Search

While Alzheimers disease and other neurodegenerative diseases have traditionally been viewed as brain disorders, there is growing evidence indicating their manifestation in the eyes as well. The retina, being a developmental extension of the brain, represents the only part of the central nervous system that can be noninvasively imaged at a high spatial resolution. The discovery of the specific pathological hallmarks of Alzheimers disease in the retina of patients holds great promise for disease diagnosis and monitoring, particularly in the early stages where disease progression can potentially be slowed. Among various retinal imaging methods, hyperspectral imaging has garnered significant attention in this field. It offers a label-free approach to detect disease biomarkers, making it especially valuable for large-scale population screening efforts. In this review, we discuss recent advances in the field and outline the current bottlenecks and enabling technologies that could propel this field toward clinical translation.

SIGNIFICANCE: Microcirculation and neurovascular coupling are important parameters to study in neurological and neuro-ophthalmic conditions. As the retina shares many similarities with the cerebral cortex and is optically accessible, a special focus is directed to assessing the chorioretinal structure, microvasculature, and hemodynamics of mice, a vital animal model for vision and neuroscience research. AIM: We aim to introduce an optical imaging tool enabling in vivo volumetric mouse retinal monitoring of vascular hemodynamics with high temporal resolution. APPROACH: We translated the spatio-temporal optical coherence tomography (STOC-T) technique into the field of small animal imaging by designing a new optical system that could compensate for the mouse eye refractive error. We also developed post-processing algorithms, notably for the assessment of (i) localized hemodynamics from the analysis of pulse wave-induced Doppler artifact modulation and (ii) retinal tissue displacement from phase-sensitive measurements. RESULTS: We acquired high-quality, in vivo volumetric mouse retina images at a rate of 113 Hz over a lateral field of view of ∼ 500    μ m . We presented high-resolution en face images of the retinal and choroidal structure and microvasculature from various layers, after digital aberration correction. We were able to measure the pulse wave velocity in capillaries of the outer plexiform layer with a mean speed of 0.35 mm/s and identified venous and arterial pulsation frequency and phase delay. We quantified the modulation amplitudes of tissue displacement near major vessels (with peaks of 150 nm), potentially carrying information about the biomechanical properties of the retinal layers involved. Last, we identified the delays between retinal displacements due to the passing of venous and arterial pulse waves. CONCLUSIONS: The developed STOC-T system provides insights into the hemodynamics of the mouse retina and choroid that could be beneficial in the study of neurovascular coupling and vasculature and flow speed anomalies in neurological and neuro-ophthalmic conditions.

BACKGROUND: Atherosclerotic cardiovascular disease (ASCVD) is a leading cause of death globally, and early detection of high-risk individuals is essential for initiating timely interventions. The authors aimed to develop and validate a deep learning (DL) model to predict an individuals elevated 10-year ASCVD risk score based on retinal images and limited demographic data. METHODS: The study used 89,894 retinal fundus images from 44,176 UK Biobank participants (96% non-Hispanic White, 5% diabetic) to train and test the DL model. The DL model was developed using retinal images plus age, race/ethnicity, and sex at birth to predict an individuals 10-year ASCVD risk score using the pooled cohort equation (PCE) as the ground truth. This model was then tested on the US EyePACS 10K dataset (5.8% non-Hispanic White, 99.9% diabetic), composed of 18,900 images from 8969 diabetic individuals. Elevated ASCVD risk was defined as a PCE score of ≥7.5%. RESULTS: In the UK Biobank internal validation dataset, the DL model achieved an area under the receiver operating characteristic curve of 0.89, sensitivity 84%, and specificity 90%, for detecting individuals with elevated ASCVD risk scores. In the EyePACS 10K and with the addition of a regression-derived diabetes modifier, it achieved sensitivity 94%, specificity 72%, mean error -0.2%, and mean absolute error 3.1%. CONCLUSION: This study demonstrates that DL models using retinal images can provide an additional approach to estimating ASCVD risk, as well as the value of applying DL models to different external datasets and opportunities about ASCVD risk assessment in patients living with diabetes.

Objective: We describe a patient affected by congenital stationary night blindness (CSNB) secondary to CACNA1F and optic neuropathy associated with an AFG3L2 variant. Methods: We performed comprehensive neuro‐ophthalmologic examinations, retinal imaging, complete ocular electrophysiology, and brain and optic nerve MRI. Genomic DNA was extracted from the peripheral blood. The patient's DNA was then investigated by next‐generation sequencing (NGS) with a panel including 32 genes associated with retinal dystrophy and therefore with a panel including seven genes associated with genetic forms of optic atrophy. Results: The genetic analysis identified a pathogenetic CACNA1F variant causing CSNB and a heterozygous variant in AFG3L2 that alters OPA1 processing and is known to be associated with OPA1‐like optic neuropathy. Conclusion: Optic disc atrophy has been previously described as an atypical feature in the phenotype of CSNB CACNA1F‐related. In this patient, we found a variant of the AFG3L2 gene that presumably explains the presence of optic atrophy in a subject affected by CSNB. Clinical Relevance: The clinical evidence of optic atrophy, which is atypical in CSNB, should raise the suspicion of concomitant hereditary optic neuropathy and emphasize the importance of broad genetic diagnostic testing to better define the genotype–phenotype correlation. [ABSTRACT FROM AUTHOR]

Owing to the epidemic growth of diabetes, ophthalmologists need to examine the huge fundus images for diagnosing the disease of Diabetic Retinopathy (DR). Without proper knowledge, people are too lethargic to detect the DR. Therefore, the early diagnosis system is requisite for treating ailments in the medical industry. Therefore, a novel deep model-based DR detection structure is recommended to fix the aforementioned difficulties. The developed deep model-based diabetic retinopathy detection process is performed adaptively. The DR detection process is imitated by garnering the images from benchmark sources. The gathered images are further preceded by the abnormality segmentation phase. Here, the Residual TransUNet with Enhanced loss function is used to employ the abnormality segmentation, and the loss function in this structure may be helpful to lessen the error in the segmentation procedure. Further, the segmented images are passed to the final phase of retinopathy detection. At this phase, the detection is carried out through the Adaptive Multiscale MobileNet. The variables in the AMMNet are optimized by the Adaptive Puzzle Optimization to obtain better detection performance. Finally, the effectiveness of the offered approach is confirmed by the experimentation procedure over various performance indices. [ABSTRACT FROM AUTHOR]

Self-supervised learning has become the cornerstone of building generalizable and transferable artificial intelligence systems in medical imaging. In particular, contrastive representation learning techniques trained on large multi-modal datasets have demonstrated impressive capabilities of producing highly transferable representations for different downstream tasks. In ophthalmology, large multi-modal datasets are abundantly available and conveniently accessible as modern retinal imaging scanners acquire both 2D fundus images and 3D optical coherence tomography (OCT) scans to assess the eye. In this context, we introduce a novel multi-modal contrastive learning-based pipeline to facilitate learning joint representations for the two retinal imaging modalities. After self-supervised pre-training on 153,306 scan pairs, we show that such a pre-training framework can provide both a retrieval system and encoders that produce comprehensive OCT and fundus image representations that generalize well for various downstream tasks on three independent external datasets, explicitly focusing on clinically pertinent prediction tasks. In addition, we show that interchanging OCT with lower-cost fundus imaging can preserve the predictive power of the trained models. [ABSTRACT FROM AUTHOR]

Age-related macular degeneration (AMD) is a leading cause of vision impairment in people over 50 years of age and has a great impact on quality of life as it affects central vision. Although there have been treatments available for the neovascular form of late AMD for decades, until now there have not been treatments available for the atrophic form of late AMD – geographic atrophy (GA). Recently, treatments acting on the complement pathway have been approved by the United States Food and Drug Administration, with other jurisdictions such as Australia considering their approval. Furthermore, there are many more potential treatments for GA currently in clinical trials. However, unlike the treatments for neovascular AMD, where clinicians have treated virtually all patients presenting with exudation, the decision to treat those with GA will be more nuanced and individualised. Longitudinal retinal imaging will be one important asset that will help tremendously when counselling patients, as through these images, the growth pattern of the atrophy can be determined. Even without individual prior imaging history, there are other imaging clues to help predict growth rates and threats to the fovea, and hence imminent vision loss. Optometrists have a critical role in this new era where GA treatments will be available, as they are often the first to have contact with GA patients. Insightful, well-informed counselling and appropriate referral for those seeking more information on potential treatment to confirm the diagnosis and perform baseline imaging at a location likely to undertake any future treatment will ensure that appropriate patients have had the best workup to be individually managed once these treatments arrive in Australia. [ABSTRACT FROM AUTHOR]

The parafoveal area, with its high concentration of photoreceptors and fine retinal capillaries, is crucial for central vision and often exhibits early signs of pathological changes. The current adaptive optics scanning laser ophthalmoscope (AOSLO) provides an excellent tool to acquire accurate and detailed information about the parafoveal area with cellular resolution. However, limited by the scanning speed of two-dimensional scanning, the field of view (FOV) in the AOSLO system was usually less than or equal to 2∘, and the stitching for the parafoveal area required dozens of images, which was time-consuming and laborious. Unfortunately, almost half of patients are unable to obtain stitched images because of their poor fixation. To solve this problem, we integrate AO technology with the line-scan imaging method to build an adaptive optics line scanning ophthalmoscope (AOLSO) system with a larger FOV. In the AOLSO, afocal spherical mirrors in pairs are nonplanar arranged and the distance and angle between optical elements are optimized to minimize the aberrations, two cylinder lenses are orthogonally placed before the imaging sensor to stretch the point spread function (PSF) for sufficiently digitizing light energy. Captured human retinal images show the whole parafoveal area with 5 × 5 ∘ FOV, 60 Hz frame rate and cellular resolutions. Take advantage of the 5∘ FOV of the AOLSO, only 9 frames of the retina are captured with several minutes to stitch a montage image with an FOV of 9 × 9 ∘ , in which photoreceptor counting is performed within approximately 5∘ eccentricity. The AOLSO system not only provides cellular resolution but also has the capability to capture the parafoveal region in a single frame, which offers great potential for noninvasive studying of the parafoveal area. [ABSTRACT FROM AUTHOR]

Purpose: To elaborate a deep learning (DL) model for automatic prediction of late recurrence (LR) of rhegmatogenous retinal detachment (RRD) using pseudocolor and fundus autofluorescence (AF) ultra‐wide field (UWF) images obtained preoperatively and postoperatively. Materials and Methods: We retrospectively included patients >18 years who underwent either scleral buckling (SB) or pars plana vitrectomy (PPV) for primary or recurrent RRD with a post‐operative follow‐up >2 years. Records of RRD recurrence between 6 weeks and 2 years after surgery served as a ground truth for the training of the deep learning (DL) models. Four separate DL models were trained to predict LR within the 2 postoperative years (binary outputs) using, respectively, UWF preoperative and postoperative pseudocolor images and UWF preoperative and postoperative AF images. Results: A total of 412 eyes were included in the study (332 eyes treated with PPV and 80 eyes with SB). The mean follow‐up was 4.0 ± 2.1 years. The DL models based on preoperative and postoperative pseudocolor UWF imaging predicted recurrence with 85.6% (sensitivity 86.7%, specificity 85.4%) and 90.2% accuracy (sensitivity 87.0%, specificity 90.8%) in PPV‐treated eyes, and 87.0% (sensitivity 86.7%, specificity 87.0%) and 91.1% (sensitivity 88.2%, specificity 91.9%) in SB‐treated eyes, respectively. The DL models using preoperative and postoperative AF‐UWF imaging predicted recurrence with 87.6% (sensitivity 84.0% and specificity 88.3%) and 91.0% (sensitivity 88.9%, specificity 91.5%) accuracy in PPV eyes, and 86.5% (sensitivity 87.5%; specificity 86.2%) and 90.6% (sensitivity 90.0%, specificity 90.7%) in SB eyes, respectively. Among the risk factors detected with visualisation methods, potential novel ones were extensive laser retinopexy and asymmetric staphyloma. Conclusions: DL can accurately predict the LR of RRD based on UWF images (especially postoperative ones), which can help refine follow‐up strategies. Saliency maps might provide further insight into the dynamics of RRD recurrence. [ABSTRACT FROM AUTHOR]

Currently, Diabetic Retinopathy (DR) is one of the most common causes of blindness among people who suffer from diabetes for a long period. But, the existing detection models have drawbacks in classifying different stages of DR due to the similarity between some classes. To overcome this problem, a Dual Attention Mechanism with EfficientNet-B0 (DAM-EfficientNet-B0) is proposed for precise detection and classification of DR. Initially, retinal images are obtained from EyePACS and Messidor datasets and fed for preprocessing by Contrast Limited Adaptive Histogram Equalization (CLAHE) and data augmentation. After that, the preprocessed images are segmented based on a modified expectation maximization technique, while the segmented images are fed to the feature extraction process by ResNet-50. Finally, the extracted features are forwarded to the proposed classifier approach for accurate detection and classification of DR as per severity levels. The utilization of DAM helps select the relevant and important features which further leads to learning differences between classes. The experimental results show that DAM-EfficientNet-B0 attains an accuracy of 99.21% which is greater than the previous methods such as MultiStream-Deep Neural Network (MS-DNN), Concatenated Convolutional Neural Network (CCNN), and DNN-Butterfly Optimization Algorithm (DNN-BOA). [ABSTRACT FROM AUTHOR]

We propose a novel and versatile computational approach, based on hierarchical COSFIRE filters, that addresses the challenge of explainable retina and palmprint recognition for automatic person identification. Unlike traditional systems that treat these biometrics separately, our method offers a unified solution, leveraging COSFIRE filters' trainable nature for enhanced selectivity and robustness, while exhibiting explainability and resilience to decision-based black-box adversarial attack and partial matching. COSFIRE filters are trainable, in that their selectivity can be determined with a one-shot learning step. In practice, we configure a COSFIRE filter that is selective for the mutual spatial arrangement of a set of automatically selected keypoints of each retina or palmprint reference image. A query image is then processed by all COSFIRE filters and it is classified with the reference image that was used to configure the COSFIRE filter that gives the strongest similarity score. Our approach, tested on the VARIA and RIDB retina datasets and the IITD palmprint dataset, achieved state-of-the-art results, including perfect classification for retina datasets and a 97.54% accuracy for the palmprint dataset. It proved robust in partial matching tests, achieving over 94% accuracy with 80% image visibility and over 97% with 90% visibility, demonstrating effectiveness with incomplete biometric data. Furthermore, while effectively resisting a decision-based black-box adversarial attack and impervious to imperceptible adversarial images, it is only susceptible to highly perceptible adversarial images with severe noise, which pose minimal concern as they can be easily detected through histogram analysis in preprocessing. In principle, the proposed learning-free hierarchical COSFIRE filters are applicable to any application that requires the identification of certain spatial arrangements of moderately complex features, such as bifurcations and crossovers. Moreover, the selectivity of COSFIRE filters is highly intuitive; and therefore, they provide an explainable solution. [ABSTRACT FROM AUTHOR]

The utilization of retinal vessel images is prevalent in the diagnosis of numerous diseases, such as chronic vascular diseases, diabetic retinopathy, and glaucoma. Despite the U-Net model's effective performance in retinal vessel imaging, it has been observed that the U-Net model fails to adequately detect dense blood vessels and vascular bifurcation. This research proposes a retinal vessel segmentation model that is both efficient and straightforward. The study introduces two significant contributions. Firstly, the multi-scale structure is improved, and a feature circulation module is added to enhance the detection of blood vessels. Secondly, an AGM module is incorporated into the jump connection to boost the transmission of context characteristics and suppress background noise. The presented model is assessed using publicly available DRIVE and STARE datasets. The proposed method and U-Net evaluation results are quantitatively and qualitatively analyzed, alongside a comparison with current state-of-the-art algorithms. By conducting a comparative analysis, certain algorithmic indices have exhibited state-of-the-art performance. The AttMSFCU-Net model, as introduced in this study, achieves advanced or comparable accuracy levels when evaluated on both the DRIVE and STARE datasets. Of particular note is the enhanced segmentation performance of dense and bifurcation vessels. The evaluation of the proposed model on the DRIVE and STARE datasets highlights its robustness and accuracy, affirming its utility in practical applications. [ABSTRACT FROM AUTHOR]

Purpose Methods Results Conclusion To validate the performance of autonomous diabetic retinopathy (DR) grading by comparing a human grader and a self‐developed deep‐learning (DL) algorithm with gold‐standard evaluation.We included 500, 6‐field retinal images graded by an expert ophthalmologist (gold standard) according to the International Clinical Diabetic Retinopathy Disease Severity Scale as represented with DR levels 0–4 (97, 100, 100, 103, 100, respectively). Weighted kappa was calculated to measure the DR classification agreement for (1) a certified human grader without, and (2) with assistance from a DL algorithm and (3) the DL operating autonomously. Using any DR (level 0 vs. 1–4) as a cutoff, we calculated sensitivity, specificity, as well as positive and negative predictive values (PPV and NPV). Finally, we assessed lesion discrepancies between Model 3 and the gold standard.As compared to the gold standard, weighted kappa for Models 1–3 was 0.88, 0.89 and 0.72, sensitivities were 95%, 94% and 78% and specificities were 82%, 84% and 81%. Extrapolating to a real‐world DR prevalence of 23.8%, the PPV were 63%, 64% and 57% and the NPV were 98%, 98% and 92%. Discrepancies between the gold standard and Model 3 were mainly incorrect detection of artefacts (n = 49), missed microaneurysms (n = 26) and inconsistencies between the segmentation and classification (n = 51).While the autonomous DL algorithm for DR classification only performed on par with a human grader for some measures in a high‐risk population, extrapolations to a real‐world population demonstrated an excellent 92% NPV, which could make it clinically feasible to use autonomously to identify non‐DR patients. [ABSTRACT FROM AUTHOR]

An event camera is a neuromimetic sensor inspired by the human retinal imaging principle, which has the advantages of high dynamic range, high temporal resolution, and low power consumption. Due to the interference of hardware and software and other factors, the event stream output from the event camera usually contains a large amount of noise, and traditional denoising algorithms cannot be applied to the event stream. To better deal with different kinds of noise and enhance the robustness of the denoising algorithm, based on the spatio-temporal distribution characteristics of effective events and noise, an event stream noise reduction and visualization algorithm is proposed. The event stream enters fine filtering after filtering the BA noise based on spatio-temporal density. The fine filtering performs time sequence analysis on the event pixels and the neighboring pixels to filter out hot noise. The proposed visualization algorithm adaptively overlaps the events of the previous frame according to the event density difference to obtain clear and coherent event frames. We conducted denoising and visualization experiments on real scenes and public datasets, respectively, and the experiments show that our algorithm is effective in filtering noise and obtaining clear and coherent event frames under different event stream densities and noise backgrounds. [ABSTRACT FROM AUTHOR]

PurposeMethodsResultsConclusionAbetalipoproteinemia (ABL, MIM 200,100) is a rare autosomal recessive disorder caused by nonfunctional microsomal triglyceride transfer protein leading to absence of apolipoprotein B-containing lipoproteins in plasma and a retinitis pigmentosa-like fundus. The MTTP gene is expressed in retinal pigment epithelium (RPE) and ganglion cells of the human retina. Understanding ABL pathophysiology would benefit from new cellular-level clinical imaging of affected retinas.We report multimodal retinal imaging in two patients with ABL. Case 1 (67-year-old woman) exhibited a bilateral decline of vision due to choroidal neovascularization (CNV) associated with angioid streaks and calcified Bruch membrane. Optical coherence tomography were consistent with basal laminar deposits and subretinal drusenoid deposits (SDD).Case 2 (46-year-old woman) exhibited unusual hyperpigmentation at the right fovea with count-fingers vision and a relatively unremarkable left fundus with 20/30 vision. The left eye exhibited the presence of nodular drusen and SDD and the absence of macular xanthophyll pigments.We propose that mutated MTTP within the retina may contribute to ABL retinopathy in addition to systemic deficiencies of fat-soluble vitamins. This concept is supported by a new mouse model with RPE-specific MTTP deficiency and a retinal degeneration phenotype. The observed range of human pathology, including angioid streaks, underscores the need for continued monitoring in adulthood, especially for CNV, a treatable condition. [ABSTRACT FROM AUTHOR]

Purpose: To evaluate the baseline characteristics of fundus autofluorescence (FAF) in patients with submacular hemorrhage (SMH). Methods: This retrospective study included patients diagnosed with treatment-naive, foveal-involving subretinal hemorrhage (size > 2-disc diameters) of any etiology, presenting between June 2017 and June 2023. Only cases with good-quality color fundus photographs, optical coherence tomography (OCT) scans, and blue-light FAF images at baseline were included. SMH imaging characteristics were documented and correlated with treatment outcomes. A successful treatment outcome was defined as the reduction, displacement or clearance of the SMH from beneath the fovea. Results: Nineteen cases of SMH (13 males, 6 females), ranging from 14 to 85 years, were analyzed. Neovascular age-related macular degeneration (nAMD) was the most common etiology (n = 11, 58%). Baseline visual acuity ranged from 6/9 to counting fingers at ½ meter, with a median presentation time of 7 days from symptom onset (range: 1–57 days). Treatment success was observed in 13 eyes (68%). Hypoautofluoroscence on FAF was significantly associated with SMH resolution (p = 0.021). However, no association was found between treatment success and clinical hemorrhage characteristics (p = 0.222), OCT findings (p = 0.222), or specific treatments (p > 0.05). Hypoautofluoroscence on FAF was the sole predictor of treatment success, as demonstrated by Spearman's correlation (r = 0.637; p = 0.003) and linear regression analysis (p = 0.003). Conclusion: FAF, in conjunction with color fundus photography and OCT, may provide valuable insights for clinicians in formulating treatment strategies for patients with SMH. Hypoautofluoroscence on FAF was a significant predictor of successful SMH resolution in this study. [ABSTRACT FROM AUTHOR]

The increasing incidence and mortality rates of cerebrovascular disease impose a heavy burden on both patients and society. Retinal imaging techniques, such as fundus photography, optical coherence tomography, and optical coherence tomography angiography, can be used for rapid, non-invasive evaluation of cerebral microcirculation and brain function since the retina and the central nervous system share similar embryonic origin characteristics and physiological features. This article aimed to review retinal imaging biomarkers related to cerebrovascular diseases and their applications in cerebrovascular diseases (stroke, cerebral small vessel disease [CSVD], and vascular cognitive impairment [VCI]), thus providing reference for early diagnosis and prevention of cerebrovascular diseases. [ABSTRACT FROM AUTHOR]

Purpose: To develop a reliable method to generate a mouse model of branch retinal artery occlusion (BRAO) using laser‐induced thrombosis of a major artery in the mouse retina. Also, to develop a reliable method to detect retinal hypoxia as predictive biomarker for the risk of neuronal cell damage in BRAO. Methods: A reliable and reproducible model of laser‐induced BRAO was developed in mouse retina using Rose Bengal. To characterize retinal hypoxia in BRAO, pimonidazole immunostaining and HYPOX‐4 molecular imaging methods were used. Terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) was used to characterize neuronal cell damage in the BRAO retina. Expression of mRNA in retinal tissues from BRAO and age‐matched control retinas were analyzed using qRT‐PCR. Results: Occlusion of a branch retinal artery near the optic nerve head (ONH) caused a pattern of retinal tissue hypoxia covering about 12.5% of the entire retina. TUNEL‐positive cells were localized in all layers in BRAO retinal tissue cross sections. In addition, qRT‐PCR data analysis suggests that BRAO is associated with both inflammation and hypoxia. Conclusions: This study provides a reliable method for BRAO in mouse retina and demonstrates the utility of molecular imaging method to detect retinal hypoxia as predictive biomarker for the risk of neuronal cell damage in BRAO. In addition, our data suggest that BRAO retinas are associated with inflammation and also associated with hypoxia‐related neuronal cell damage. Perspectives: Imaging areas of retinal hypoxia may provide accurate diagnosis, evaluating retinal tissue injury from BRAO. [ABSTRACT FROM AUTHOR]

The segmentation of retinal blood vessels in fundus images is critical for the diagnosis and analysis of various ocular, circulatory, and neurological conditions. Accurate segmentation aids in the early detection of diseases such as diabetic retinopathy and glaucoma. Traditional automated segmentation methods often rely on extensive labeled datasets for model pre-training, which limits their generalization capacity. Recent advancements in artificial intelligence and computer vision have introduced foundational models, such as the Segment Anything Model (SAM), which demonstrate strong zero-shot segmentation performance and transferability in natural image processing. However, SAM's application to medical imaging, particularly in retinal vessel segmentation, has produced suboptimal results. This study proposes an improved approach to retinal fundus image segmentation by integrating a Generative Adversarial Network (GAN)-based data augmentation technique to enhance training data diversity. Additionally, a dynamic batch size mechanism was introduced, optimizing the loss function for mask prediction and allowing flexible control over slice selection during loss calculation. This dual enhancement aims to improve the precision of blood vessel segmentation in retinal fundus images, overcoming the limitations observed in previous applications of SAM to medical image segmentation. The proposed method demonstrates potential for advancing the accuracy of retinal vessel segmentation, providing a robust tool for clinical diagnosis. [ABSTRACT FROM AUTHOR]

The escalating global prevalence of blindness remains a pressing concern, with eye diseases representing the primary culprits behind this issue. Vision is integral to various aspects of human life, underscoring the significance of effective eye disease detection. Presently, disease detection relies largely on manual methods, which are susceptible to misdiagnosis. However, the advent of technology has paved the way for disease detection through the application of deep learning methodologies. Deep learning exhibits substantial potential in disease detection, particularly when applied to image data, as attested by its accuracy in algorithmic assessments. This research introduces a novel approach to disease detection, specifically transfer learning-based deep learning. The study seeks to evaluate and compare the performance of various models, including EfficientNetB3, DenseNet-121, VGG-16, and ResNet-152, in identifying three prevalent eye diseases: cataract, diabetic retinopathy, and glaucoma, utilizing retinal fundus image data. Extensive experimentation reveals that the DenseNet-121 model achieves the highest accuracy levels, boasting precision, recall, F1-score, and accuracy values of 96.5%, 96%, 96.25%, and 96.20%, respectively. These results demonstrate the superior performance of the employed transfer learning model, signifying its efficacy in detecting eye diseases. [ABSTRACT FROM AUTHOR]

Glaucoma is a significant eye disease that can lead to irreversible vision loss if not detected and treated early. This research focuses on developing an automated glaucoma detection system using a combination of a convolutional neural network (CNN) with the residual network 18 (ResNet18) architecture, locality sensitive hashing (LSH), and Hamming distance calculation. The CNN model is trained to extract meaningful features from retinal images, while LSH enables efficient indexing and retrieval of similar images. Hamming distance calculations are utilized to measure the dissimilarity between binary codes obtained from LSH. A dataset of 506 retinal images, consisting of 117 glaucoma images, 19 glaucoma suspect images, and 370 healthy images. The proposed glaucoma detection system achieved an average accuracy of 99.96%, sensitivity of 99.97%, and specificity of 99.94% during training, and 82.37% accuracy, 86.78% sensitivity, and 73.55% specificity during testing. Comparative analysis demonstrated its superiority over traditional methods. Further research should focus on larger datasets and explore multi-class classification for different glaucoma stages. The proposed system has potential for early glaucoma detection, facilitating timely intervention, and preventing vision loss. [ABSTRACT FROM AUTHOR]

Ophthalmologists rely heavily on retinal fundus imaging to diagnose retinal diseases. Early detection can enhance the likelihood of a cure and also prevent blindness. Retinal fundus images can be used by medical professionals to diagnose retinal conditions such as diabetic retinopathy and retinitis pigmentosa. This study proposes an automated diagnostic approach using a Deep Learning (DL) model to identify fundus images with a high prediction rate. This study aims to use multilabel classification to identify diseases in fundus images. An EfficientNet-B5-based model was trained on a fundus image dataset to classify images as normal, NPDR, and PDR. Image preprocessing was used, including conversion to RGB format, resizing to 224x224, and image filtering using the Gaussian blur algorithm. Additionally, 10-fold cross-validation was used to train and validate the proposed approach. The enhanced EfficientNet-B5 model demonstrated superior validation and training accuracy for eye disease classification compared to existing techniques, achieving 96.04% and 99.54%, respectively. This technology enables early detection and treatment of eye conditions, potentially improving patient outcomes. [ABSTRACT FROM AUTHOR]

Background Results and Conclusion Shared anatomical, histological and physiological pathways between the kidney and the eye are well documented, demonstrating that ocular manifestations serve as valuable prognostic indicators in chronic kidney disease (CKD), providing insights into disease severity and progression. Through non‐invasive imaging modalities such as retinal fundus photography, early retinal microvascular alterations indicative of CKD progression can be detected, enabling timely intervention and risk stratification. However, the conclusions drawn from the review primarily demonstrate a strong or independent association between glaucoma or retinopathy and CKD.Multiple shared pathophysiological events have been implicated in the pathogenesis in the alterations at eye and kidney including renin‐angiotensin‐aldosterone system. Patients with CKD are more likely to experience glaucoma, age‐related macular degeneration, cataracts, uremic optic neuropathy and retinopathy. To establish the role of ocular manifestations in predicting CKD progression, it is crucial to address the limitations of correlation and explore the underlying causality with further research on common disease pathogenesis. Additionally, specific methods for risk stratification based on retinal changes, the effectiveness of timely interventions, and the development of predictive tools combining ocular and renal data are of utmost importance research topics to enlighten the bidirectional causality. [ABSTRACT FROM AUTHOR]

Purpose: To investigate the changes in retinal vascular structures associated with various stages of myopia by designing automated software based on an artificial intelligence model. Methods: The study involved 1324 pediatric participants from the National Children's Medical Center in China, and 2366 high-quality retinal images and corresponding refractive parameters were obtained and analyzed. Spherical equivalent refraction (SER) degree was calculated. We proposed a data analysis model based on a combination of the Convolutional Neural Networks (CNN) model and the attention module to classify images, segment vascular structures, and measure vascular parameters, such as main angle (MA), branching angle (BA), bifurcation edge angle (BEA) and bifurcation edge coefficient (BEC). One-way ANOVA compared parameter measurements between the normal fundus, low myopia, moderate myopia, and high myopia groups. Results: The mean age was 9.85 ± 2.60 years, with an average SER of -1.49 ± 3.16D in the right eye and − 1.48 ± 3.13D in the left eye. There were 279 (12.38%) images in the normal group and 384 (16.23%) images in the high myopia group. Compared with normal fundus, the MA of fundus vessels in different myopic refractive groups was significantly reduced (P = 0.006, P = 0.004, P = 0.019, respectively), and the performance of the venous system was particularly obvious (P < 0.001). At the same time, the BEC decreased disproportionately (P < 0.001). Further analysis of fundus vascular parameters at different degrees of myopia showed that there were also significant differences in BA and branching coefficient (BC). The arterial BA value of the fundus vessel in the high myopia group was lower than that of other groups (P = 0.032, 95% confidence interval [CI], 0.22–4.86), while the venous BA values increased (P = 0.026). The BEC values of high myopia were higher than those of low and moderate myopia groups. When the loss function of our data classification model converged to 0.09, the model accuracy reached 94.19%. Conclusion: The progression of myopia is associated with a series of quantitative retinal vascular parameters, particularly the vascular angles. As the degree of myopia increases, the diversity of vascular characteristics represented by these parameters also increases. [ABSTRACT FROM AUTHOR]

Retinal imaging often falls short in image quality due to limitations in imaging conditions. Issues such as low contrast and inadequate brightness are frequently encountered. However, fundus pictures play a crucial role in diagnosing various retinal diseases within the field of ophthalmology. Nonetheless, specific ocular abnormalities and capturing environments result in low-grade fundus images, hampering the diagnostic abilities of both human experts and machines. Analyzing color fundus images to detect retinal abnormalities necessitates enhanced representation of image properties, including contrast, illumination, and precise edge points. The proposed method introduces a new technique for improving color fundus photos. The algorithm comprises three stages. Firstly, a feed-forward denoising convolutional neural network (DnCNN) removes noise. Subsequently, a contrast enhancement method, recursive separated weighted histogram equalization (RSWHE), addresses low contrast issues. Finally, adaptive Gamma correction (AGC) improves uneven luminosity. Experiments were conducted using the STARE benchmark datasets to evaluate the algorithm. The suggested algorithm's output is equated against state-of-the-art enhancement methods. Objective validation was performed using performance parameters such as NIQE, PCQI, CEIQ, MEME, and PSNR. It suggests that the algorithm has the potential to serve as an efficient method for enhancing retinal images, thereby improving diagnostic capabilities in the field of ophthalmology. [ABSTRACT FROM AUTHOR]

Diabetic Retinopathy (DR) is termed as ever-lasting retinal disorders which can cause loss of vision and even blindness in most of the cases. The procedure to classify the level of severity in DR and is complex process as lesion features are complex to examine. The task of screening needs an effectual detection technique for classifying the subtle pathologies in retina. The deep models is considered as an imperative domain treating the disease caused in eye and help the ophthalmologists to make timely decisions. A new LeNet model is developed with Chronological Rat Swarm Optimization (CRSO) for detecting DR. The first step is the capturing retinal images through fundus photography, which further undergoes pre-processing with median filter. The segmentation is then achieved wherein lesion segmentation is completed with U-Net and blood vessel segmentation is performed with Dense U-Net. From input fundus image, lesion segmented output, and blood vessel segmented output, extraction of essential feature is performed. Finally, DR detection is executed with CRSO-based LeNet. The CRSO-based LeNet attained superior performance with highest accuracy of 89%, NPV of 87.2%, PPV of 87.0%, TNR of 89.0% and TPR of 89.5%. The proposed method is useful in various applications, such as medical applications and healthcare applications. [ABSTRACT FROM AUTHOR]

Background: Optical coherence tomography (OCT) is a leading ocular imaging modality, known for delivering high-resolution volumetric morphological images. However, conventional OCT systems are limited by their narrow field-of-view (FOV) and their reliance on scattering contrast, lacking molecular specificity. Methods: To address these limitations, we developed a custom-built 105∘ ultra-widefield polarization-diversity OCT (UWF PD-OCT) system for assessing various retinal and choroidal conditions, which is particularly advantageous for visualizing peripheral retinal abnormalities. Patients with peripheral lesions or pigmentary changes were imaged using the UWF PD-OCT to evaluate the system's diagnostic capabilities. Comparisons were made with conventional swept-source OCT and other standard clinical imaging modalities to highlight the benefits of depolarization contrast for identifying pathological changes. Results: The molecular-specific contrast offered by UWF PD-OCT enhanced the detection of disease-specific features, particularly in the peripheral retina, by capturing melanin distribution and pigmentary changes in a single shot. This detailed visualization allows clinicians to monitor disease progression with greater precision, offering more accurate insights into retinal and choroidal pathologies. Conclusions: Integrating UWF PD-OCT into clinical practice represents a major advancement in ocular imaging, enabling comprehensive views of retinal pathologies that are difficult to capture with current modalities. This technology holds great potential to transform the diagnosis and management of retinal and choroidal diseases by providing unique insights into peripheral retinal abnormalities and melanin-specific changes, critical for early detection and timely intervention. [ABSTRACT FROM AUTHOR]

(200/200) Purpose: Our aim was to evaluate structural alterations of retinal arterioles due to type 1 diabetes (T1D) and/or diabetic retinopathy (DR) under AOSLO imaging. Methods: Each study eye underwent mydriasis and AOSLO imaging in a single-visit study. The instrument's arrangement of four offset aperture images provided two orthogonal split-detector images and enabled isotropic analysis of the arteriolar boundaries. For each arteriole, we calculated the wall-to-lumen ratio (WLR), mean wall thickness, and luminal and external diameters. Results: In total, we enrolled 5 (20.8%) healthy control eyes and 19 eyes of patients with T1D. The DR distribution was: four (16.7%) no-DR, nine (37.5%%) mild or moderate nonproliferative DR (NPDR), and six (25%) severe NPDR or proliferative DR. Mean wall thickness increased significantly in eyes with T1D compared to healthy controls (p = 0.0006) and in eyes with more advanced DR (p = 0.0004). The WLR was significantly higher in eyes with T1D (p = 0.002) or more severe DR (p = 0.004). There was no significant relationship between T1D status or DR severity and any of the arteriolar diameters. Conclusions: In this preliminary study, there appeared to be increases in the WLR and mean wall thickness in eyes with T1D and more severe DR than in the controls and eyes with no/less severe DR. Future studies may further elucidate the relationship between the retinal arteriolar structure and physiologic alterations in DR. [ABSTRACT FROM AUTHOR]

The difficulty of classifying retinal fundus images with one or more illnesses present or missing is known as fundus multi-lesion classification. The challenges faced by current approaches include the inability to extract comparable morphological features from images of different lesions and the inability to resolve the issue of the same lesion, which presents significant feature variances due to grading disparities. This paper proposes a multi-disease recognition network model, Fundus-DANet, based on the dilated convolution. It has two sub-modules to address the aforementioned issues: the interclass learning module (ILM) and the dilated-convolution convolutional block attention module (DA-CBAM). The DA-CBAM uses a convolutional block attention module (CBAM) and dilated convolution to extract and merge multiscale information from images. The ILM uses the channel attention mechanism to map the features to lower dimensions, facilitating exploring latent relationships between various categories. The results demonstrate that this model outperforms previous models in classifying fundus multilocular lesions in the OIA-ODIR dataset with 93% accuracy. [ABSTRACT FROM AUTHOR]

Diabetes mellitus (DM) significantly impacts ocular health, particularly the retina and choroid, leading to diabetic retinopathy (DR) and diabetic macular edema (DME). This study aimed to evaluate the effects of intravitreal dexamethasone (IVD) on the choroidal vascularity index (CVI) in order to evaluate its potential as a biomarker. This was a retrospective observational study in which 37 eyes with DME treated with IVD were included. Baseline and follow-up (2 and 4 months post-injection) comprehensive ophthalmological examinations and spectral domain optical coherence tomography (SD-OCT) were performed. The CVI, luminal area (LA), stromal area (SA), and total area (TA) were assessed. No statistically significant changes in CVI, LA, SA, and TA were observed in the overall patient sample from baseline to 2 and 4 months post-IVD injection. However, a significant CVI reduction was noted in responder patients at 4 months. Positive correlations were found between the CVI and central macular thickness (CMT) at 2 months, suggesting that higher CMT values are associated with higher CVI values. This study indicated that IVD influences choroidal circulation, as evidenced by CVI reduction in responder patients. The CVI could potentially serve as an objective biomarker for treatment response in DME. Larger, prospective studies are warranted to clarify these findings further. [ABSTRACT FROM AUTHOR]

Retinal fundus image classification is pivotal for the early detection of various eye disorders. Leveraging advancements in deep learning, this research introduces a robust framework for retinal fundus image classification. In this research paper, we introduce a novel methodology that integrates the Active Gradient Deep Convolutional Neural Network (AG-DCNN) architecture with the Red Spider Optimization (RSO) algorithm. Extensive evaluations conducted on well-established retinal image datasets such as DRIVE, STARE, CHASE DB1, HRF, DRISHTI-GS, and RFMiD illustrate substantial enhancements in classification accuracy, sensitivity, and specificity when compared to conventional approaches. The AG-DCNN with RSO exhibits superior performance and remarkable generalization abilities across diverse datasets. This novel approach not only enhances the latest advances in retinal image classification but also holds promise for early disease diagnosis and improved patient care. Through extensive examinations, our proposed method consistently outperforms established techniques, establishing itself as the benchmark in retinal fundus image classification. The integration of AG-DCNN and RSO showcases its efficacy and potential impact in advancing the field of medical image analysis. [ABSTRACT FROM AUTHOR]

PurposeMethodsResultsConclusionTo report a case of Vogt-Koyanagi-Harada (VKH)-like uveitis under treatment with Dabrafenib and Trametinib for metastatic malignant melanoma, representing the longest follow-up (49 months) with retinal multimodal imaging.Retrospective case report.A 49-year-old female with metastatic relapsing cutaneous malignant melanoma presented with blurry vision in both eyes for 1 week. She had been treated with Dabrafenib and Trametinib for 2 months. Fundus examination detected serous retinal detachments (SRDs) and hyperemic optic disks. Spectral-domain optical coherence tomography (SD-OCT) revealed SRDs, retinal pigment epithelium undulations, choroidal thickening, and loss of normal choroidal vascular architecture. The patient was diagnosed with VKH-like uveitis secondary to targeted agents since systemic investigations were unremarkable. Dabrafenib and Trametinib were discontinued, and pulse steroid treatment was started. Following the improvement of retinal and choroidal signs, the same targeted agents were restarted 6 weeks later. No recurrence of uveitis occurred during 49 months of follow-up; however, the convalescent phase findings of VKH were observed in the fundus examination. The systemic status of the patient, who is still using Dabrafenib and Trametinib, is stable.Although the mechanism is still unknown, the development of VKH-like uveitis secondary to targeted therapy may indicate successful tumor control in patients with metastatic melanoma. Providing effective immunosuppression with corticosteroids and making necessary dose modifications with a multidisciplinary approach may extend the survival of patients. [ABSTRACT FROM AUTHOR]

Persistent subretinal fluid (PSF) after scleral bucking of rhegmatogenous retinal detachment may delay recovery and affect the final visual quality, but with no effective treatment. This study firstly investigated the safety and efficacy of 577 nm yellow subthreshold micropulse laser (SML) on PSF after scleral bucking surgery. This double-masked randomized clinical trial was conducted from December 2020 to October 2022 at Chongqing Aier Eye Hospital. Participants with PSF last for 1 month after scleral bucking surgery with break closed and retina reattachment were recruitment. These participants were treated by 577 nm yellow SML or sham treatment. Funduscopy, optical coherence tomography (OCT) volume change, best corrected vision acuity (BCVA) and visual field test were evaluated for six mouths follow-up. A total of 24 participants were randomized into SML group or Sham group equally. Compared with Sham group, the OCT volume within 6 mm of macular fovea was significantly less in SML group 6 months after therapy (P = 0.048). There were no statistically significant differences of OCT volume at 1, 2 and 3 months from baseline between groups. BCVA of ETDRS letters had no statistically significant difference. Pattern Standard Deviation amelioration (P = 0.039) had statistically significance in SML group compared with Sham group. There were no complications in the 2 groups. These preliminary findings suggest that 577 nm yellow SML therapy could accelerate PSF absorption after scleral bucking surgery. Trial registration: Chinese Clinical Trial Registry No. ChiCTR2000037838, 02/09/2020, https://www.chictr.org.cn/showproj.html?proj=51885. [ABSTRACT FROM AUTHOR]

Purpose: To delineate the genotype and phenotype of RH in a Chinese cohort. Methods: A group of 51 Chinese probands with RH across 76 eyes was assembled and underwent complete retinal imaging examinations. Sanger sequencing and universal primer quantitative fluorescent multiplex–polymerase chain reaction (UPQFM-PCR) were employed for mutation detection in the coding region of the Von Hippel–Lindal (VHL) gene. For frequency calculation, our series was combined with three large cohorts of East Asian descent through a literature review. Results: The Von Hippel–Lindal (VHL) syndrome was excluded in fifteen patients (median age: 32.00 years) with unilateral solitary RH. Thirty-six patients of younger ages (median: 22.00 years, p = 0.008, Mann–Whitney test) conformed to the diagnostic criteria of the VHL syndrome, and thirty-four patients were genetically confirmed. There were four novel variants identified in the VHL gene. Codons 167, 161 and 86 exhibited a mutation occurrence of more than 5% after pooling with literature data, and the large genomic deletion demonstrated a frequency of 17.65%. The RHs were classified as "extrapapillary", "juxtapapillary" and "mixed" types in 53, 7 and 5 eyes, respectively. Almost all extrapapillary RH lesions were found in the peripheral retina. Hemangioblastomas in the central nervous system (CNS) were observed in 25 out of 31 kindreds (80.65%) with full systemic evaluation data. Conclusions: VHL-associated RH might exhibit earlier onset than non-VHL RH. Large genomic deletions were observed at a notably high frequency in the Chinese series with VHL-associated RH, which might be associated with East Asian ethnicity background. RH could potentially serve as an early indicator of CNS hemangioblastoma. [ABSTRACT FROM AUTHOR]

Achromatopsia is the most common cone dysfunction syndrome, affecting 1 in 30,000 people. It is an autosomal recessive disorder with a heterogeneous genetic background with variants reported in CNGA3, CNGB3, GNAT2, PDE6C, PDE6H, and ATF6. Up to 90% of achromatopsia patients harbour mutations in CNGA3 or CNB3, which encode for the alpha and beta subunits of the cone cyclic nucleotide-gated (CNG) channel in cone-specific phototransduction. The condition presents at birth or early infancy with poor visual acuity, nystagmus, photophobia, and colour vision loss in all axes. Multimodal retinal imaging has provided insightful information to characterise achromatopsia patients based on their genotype. There is no FDA-approved treatment for achromatopsia; however, studies have reported several preclinical gene therapies with anatomical and functional improvements reported in vivo. There are currently five gene therapy clinical trials registered for human patients at the phase I/II stage and for CNGA3 or CNGB3 causing achromatopsia. This review aims to discuss the genetics of achromatopsia, genotypic and phenotypic correlations in multimodal retinal imaging, and the developments and challenges in gene therapy clinical trials. [ABSTRACT FROM AUTHOR]

The incorporation of gold nanoparticles (GNPs) into retinal imaging signifies a notable advancement in ophthalmology, offering improved accuracy in diagnosis and patient outcomes. This review explores the synthesis and unique properties of GNPs, highlighting their adjustable surface plasmon resonance, biocompatibility, and excellent optical absorption and scattering abilities. These features make GNPs advantageous contrast agents, enhancing the precision and quality of various imaging modalities, including photoacoustic imaging, optical coherence tomography, and fluorescence imaging. This paper analyzes the unique properties and corresponding mechanisms based on the morphological features of GNPs, highlighting the potential of GNPs in retinal disease diagnosis and management. Given the limitations currently encountered in clinical applications of GNPs, the approaches and strategies to overcome these limitations are also discussed. These findings suggest that the properties and efficacy of GNPs have innovative applications in retinal disease imaging. [ABSTRACT FROM AUTHOR]

Background: This study aimed to test an artificial intelligence-based reading system (AIRS) capable of reading retinographies of type 2 diabetic (T2DM) patients and a predictive algorithm (DRPA) that predicts the risk of each patient with T2DM of developing diabetic retinopathy (DR). Methods: We tested the ability of the AIRS to read and classify 15,297 retinal photographs from our database of diabetics and 1200 retinal images taken with Messidor-2 into the different DR categories. We tested the DRPA in a sample of 40,129 T2DM patients. The results obtained by the AIRS and the DRPA were then compared with those provided by four retina specialists regarding sensitivity (S), specificity (SP), positive predictive value (PPV), negative predictive value (NPV), accuracy (ACC), and area under the curve (AUC). Results: The results of testing the AIRS for identifying referral DR (RDR) in our database were ACC = 98.6, S = 96.7, SP = 99.8, PPV = 99.0, NPV = 98.0, and AUC = 0.958, and in Messidor-2 were ACC = 96.78%, S = 94.64%, SP = 99.14%, PPV = 90.54%, NPV = 99.53%, and AUC = 0.918. The results of our DRPA when predicting the presence of any type of DR were ACC = 0.97, S = 0.89, SP = 0.98, PPV = 0.79, NPV = 0.98, and AUC = 0.92. Conclusions: The AIRS performed well when reading and classifying the retinographies of T2DM patients with RDR. The DRPA performed well in predicting the absence of DR based on some clinical variables. [ABSTRACT FROM AUTHOR]

This article reviews innovative diagnostic approaches for diabetic retinopathy as the prevalence of diabetes mellitus and its complications continue to escalate. Novel techniques focus on early disease detection. Technological innovations, such as teleophthalmology, smartphone-based photography, artificial intelligence with deep learning, or widefield photography, can enhance diagnostic accuracy and accelerate the treatment. The review highlights teleophthalmology and handheld photography as promising solutions for remote eye care. These methods revolutionize diabetic retinopathy screening, offering cost-effective and accessible solutions. However, the use of these techniques may be limited by insurance coverage in certain world regions. Ultra-widefield photography offers a comprehensive view of up to 80.0% of the retina in a single image, compared to the 34.0% coverage of the traditional seven-field imaging protocol. It allows retinal imaging without pupil dilation, especially for individuals with compromised mydriasis. However, they also have drawbacks, including high costs, artifacts from eyelashes, eyelid margins, and peripheral distortion. Recent advances in artificial intelligence and machine learning, particularly through convolutional neural networks, are revolutionizing diabetic retinopathy diagnostics, enhancing screening efficiency and accuracy. FDA-approved Artificial Intelligence-powered devices such as LumineticsCore™, EyeArt, and AEYE Diagnostic Screening demonstrate high sensitivity and specificity in diabetic retinopathy detection. While Artificial Intelligence offers the potential to improve patient outcomes and reduce treatment costs, challenges such as dataset biases, high initial costs, and cybersecurity risks must be considered to ensure safety and efficiency. Nanotechnology advancements further enhance diagnosis, offering highly branched polyethyleneimine particles with fluorescein sodium (PEI-NHAc-FS) for better fluorescein angiography or vanadium oxide-based metabolic fingerprinting for early detection. [ABSTRACT FROM AUTHOR]

Purpose: Women with hypertensive disorders of pregnancy (HDP) are at increased risk of developing premature cardiovascular disease (CVD). The mechanisms behind this are not fully understood, but microvascular alterations have been documented in retinal arterioles and venules. The aim of this study was to use non‐invasive retinal imaging to investigate the structural and functional properties of arterioles, venules and capillaries in this patient group. Methods: We examined 27 women with previous HDP and 23 controls at 3 years postpartum. The retinal microvasculature was assessed by vessel calibre measurements, retinal oximetry and optical coherence tomography angiography. Differences were analysed using non‐parametric tests and multiple regression analyses, adjusted for age and body mass index. Results: Median arteriolar oxygen saturation (SaO2; 94.2% vs. 93.0%), venular oxygen saturation (SvO2; 60.1% vs. 62.4%) and arteriovenous saturation difference (AV‐difference; 32.8% vs. 32.3%) were similar across groups. Capillary vessel density (VD; 46.2% vs. 46.3%), skeletonised VD (VSD; 21.3 vs. 21.1 mm/mm2) and vessel diameter index (21.65 vs. 21.86) were also comparable. In the HDP group, mean arterial pressure (MAP) was positively correlated with AV‐difference (R2 = 0.209) and negatively correlated with arteriolar diameter (CRAE; r2 = 0.382). Conclusions: Structural microvascular alterations appear not to be key biomarkers for CVD risk after HDP as early as 3 years postpartum in otherwise healthy women. Further studies are needed to evaluate whether such changes occur later in life. MAP was associated with AV‐difference only in the HDP group, suggesting specific mechanisms affecting functional microvascular properties in these women. [ABSTRACT FROM AUTHOR]

Introduction: This study reports our experiences with systematic retinal screening in Denmark through optometrists with access to tele-ophthalmological services before, during, and after the COVID-19 pandemic. Methods: We evaluated an optometrist-based retinal screening system with a referral option for tele-ophthalmological service by a consultant ophthalmologist within the time period of August 1, 2018 to September 30, 2023. The optometrist collected patient history, refraction, best-corrected visual acuity, intraocular pressure, basic slit-lamp examination, 4-in-1 visual field report, and retinal imaging using color fundus 45° photography. Tele-ophthalmological services were provided by consultant ophthalmologists. Within pre-defined periods of pre-COVID-19, COVID-19, and post-COVID-19, we evaluated the rate of referrals to the tele-ophthalmological service, diagnoses made, and referrals to the public healthcare system. Results: A total of 1,142,028 unique individuals, which corresponded to 19.1% of the entire population of Denmark, underwent screening by the optometrists; 50,612 (4.4%) of these individuals were referred to the tele-ophthalmological examination by consultant ophthalmologists. A referral for further ophthalmic examination, either at hospital or at an ophthalmic practice, was made for 10,300 individuals (20.4% of those referred for tele-ophthalmology, corresponding to 0.9% of the population screened). The referral rate from the screening to the tele-ophthalmological service increased from before COVID-19 (3.4%) to during COVID-19 (4.3%) and further after COVID-19 (6.4%). This increase coincided with an increasing prevalence of conditions seen in the tele-ophthalmological service. Conclusion: During a period of 5 years, 19.1% of the entire population of Denmark underwent retinal screening. This provided an adjunctive health service during a period of severe strain on the public healthcare system, while limiting the number of excessive referrals to the public healthcare system. Temporal trends illustrated an increased pattern of use of a large-scale tele-ophthalmological system. [ABSTRACT FROM AUTHOR]

This commentary article delves into the transformative role of optical coherence tomography angiography (OCTA) in diagnosing and managing a wide array of eye conditions, including diabetic retinopathy, age-related macular degeneration, retinal vein occlusions, and white dot syndromes. Developed in 2005, OCTA has emerged as a non-invasive, high-resolution imaging technique that offers advantages over traditional fluorescein angiography (FA), providing quicker and safer monitoring of ocular conditions with similar diagnostic accuracy. In diabetic retinopathy, OCTA has been instrumental in early identification of retinal changes, offering quantifiable metrics including perfused capillary density (PCD) for assessing vascular alterations. For age-related macular degeneration (AMD), OCTA has deepened our understanding of non-exudative neovascular AMD, allowing for more effective monitoring and potential earlier initiation of treatment. In cases of retinal vein occlusions, OCTA can reveal specific microvascular features and allow for depth-resolved measurements of the foveal avascular zone, providing significant prognostic implications. OCTA has also been invaluable in studying rare white dot syndromes, enabling nuanced differentiation between conditions that often present similarly. Emerging research also suggests that OCTA can have potential utility in neurodegenerative diseases like Alzheimer's, where retinal vascular patterns could offer diagnostic insights. While OCTA is revolutionizing ophthalmic care, further clinical trials and standardization are needed for its broader adoption into clinical practice. [ABSTRACT FROM AUTHOR]

Glaucoma, a leading cause of permanent blindness worldwide, necessitates early detection to prevent vision loss, a task that is challenging and time-consuming when performed manually. This study proposes an automatic glaucoma detection method on enhanced retinal images using deep learning. The system analyzes retinal images, generating masks for the optic disc and optic cup, and providing a classification for glaucoma diagnosis. We employ a U-Net architecture with a pretrained residual neural network (ResNet34) for segmentation and an EfficientNetB0 for classification. The proposed framework is tested on publicly available datasets, including ORIGA, REFUGE, RIM-ONE DL, and HRF. Our work evaluated the U-Net model with five pretrained backbones (ResNet34, ResNet50, VGG19, DenseNet121, and EfficientNetB0) and examined preprocessing effects. We optimized model training with limited data using transfer learning and data augmentation techniques. The segmentation model achieves a mean intersection over union (mIoU) value of 0.98. The classification model shows remarkable performance with 99.9% training and 100% testing accuracy on ORIGA, 99.9% training and 99% testing accuracy on RIM-ONE DL, and 98% training and 100% testing accuracy on HRF. The proposed model outperforms related works and demonstrates potential for accurate glaucoma classification and detection tasks. [ABSTRACT FROM AUTHOR]

Schizophrenia (SZ) and bipolar disorder (BD) are characterized by major symptomatic, cognitive, and neuroanatomical changes. Recent studies have used optical coherence tomography (OCT) to investigate retinal changes in SZ and BD, but their unique and shared changes require further evaluation. Articles were identified using PubMed and Google Scholar. 39 studies met the inclusion criteria. Diagnostic groups were proband (SZ/BD combined), SZ, BD, and healthy control (HC) eyes. Meta-analyses utilized fixed and random effects models when appropriate, and publication bias was corrected using trim-and-fill analysis ("meta" package in R). Results are reported as standardized mean differences with 95% CIs. Data from 3145 patient eyes (1956 SZ, 1189 BD) and 3135 HC eyes were included. Studies identified thinning of the peripapillary retinal nerve fiber layer (pRNFL, overall and in 2 subregions), m-Retina (overall and all subregions), mGCL-IPL, mIPL, and mRPE in SZ patients. BD showed thinning of the pRNFL (overall and in each subregion), pGCC, and macular Retina (in 5 subregions), but no changes in thickness or volume for the total retina. Neither SZ nor BD patients demonstrated significant changes in the fovea, mRNFL, mGCL, mGCC, mINL, mOPL, mONL, or choroid thicknesses. Moderating effects of age, illness duration, and smoking on retinal structures were identified. This meta-analysis builds upon previous literature in this field by incorporating recent OCT studies and examining both peripapillary and macular retinal regions with respect to psychotic disorders. Overall, this meta-analysis demonstrated both peripapillary and macular structural retinal abnormalities in people with SZ or BD compared with HCs. [ABSTRACT FROM AUTHOR]

Cardiovascular diseases (CVDs) are a major cause of mortality globally, emphasizing the need for early detection and effective risk assessment to improve patient outcomes. Advances in oculomics, which utilize the relationship between retinal microvascular changes and systemic vascular health, offer a promising non-invasive approach to assessing CVD risk. Retinal fundus imaging and optical coherence tomography/angiography (OCT/OCTA) provides critical information for early diagnosis, with retinal vascular parameters such as vessel caliber, tortuosity, and branching patterns identified as key biomarkers. Given the large volume of data generated during routine eye exams, there is a growing need for automated tools to aid in diagnosis and risk prediction. The study demonstrates that AI-driven analysis of retinal images can accurately predict cardiovascular risk factors, cardiovascular events, and metabolic diseases, surpassing traditional diagnostic methods in some cases. These models achieved area under the curve (AUC) values ranging from 0.71 to 0.87, sensitivity between 71% and 89%, and specificity between 40% and 70%, surpassing traditional diagnostic methods in some cases. This approach highlights the potential of retinal imaging as a key component in personalized medicine, enabling more precise risk assessment and earlier intervention. It not only aids in detecting vascular abnormalities that may precede cardiovascular events but also offers a scalable, non-invasive, and cost-effective solution for widespread screening. However, the article also emphasizes the need for further research to standardize imaging protocols and validate the clinical utility of these biomarkers across different populations. By integrating oculomics into routine clinical practice, healthcare providers could significantly enhance early detection and management of systemic diseases, ultimately improving patient outcomes. Fundus image analysis thus represents a valuable tool in the future of precision medicine and cardiovascular health management. [ABSTRACT FROM AUTHOR]

Background: Age-related macular degeneration (AMD) is a complex eye disorder with an environmental and genetic origin, affecting millions worldwide. The study aims to explore the association between retinal morphology and the polygenic risk score (PRS) for AMD using fundus images and deep learning techniques. Methods: The study used and pre-processed 23,654 fundus images from 332 subjects (235 patients with AMD and 97 controls), ultimately selecting 558 high-quality images for analysis. The fine-tuned DenseNet121 deep learning model was employed to estimate PRS from single fundus images. After training, deep features were extracted, fused, and used in machine learning regression models to estimate PRS for each subject. The Grad-CAM technique was applied to examine the relationship between areas of increased model activity and the retina's morphological features specific to AMD. Results: Using the hybrid approach improved the results obtained by DenseNet121 in 5-fold cross-validation. The final evaluation metrics for all predictions from the best model from each fold are MAE = 0.74, MSE = 0.85, RMSE = 0.92, R2 = 0.18, MAPE = 2.41. Grad-CAM heatmap evaluation showed that the model decisions rely on lesion area, focusing mostly on the presence of drusen. The proposed approach was also shown to be sensitive to artifacts present in the image. Conclusions: The findings indicate an association between fundus images and AMD PRS, suggesting that deep learning models may effectively estimate genetic risk for AMD from retinal images, potentially aiding in early detection and personalized treatment strategies. [ABSTRACT FROM AUTHOR]

Background: Early diagnosis and treatment of diabetic eye disease (DED) improve prognosis and lessen the possibility of permanent vision loss. Screening of retinal fundus images is a significant process widely employed for diagnosing patients with DED or other eye problems. However, considerable time and effort are required to detect these images manually. Methods: Deep learning approaches in machine learning have attained superior performance for the binary classification of healthy and pathological retinal fundus images. In contrast, multi-class retinal eye disease classification is still a difficult task. Therefore, a two-phase transfer learning approach is developed in this research for automated classification and segmentation of multi-class DED pathologies. Results: In the first step, a Modified ResNet-50 model pre-trained on the ImageNet dataset was transferred and learned to classify normal diabetic macular edema (DME), diabetic retinopathy, glaucoma, and cataracts. In the second step, the defective region of multiple eye diseases is segmented using the transfer learning-based DenseUNet model. From the publicly accessible dataset, the suggested model is assessed using several retinal fundus images. Our proposed model for multi-class classification achieves a maximum specificity of 99.73%, a sensitivity of 99.54%, and an accuracy of 99.67%. [ABSTRACT FROM AUTHOR]

Abstract Self-supervised learning has become the cornerstone of building generalizable and transferable artificial intelligence systems in medical imaging. In particular, contrastive representation learning techniques trained on large multi-modal datasets have demonstrated impressive capabilities of producing highly transferable representations for different downstream tasks. In ophthalmology, large multi-modal datasets are abundantly available and conveniently accessible as modern retinal imaging scanners acquire both 2D fundus images and 3D optical coherence tomography (OCT) scans to assess the eye. In this context, we introduce a novel multi-modal contrastive learning-based pipeline to facilitate learning joint representations for the two retinal imaging modalities. After self-supervised pre-training on 153,306 scan pairs, we show that such a pre-training framework can provide both a retrieval system and encoders that produce comprehensive OCT and fundus image representations that generalize well for various downstream tasks on three independent external datasets, explicitly focusing on clinically pertinent prediction tasks. In addition, we show that interchanging OCT with lower-cost fundus imaging can preserve the predictive power of the trained models.