting GA lesion trajectories, potentially supporting more efficient clinical trial designs and personalized patient monitoring.

Background/Objectives: Alzheimer’s disease (AD) is the leading cause of dementia, which is an inevitable consequence of aging. Early detection of AD, or detection during the pre-AD stage, is beneficial, as it enables timely intervention to reduce modifiable risk factors, which may help prevent or delay the progression to dementia. On the one hand, plasma biomarkers have demonstrated great promise in predicting cognitive decline. On the other hand, in recent years, ocular imaging features, particularly the thickness of retinal layers measured by spectral-domain optical coherence tomography (SD-OCT), are emerging as possible non-invasive, non-contact surrogate markers for early detection and monitoring of neurodegeneration. This pilot study aims to identify retinal layer thickness changes across the entire retina linked to plasma AD biomarkers in cognitively healthy (CH) elderly individuals at risk for AD. Methods: Eleven CH individuals (20 eyes total) were classified in the pre-AD stage by plasma β-amyloid (Aβ)42/40 ratio < 0.10 and underwent SD-OCT. A deep-learning-derived automated algorithm was used to segment retinal layers on OCT (with manual correction when needed). Multiple layer thicknesses throughout the entire retina (including the inner retina, the outer retina, and the choroid) were measured in the inner ring (1–3 mm) and outer ring (3–6 mm) of the Early Treatment Diabetic Retinopathy Study (ETDRS). Relationships between retinal layers and plasma biomarkers were analyzed by ridge regression/bootstrapping. Results: Results showed that photoreceptor inner segment (PR-IS) thinning had the largest size effect with neurofilament light chain. Additional findings revealed thinning or thickening of the other retinal layers in association with increasing levels of glial fibrillary acidic protein and phosphorylated tau at threonine 181 and 217 (p-tau181 and p-tau217). Conclusions: This pilot study suggests that retinal layer-specific signatures exist, with PR-IS thinning as the largest effect, indicating neurodegeneration in pre-AD. Further research is needed to confirm the findings of this pilot study using larger longitudinal pre-AD cohorts and comparative analyses with healthy aging adults.

Optical coherence tomography angiography (OCTA) is a noninvasive imaging technique used to visualize retinal blood flow and identify changes in vascular density and enlargement or distortion of the foveal avascular zone (FAZ), which are indicators of various eye diseases. Although several automated FAZ detection and segmentation algorithms have been developed for use with OCTA, their performance can vary significantly due to differences in data accessibility of OCTA in different retinal pathologies, and differences in image quality in different subjects and/or different OCTA devices. For example, data from subjects with direct macular damage, such as in age-related macular degeneration (AMD), are more readily available in eye clinics, while data on macular damage due to systemic diseases like Alzheimer’s disease are often less accessible; data from healthy subjects may have better OCTA quality than subjects with ophthalmic pathologies. Typically, segmentation algorithms make use of convolutional neural networks and, more recently, vision transformers, which make use of both long-range context and fine-grained detail. However, transformers are known to be data-hungry, and may overfit small datasets, such as those common for FAZ segmentation in OCTA, to which there is limited access in clinical practice. To improve model generalization in low-data or imbalanced settings, we propose a multi-condition transformer-based architecture that uses four teacher encoders to distill knowledge into a shared base model, enabling the transfer of learned features across multiple datasets. These include intra-modality distillation using OCTA datasets from four ocular conditions: healthy aging eyes, Alzheimer’s disease, AMD, and diabetic retinopathy; and inter-modality distillation incorporating color fundus photographs of subjects undergoing laser photocoagulation therapy. Our multi-condition model achieved a mean Dice Index of 83.8% with pretraining, outperforming single-condition models (mean of 83.1%) across all conditions. Pretraining on color fundus photocoagulation images improved the average Dice Index by a small margin on all conditions except AMD (1.1% on single-condition models, and 0.1% on multi-condition models). Our architecture demonstrates potential for broader applications in detecting and analyzing ophthalmic and systemic diseases across diverse imaging datasets and settings.

The diagnosis of primary open-angle glaucoma (POAG) progression based on structural imaging such as color fundus photos (CFPs) is challenging due to the limited number of early biomarkers, as commonly determined by clinicians, and the inherent variability in optic nerve heads (ONHs) between individuals. Moreover, while visual function is the main concern for glaucoma patients, and the ability to infer future visual outcome from imaging will benefit patients by early intervention, there is currently no available tool for this. To detect glaucoma progression from ocular hypertension both structurally and functionally, and identify potential objective early biomarkers associated with progression, we developed and evaluated deep convolutional long short-term memory (CNN-LSTM) neural network models using longitudinal CFPs from the Ocular Hypertension Treatment Study (OHTS). Patients were categorized into four diagnostic groups for model input: healthy, POAG with optic disc changes, POAG with visual field (VF) changes, and POAG with both optic disc and VF changes. Gradient-weighted class activation mapping (Grad-CAM) was employed for the post hoc visualization of image features, which may be associated with the objective POAG biomarkers (rather than the biomarkers determined by clinicians). The CNN-LSTM models for the detection of POAG progression achieved promising performance results both for the structural and functional models, with an area under curve (AUC) performance of 0.894 for the disc-only group, 0.911 for the VF-only group, and 0.939 for the disc and VF group. The model demonstrated high precision (0.984) and F1-score (0.963) in the both-changes group (disc + VF). Our preliminary investigation for early POAG biomarkers with Grad-CAM feature visualization signified that retinal vasculature could serve as an early and objective biomarker for POAG progression, complementing the traditionally used optic disc features and improving clinical workflows.

Stargardt atrophy and geographic atrophy (GA) represent pivotal endpoints in FDA-approved clinical trials. Predicting atrophy progression is crucial for evaluating drug efficacy. Fundus autofluorescence (FAF), the standard 2D imaging modality in these trials, has limitations in patient comfort. In contrast, spectral-domain optical coherence tomography (SD-OCT), a 3D imaging modality, is more patient friendly but suffers from lower image quality. This study has two primary objectives: (1) develop an efficient predictive modeling for the generation of future FAF images and prediction of future Stargardt atrophic (as well as GA) regions and (2) develop an efficient predictive modeling with advanced 3D OCT features at ellipsoid zone (EZ) for the comparative performance in the generation of future enface EZ maps and prediction of future Stargardt atrophic regions on OCT as on FAF. To achieve these goals, we propose two deep neural networks (termed ReConNet and ReConNet-Ensemble) with recurrent learning units (long short-term memory, LSTM) integrating with a convolutional neural network (CNN) encoder–decoder architecture and concurrent learning units integrated by ensemble/multiple recurrent learning channels. The ReConNet, which incorporates LSTM connections with CNN, is developed for the first goal on longitudinal FAF. The ReConNet-Ensemble, which incorporates multiple recurrent learning channels based on enhanced EZ enface maps to capture higher-order inherent OCT EZ features, is developed for the second goal on longitudinal OCT. Using FAF images at months 0, 6, and 12 to predict atrophy at month 18, the ReConNet achieved mean (±standard deviation, SD) and median Dice coefficients of 0.895 (±0.086) and 0.922 for Stargardt atrophy and 0.864 (±0.113) and 0.893 for GA. Using SD-OCT images at months 0 and 6 to predict atrophy at month 12, the ReConNet-Ensemble achieved mean and median Dice coefficients of 0.882 (±0.101) and 0.906 for Stargardt atrophy. The prediction performance on OCT images is comparably good to that on FAF. These results underscore the potential of SD-OCT for efficient and practical assessment of atrophy progression in clinical trials and retina clinics, complementing or surpassing the widely used FAF imaging technique.

Stargardt disease and age-related macular degeneration are the leading causes of blindness in the juvenile and geriatric populations, respectively. The formation of atrophic regions of the macula is a hallmark of the end-stages of both diseases. The progression of these diseases is tracked using various imaging modalities, two of the most common being fundus autofluorescence (FAF) imaging and spectral-domain optical coherence tomography (SD-OCT). This study seeks to investigate the use of longitudinal FAF and SD-OCT imaging (month 0, month 6, month 12, and month 18) data for the predictive modelling of future atrophy in Stargardt and geographic atrophy. To achieve such an objective, we develop a set of novel deep convolutional neural networks enhanced with recurrent network units for longitudinal prediction and concurrent learning of ensemble network units (termed ReConNet) which take advantage of improved retinal layer features beyond the mean intensity features. Using FAF images, the neural network presented in this paper achieved mean (± standard deviation, SD) and median Dice coefficients of 0.895 (± 0.086) and 0.922 for Stargardt atrophy, and 0.864 (± 0.113) and 0.893 for geographic atrophy. Using SD-OCT images for Stargardt atrophy, the neural network achieved mean and median Dice coefficients of 0.882 (± 0.101) and 0.906, respectively. When predicting only the interval growth of the atrophic lesions with FAF images, mean (± SD) and median Dice coefficients of 0.557 (± 0.094) and 0.559 were achieved for Stargardt atrophy, and 0.612 (± 0.089) and 0.601 for geographic atrophy. The prediction performance in OCT images is comparably good to that using FAF which opens a new, more efficient, and practical door in the assessment of atrophy progression for clinical trials and retina clinics, beyond widely used FAF. These results are highly encouraging for a high-performance interval growth prediction when more frequent or longer-term longitudinal data are available in our clinics. This is a pressing task for our next step in ongoing research.

Purpose: Lack of valid end points impedes developing therapeutic strategies for early age-related macular degeneration (AMD). Delayed rod-mediated dark adaptation (RMDA) is the first functional biomarker for incident early AMD. The relationship between RMDA and the status of outer retinal bands on optical coherence tomography (OCT) have not been well defined. This study aims to characterize these relationships in early and intermediate AMD. Methods: Baseline data from 476 participants was assessed including eyes with early AMD (n = 138), intermediate AMD (n = 101), and normal aging (n = 237). Participants underwent volume OCT imaging of the macula and rod intercept time (RIT) was measured. The ellipsoid zone (EZ) and interdigitation zone (IZ) on all OCT B-scans of the volumes were segmented. The area of detectable EZ and IZ, and mean thickness of IZ within the Early Treatment Diabetic Retinopathy Study (ETDRS) grid were computed and associations with RIT were assessed by Spearman's correlation coefficient and age adjusted. Results: Delayed RMDA (longer RIT) was most strongly associated with less preserved IZ area (r = −0.591; P < 0.001), followed by decreased IZ thickness (r = −0.434; P < 0.001), and EZ area (r = −0.334; P < 0.001). This correlation between RIT and IZ integrity was not apparent when considering normal eyes alone within 1.5 mm of the fovea. Conclusions: RMDA is correlated with the status of outer retinal bands in early and intermediate AMD eyes, particularly, the status of the IZ. This correlation is consistent with a previous analysis of only foveal B-scans and is biologically plausible given that retinoid availability, involving transfer at the interface attributed to the IZ, is rate-limiting for RMDA.

Purpose: While intermediate and late age-Related Macular Degeneration (AMD) have been widely investigated, rare studies were focused on the pathophysiologic mechanism of early AMD. Delayed rod-mediated dark adaptation (RMDA) is the first functional biomarker for incident early AMD. The status of outer retinal bands on optical coherence tomography (OCT) may be potential imaging biomarkers and the purpose is to investigate the hypothesis that the integrity of interdigitation zone (IZ) may provide insight into the health of photoreceptors and retinal pigment epithelium (RPE) in early AMD. Methods: We establish the structure-function relationship between ellipsoid zone (EZ) integrity and RMDA, and IZ integrity and RMDA in a large-scale OCT dataset from eyes with normal aging (n=237), early AMD (n=138), and intermediate AMD (n=101) by utilizing a novel deep-learning-derived algorithm with manual correction when needed to segment the EZ and IZ on OCT B-scans (57,596 B-scans), and utilizing the AdaptDx device to measure RMDA. Results: Our data demonstrates that slower RMDA is associated with less preserved EZ (r = -0.334; p<0.001) and IZ area (r = -0.591; p<0.001), and decreased IZ thickness (r = -0.434; p<0.001). These associations are not apparent when considering normal eyes alone. Conclusions: The association with IZ area and RMDA in large-scale data is biologically plausible because retinoid availability and transfer at the interface attributed to IZ is rate-limiting for RMDA. This study supports the hypothesis that the IZ integrity provides insight into the health of photoreceptors and RPE in early AMD and is a potential new imaging biomarker.