Imaging and Biomarkers in Diabetic Macular Edema and Diabetic Retinopathy

A deep learning approach to hard exudates detection and disorganization of retinal inner layers identification on OCT images

The purpose of the study was to detect Hard Exudates (HE) and classify Disorganization of Retinal Inner Layers (DRIL) implementing a Deep Learning (DL) system on optical coherence tomography (OCT) images of eyes with diabetic macular edema (DME). We collected a dataset composed of 442 OCT images on which we annotated 6847 HE and the presence of DRIL. A complex operational pipeline was defined to implement data cleaning and image transformations, and train two DL models. The state-of-the-art neural network architectures (Yolov7, ConvNeXt, RegNetX) and advanced techniques were exploited to aggregate the results (Ensemble learning, Edge detection) and obtain a final model. The DL approach reached good performance in detecting HE and classifying DRIL. Regarding HE detection the model got an AP@0.5 score equal to 34.4% with Precision of 48.7% and Recall of 43.1%; while for DRIL classification an Accuracy of 91.1% with Sensitivity and Specificity both of 91.1% and AUC and AUPR values equal to 91% were obtained. The P-value was lower than 0.05 and the Kappa coefficient was 0.82. The DL models proved to be able to identify HE and DRIL in eyes with DME with a very good accuracy and all the metrics calculated confirmed the system performance. Our DL approach demonstrated to be a good candidate as a supporting tool for ophthalmologists in OCT images analysis.

Generative adversarial networks synthetic optical coherence tomography images as an education tool for image diagnosis of macular diseases: a randomized trial

Purpose

This study aimed to evaluate the effectiveness of generative adversarial networks (GANs) in creating synthetic OCT images as an educational tool for teaching image diagnosis of macular diseases to medical students and ophthalmic residents.

Methods

In this randomized trial, 20 fifth-year medical students and 20 ophthalmic residents were enrolled and randomly assigned (1:1 allocation) into Group real OCT and Group GANs OCT. All participants had a pretest to assess their educational background, followed by a 30-min smartphone-based education program using GANs or real OCT images for macular disease recognition training. Two additional tests were scheduled: one 5 min after the training to assess short-term performance, and another 1 week later to assess long-term performance. Scores and time consumption were recorded and compared. After all the tests, participants completed an anonymous subjective questionnaire.

Results

Group GANs OCT scores increased from 80.0 (46.0 to 85.5) to 92.0 (81.0 to 95.5) 5 min after training (p < 0.001) and 92.30 ± 5.36 1 week after training (p < 0.001). Similarly, Group real OCT scores increased from 66.00 ± 19.52 to 92.90 ± 5.71 (p < 0.001), respectively. When compared between two groups, no statistically significant difference was found in test scores, score improvements, or time consumption. After training, medical students had a significantly higher score improvement than residents (p < 0.001).

Conclusion

The education tool using synthetic OCT images had a similar educational ability compared to that using real OCT images, which improved the interpretation ability of ophthalmic residents and medical students in both short-term and long-term performances. The smartphone-based educational tool could be widely promoted for educational applications.Clinical trial registration: https://www.chictr.org.cn, Chinese Clinical Trial Registry [No. ChiCTR 2100053195].

Ergodic speckle contrast optical coherence tomography velocimetry of rapid blood flow

Visualizing a 3D blood flow velocity field through noninvasive imaging is crucial for analyzing hemodynamic mechanisms in areas prone to disorders. However, traditional correlation-based optical coherence tomography (OCT) velocimetry techniques have a maximum measurable flow velocity depending on the A-line rate. We presented the ergodic speckle contrast OCT (ESCOCT) to break the bottleneck in measuring the rapid blood flow velocity. It achieved a measurement of blood flow velocity ranging from 9.5 to 280 mm/s using a 100 kHz swept-source (SS) OCT based on 100 A-repeats scanning mode. Addressing the non-ergodic problem of temporal OCT signals by integrating more consecutive A-scans, ESCOCT can enable the estimation for lower velocity flows by increasing A-repeats. ESCOCT provided a wide dynamic range with no upper limit on measuring blood flow velocity with an adequate signal-to-noise ratio and improved the sensitivity and accuracy of the hemodynamic assessment.

Segmentation of retinal microaneurysms in fluorescein fundus angiography images by a novel three-step model

Introduction

Microaneurysms serve as early signs of diabetic retinopathy, and their accurate detection is critical for effective treatment. Due to their low contrast and similarity to retinal vessels, distinguishing microaneurysms from background noise and retinal vessels in fluorescein fundus angiography (FFA) images poses a significant challenge.

Methods

We present a model for automatic detection of microaneurysms. FFA images were pre-processed using Top-hat transformation, Gray-stretching, and Gaussian filter techniques to eliminate noise. The candidate microaneurysms were coarsely segmented using an improved matched filter algorithm. Real microaneurysms were segmented by a morphological strategy. To evaluate the segmentation performance, our proposed model was compared against other models, including Otsu's method, Region Growing, Global Threshold, Matched Filter, Fuzzy c-means, and K-means, using both self-constructed and publicly available datasets. Performance metrics such as accuracy, sensitivity, specificity, positive predictive value, and intersection-over-union were calculated.

Results

The proposed model outperforms other models in terms of accuracy, sensitivity, specificity, positive predictive value, and intersection-over-union. The segmentation results obtained with our model closely align with benchmark standard. Our model demonstrates significant advantages for microaneurysm segmentation in FFA images and holds promise for clinical application in the diagnosis of diabetic retinopathy.

Conclusion

The proposed model offers a robust and accurate approach to microaneurysm detection, outperforming existing methods and demonstrating potential for clinical application in the effective treatment of diabetic retinopathy.

OCT Intensity of the Region between Outer Retina Band 2 and Band 3 as a Biomarker for Retinal Degeneration and Therapy

Optical coherence tomography (OCT) is widely used to probe retinal structure and function. This study investigated the outer retina band (ORB) pattern and reflective intensity for the region between bands 2 and 3 (Dip) in three mouse models of inherited retinal degeneration (Rs1KO, TTLL5KO, RPE65KO) and in human AMD patients from the A2A database. OCT images were manually graded, and reflectivity signals were used to calculate the Dip ratio. Qualitative analyses demonstrated the progressive merging band 2 and band 3 in all three mouse models, leading to a reduction in the Dip ratio compared to wildtype (WT) controls. Gene replacement therapy in Rs1KO mice reverted the ORB pattern to one resembling WT and increased the Dip ratio. The degree of anatomical rescue in these mice was highly correlated with level of transgenic RS1 expression and with the restoration of ERG b-wave amplitudes. While the inner retinal cavity was significantly enlarged in dark-adapted Rs1KO mice, the Dip ratio was not altered. A reduction of the Dip ratio was also detected in AMD patients compared with healthy controls and was also positively correlated with AMD severity on the AMD score. We propose that the ORB and Dip ratio can be used as non-invasive early biomarkers for retina health, which can be used to probe therapeutic gene expression and to evaluate the effectiveness of therapy.

Peripapillary hyperreflective ovoid mass-like structures: multimodal imaging and associated diseases

Growing evidence has demonstrated that peripapillary hyperreflective ovoid mass-like structures (PHOMS) are novel structures rather than a subtype of optic disc drusen. They correspond to the laterally bulging herniation of optic nerve fibers and are believed to be the marker of axoplasmic stasis. PHOMS present in a broad spectrum of diseases, including optic disc drusen, tilted disc syndrome, papilloedema, multiple sclerosis, non-arteritic anterior ischemic optic neuropathy, optic neuritis, Leber hereditary optic neuropathy, and so on. We focus on the multimodal imaging features, pathophysiological mechanisms of PHOMS, and their association with multiple diseases and healthy people in this review to deepen the ophthalmologists' understanding of PHOMS. Additionally, we provide some new directions for future research.

Assessment of area and structural irregularity of retinal layers in diabetic retinopathy using machine learning and image processing techniques

Diabetes retinopathy prevention necessitates early detection, monitoring, and treatment. Non-invasive optical coherence tomography (OCT) shows structural changes in the retinal layer. OCT image evaluation necessitates retinal layer segmentation. The ability of our automated retinal layer segmentation to distinguish between normal, non-proliferative (NPDR), and proliferative diabetic retinopathy (PDR) was investigated in this study using quantifiable biomarkers such as retina layer smoothness index (SI) and area (S) in horizontal and vertical OCT images for each zone (fovea, superior, inferior, nasal, and temporal). This research includes 84 eyes from 57 individuals. The study shows a significant difference in the Area (S) of inner nuclear layer (INL) and outer nuclear layer (ONL) in the horizontal foveal zone across the three groups (p < 0.001). In the horizontal scan, there is a significant difference in the smoothness index (SI) of the inner plexiform layer (IPL) and the upper border of the outer plexiform layer (OPL) among three groups (p < 0.05). There is also a significant difference in the area (S) of the OPL in the foveal zone among the three groups (p = 0.003). The area (S) of the INL in the foveal region of horizontal slabs performed best for distinguishing diabetic patients (NPDR and PDR) from normal individuals, with an accuracy of 87.6%. The smoothness index (SI) of IPL in the nasal zone of horizontal foveal slabs was the most accurate at 97.2% in distinguishing PDR from NPDR. The smoothness index of the top border of the OPL in the nasal zone of horizontal slabs was 84.1% accurate in distinguishing NPDR from PDR. Smoothness index of IPL in the temporal zone of horizontal slabs was 89.8% accurate in identifying NPDR from PDR patients. In conclusion, optical coherence tomography can assess the smoothness index and irregularity of the inner and outer plexiform layers, particularly in the nasal and temporal regions of horizontal foveal slabs, to distinguish non-proliferative from proliferative diabetic retinopathy. The evolution of diabetic retinopathy throughout severity levels and its effects on retinal layer irregularity need more study.

Optical Coherence Tomography Angiography as a Diagnostic Tool for Diabetic Retinopathy

Diabetic retinopathy (DR) is the most common microvascular complication of diabetes mellitus, leading to visual impairment if left untreated. This review discusses the use of optical coherence tomography angiography (OCTA) as a diagnostic tool for the early detection and management of DR. OCTA is a fast, non-invasive, non-contact test that enables the detailed visualisation of the macular microvasculature in different plexuses. OCTA offers several advantages over fundus fluorescein angiography (FFA), notably offering quantitative data. OCTA is not without limitations, including the requirement for careful interpretation of artefacts and the limited region of interest that can be captured currently. We explore how OCTA has been instrumental in detecting early microvascular changes that precede clinical signs of DR. We also discuss the application of OCTA in the diagnosis and management of various stages of DR, including non-proliferative diabetic retinopathy (NPDR), proliferative diabetic retinopathy (PDR), diabetic macular oedema (DMO), diabetic macular ischaemia (DMI), and pre-diabetes. Finally, we discuss the future role of OCTA and how it may be used to enhance the clinical outcomes of DR.

Predicting visual outcomes following anti-vascular endothelial growth factor treatment for diabetic macular edema

PURPOSE

To assess the utility of pre-defined imaging biomarkers on optical coherence tomography (OCT) and OCT angiography (OCTA) in patients with diabetic macular edema (DME) following anti-vascular endothelial growth factor (anti-VEGF) therapy in determining visual and anatomical outcomes.

METHODS

In this prospective, non-randomized, and interventional study, 17 patients with treatment-naive DME were included. OCT biomarkers [size/reflectivity of cysts, disorganization of retinal inner layers, integrity of ellipsoid zone or external limiting membrane, subfoveal serous retinal detachment, hyper-reflective foci (HRF)] and OCTA [vascular density (VD), foveal avascular zone (FAZ), and total micro-aneurysms in superficial capillary plexus and deep capillary plexus (DCP)] were analyzed at baseline and after three monthly intravitreal anti-VEGF injections. Response was defined as a decrease of 10% or more in central macular thickness from the baseline after three injections.

RESULTS

13/17 (76.47%) patients were categorized as responders to anti-VEGF therapy. Non-responders had significantly greater hyper-reflectivity of cysts (P = 0.015), larger cystic spaces (P = 0.023), and an increased number of HRF (P = 0.04) at baseline. On OCTA, non-responders showed larger FAZ in DCP (1.35 ± 0.21 versus 1.14 ± 0.28 mm2) (P = 0.042) and lower VD (61.17 ± 0.45 versus 62.73 ± 3.32) in DCP at baseline. At 3 months, the VD increased in responders (63.10 ± 3.42) compared to a decrease in non-responders (60.82 ± 1.13) (P = 0.032).

CONCLUSIONS

Non-responders show a higher number of micro-aneurysms, larger FAZ, and lower VD in the DCP on OCTA and higher cyst hyper-reflectivity and HRF and larger cystic spaces on OCT imaging.

Automatic interpretation and clinical evaluation for fundus fluorescein angiography images of diabetic retinopathy patients by deep learning

BACKGROUND/AIMS

Fundus fluorescein angiography (FFA) is an important technique to evaluate diabetic retinopathy (DR) and other retinal diseases. The interpretation of FFA images is complex and time-consuming, and the ability of diagnosis is uneven among different ophthalmologists. The aim of the study is to develop a clinically usable multilevel classification deep learning model for FFA images, including prediagnosis assessment and lesion classification.

METHODS

A total of 15 599 FFA images of 1558 eyes from 845 patients diagnosed with DR were collected and annotated. Three convolutional neural network (CNN) models were trained to generate the label of image quality, location, laterality of eye, phase and five lesions. Performance of the models was evaluated by accuracy, F-1 score, the area under the curve and human-machine comparison. The images with false positive and false negative results were analysed in detail.

RESULTS

Compared with LeNet-5 and VGG16, ResNet18 got the best result, achieving an accuracy of 80.79%-93.34% for prediagnosis assessment and an accuracy of 63.67%-88.88% for lesion detection. The human-machine comparison showed that the CNN had similar accuracy with junior ophthalmologists. The false positive and false negative analysis indicated a direction of improvement.

CONCLUSION

This is the first study to do automated standardised labelling on FFA images. Our model is able to be applied in clinical practice, and will make great contributions to the development of intelligent diagnosis of FFA images.

Evaluation of the efficacy of anti-vascular endothelial growth factors in diabetic macular edema with retinal inner and outer layers disorganization

AIMS

This study aims to compare the effectiveness of treatment between anti-vascular endothelial growth factor (anti-VEGF) agents in diabetic macular edema (DME) patients with disorganization of retinal inner layers (DRIL). Epiretinal membrane, serous macular detachment, ellipsoid zone (EZ) disorder, external limiting membrane (ELM) disorder, and hyperreflective foci were also examined.

METHODS

Patients treated for DME and also had DRIL were included in the study. The study design was retrospective and cross-sectional. The complete ophthalmologic records and imaging were scanned at the beginning, 3rd-month, 6th-month, and 12th-month follow-up, and the treatments administered were recorded. Anti-VEGF agents administered to the patients were examined in three groups: bevacizumab, ranibizumab, and aflibercept.

RESULT

A total of 141 eyes of 100 patients were included in our study. One hundred and fifteen eyes (81.6%) had a BCVA of 0, 5, or less at the beginning. There was no statistically significant difference between the three groups regarding initial BCVA and CMT and the change in BCVA and CMT at the beginning and the 12th month (p > 0.05). There was a negative correlation between EZ and ELM disorders in patients and the change in BCVA at 12 months (r: 0.45 p < 0.001, r: 0.32 p < 0.001, respectively). The number of injections over five was positively correlated with the change in CMT but not with BCVA (r: - 2.35 p = 0.005 and r: 0.147 p = 0.082, respectively).

CONCLUSIONS

No statistically significant difference was found between anti-VEGF agents when treating DME patients with DRIL. In addition, we have shown that anatomically better results were obtained in those who had five or more injections, although not in terms of BCVA.

Longitudinal Changes of Parafoveal Vessel Density in Diabetic Patients without Clinical Retinopathy Using Optical Coherence Tomography Angiography

PURPOSE

The purpose of this study was to identify the rate of parafoveal vessel density (VD) changes associated with the progression from non-diabetic retinopathy (NDR) to early stages of DR over a year.

METHODS

This longitudinal cohort study enrolled diabetic patients from the Guangzhou community in China. The patients with NDR at baseline were included and underwent comprehensive examinations at baseline and after 1 year. A commercial OCTA device (Triton Plus, Topcon, Tokyo, Japan) was employed to quantify the parafoveal VD in the superficial and deep capillary plexuses. The rates of change in parafoveal VD over time in the incident DR and NDR groups were compared after a year.

RESULTS

A total of 448 NDR patients were included in the study. Among them, 382 (83.2%) were stable and 66 (14.4%) developed incident DR during the 1-year follow-up. The average parafoveal VD in the superficial capillary plexus (SCP) reduced significantly more quickly in the incident DR group than in the NDR group (-1.95 ± 0.45%/year vs. -0.45 ± 0.19/year, p = 0.002). The VD reduction rate for the deep capillary plexus (DCP) was not significantly different for the groups (p = 0.156).

CONCLUSIONS

The incident DR group experienced a significantly faster reduction in parafoveal VD in the SCP compared with the stable group. Our findings further provide supporting evidence that parafoveal VD in the SCP may be used as an early indicator of the pre-clinical stages of DR.

The Present and Future of Artificial Intelligence-Based Medical Image in Diabetes Mellitus: Focus on Analytical Methods and Limitations of Clinical Use

Artificial intelligence (AI)-based diagnostic technology using medical images can be used to increase examination accessibility and support clinical decision-making for screening and diagnosis. To determine a machine learning algorithm for diabetes complications, a literature review of studies using medical image-based AI technology was conducted using the National Library of Medicine PubMed, and the Excerpta Medica databases. Lists of studies using diabetes diagnostic images and AI as keywords were combined. In total, 227 appropriate studies were selected. Diabetic retinopathy studies using the AI model were the most frequent (85.0%, 193/227 cases), followed by diabetic foot (7.9%, 18/227 cases) and diabetic neuropathy (2.7%, 6/227 cases). The studies used open datasets (42.3%, 96/227 cases) or directly constructed data from fundoscopy or optical coherence tomography (57.7%, 131/227 cases). Major limitations in AI-based detection of diabetes complications using medical images were the lack of datasets (36.1%, 82/227 cases) and severity misclassification (26.4%, 60/227 cases). Although it remains difficult to use and fully trust AI-based imaging analysis technology clinically, it reduces clinicians' time and labor, and the expectations from its decision-support roles are high. Various data collection and synthesis data technology developments according to the disease severity are required to solve data imbalance.

Supervised Contrastive Learning with Angular Margin for the Detection and Grading of Diabetic Retinopathy

Many researchers have realized the intelligent medical diagnosis of diabetic retinopathy (DR) from fundus images by using deep learning methods, including supervised contrastive learning (SupCon). However, although SupCon brings label information into the calculation of contrastive learning, it does not distinguish between augmented positives and same-label positives. As a result, we propose the concept of Angular Margin and incorporate it into SupCon to address this issue. To demonstrate the effectiveness of our strategy, we tested it on two datasets for the detection and grading of DR. To align with previous work, Accuracy, Precision, Recall, F1, and AUC were selected as evaluation metrics. Moreover, we also chose alignment and uniformity to verify the effect of representation learning and UMAP (Uniform Manifold Approximation and Projection) to visualize fundus image embeddings. In summary, DR detection achieved state-of-the-art results across all metrics, with Accuracy = 98.91, Precision = 98.93, Recall = 98.90, F1 = 98.91, and AUC = 99.80. The grading also attained state-of-the-art results in terms of Accuracy and AUC, which were 85.61 and 93.97, respectively. The experimental results demonstrate that Angular Margin is an excellent intelligent medical diagnostic algorithm, performing well in both DR detection and grading tasks.

The Disorganization of Retinal Inner Layers Is Correlated to Müller Cells Impairment in Diabetic Macular Edema: An Imaging and Omics Study

The disorganization of retinal inner layers (DRIL) is an optical coherence tomography (OCT) biomarker strictly associated with visual outcomes in patients with diabetic macular edema (DME) whose pathophysiology is still unclear. The aim of this study was to characterize in vivo, using retinal imaging and liquid biopsy, DRIL in eyes with DME. This was an observational cross-sectional study. Patients affected by center-involved DME were enrolled. All patients underwent spectral domain optical coherence tomography (SD-OCT) and proteomic analysis of aqueous humor (AH). The presence of DRIL at OCT was analyzed by two masked retinal experts. Fifty-seven biochemical biomarkers were analyzed from AH samples. Nineteen eyes of nineteen DME patients were enrolled. DRIL was present in 10 patients (52.63%). No statistically significant difference was found between DME eyes with and without DRIL, considering the AH concentration of all the analyzed biomarkers except for glial fibrillary acidic protein (GFAP), a biomarker of Müller cells dysfunction (p = 0.02). In conclusion, DRIL, in DME eyes, seems to strictly depend on a major dysfunction of Müller cells, explaining its role not only as imaging biomarker, but also as visual function Müller cells-related parameter.

Association Between Contrast Sensitivity and Central Subfield Thickness in Center-Involving Diabetic Macular Edema

Purpose: To evaluate the association between contrast sensitivity (CS) and central subfield thickness (CST) in diabetic macular edema (DME). Methods: This prospectively recruited, cross-sectional study included eyes with DME evaluated from November 2018 to March 2021. CST was measured using spectral-domain optical coherence tomography on the same day as CS testing. Only eyes with center-involving DME (CST >305 µm for women; >320 µm for men) were included. CS was evaluated using the quantitative CS function (qCSF) test. Outcomes included visual acuity (VA) and the following qCSF metrics: area under the log CS function, contrast acuity (CA), and CS thresholds at 1 to 18 cycles per degree (cpd). Pearson correlation and mixed-effects regression analyses were performed. Results: The cohort included 52 eyes of 43 patients. Pearson correlation analysis showed a stronger association between CST and CS thresholds at 6 cpd (r = -0.422, P = 0.002) than CST and VA (r = 0.293, P = 0.035). Mixed-effects univariate and multivariate regression analyses showed significant associations between CST and CA (β = -0.001, P = .030), CS at 6 cpd (β = -0.002, P = .008), and CS at 12 cpd (β = -0.001, P = .049) but no significant associations between CST and VA. Among the visual function metrics, the effect size of CST was largest on CS at 6 cpd (βStandardized = -0.37, P = .008). Conclusions: In patients with DME, CS may be more strongly associated with CST than VA. Including CS as an adjunct visual function outcome measure in eyes with DME may prove clinically valuable.

A novel model of subretinal edema induced by DL-alpha aminoadipic acid

In this study we described a new model of subretinal edema induced by single intraocular injection of DL-alpha-aminoadipic acid (DLAAA) that can be employed to study the mechanism of retinal edema and test the efficacy or potential toxicity of treatments. The progression of subretinal edema was evaluated by fundus photography, fluorescein angiography and optical coherence tomography for up to 4 weeks following DLAAA injection. The VEGF, IL-6, TNF-α, Occludin, ZO-1, AQP4, Kir4.1, GFAP and GS levels were examined in DLAAA models by immunostaining, immumohistochemical staining and Western blot. Additionally, bulk RNA-seq was used to detect the mechanism involved in DLAAA-induced retinal Müller cellular injuries. In vivo and vitro assays were further conducted to confirm the sequencing results. Subretinal edema was successfully induced by DLAAA in New Zealand White rabbits (1.29 mg/eye) and C57BL/6 mice (50 or 100 μg/eye). Our results demonstrated that the disruption of blood-retinal-barrier, including vascular hyperpermeability, inflammation, and Müller cell dysfunction of fluid clearance, was involved in subretinal edema formation in the model. Bulk RNA-seq and in vitro studies indicated the activation of p38 MAPK signaling pathway in DLAAA models. This DLAAA-induced subretinal edema model can be used for mechanistic studies or drug screening.

OCT and OCT Angiography Update: Clinical Application to Age-Related Macular Degeneration, Central Serous Chorioretinopathy, Macular Telangiectasia, and Diabetic Retinopathy

Similar to ultrasound adapting soundwaves to depict the inner structures and tissues, optical coherence tomography (OCT) utilizes low coherence light waves to assess characteristics in the eye. Compared to the previous gold standard diagnostic imaging fluorescein angiography, OCT is a noninvasive imaging modality that generates images of ocular tissues at a rapid speed. Two commonly used iterations of OCT include spectral-domain (SD) and swept-source (SS). Each comes with different wavelengths and tissue penetration capacities. OCT angiography (OCTA) is a functional extension of the OCT. It generates a large number of pixels to capture the tissue and underlying blood flow. This allows OCTA to measure ischemia and demarcation of the vasculature in a wide range of conditions. This review focused on the study of four commonly encountered diseases involving the retina including age-related macular degeneration (AMD), diabetic retinopathy (DR), central serous chorioretinopathy (CSC), and macular telangiectasia (MacTel). Modern imaging techniques including SD-OCT, TD-OCT, SS-OCT, and OCTA assist with understanding the disease pathogenesis and natural history of disease progression, in addition to routine diagnosis and management in the clinical setting. Finally, this review compares each imaging technique's limitations and potential refinements.

Quantitative approaches in multimodal fundus imaging: State of the art and future perspectives

When it first appeared, multimodal fundus imaging revolutionized the diagnostic workup and provided extremely useful new insights into the pathogenesis of fundus diseases. The recent addition of quantitative approaches has further expanded the amount of information that can be obtained. In spite of the growing interest in advanced quantitative metrics, the scientific community has not reached a stable consensus on repeatable, standardized quantitative techniques to process and analyze the images. Furthermore, imaging artifacts may considerably affect the processing and interpretation of quantitative data, potentially affecting their reliability. The aim of this survey is to provide a comprehensive summary of the main multimodal imaging techniques, covering their limitations as well as their strengths. We also offer a thorough analysis of current quantitative imaging metrics, looking into their technical features, limitations, and interpretation. In addition, we describe the main imaging artifacts and their potential impact on imaging quality and reliability. The prospect of increasing reliance on artificial intelligence-based analyses suggests there is a need to develop more sophisticated quantitative metrics and to improve imaging technologies, incorporating clear, standardized, post-processing procedures. These measures are becoming urgent if these analyses are to cross the threshold from a research context to real-life clinical practice.

An automated diabetic retinopathy disorders detection model based on pretrained MobileNetv2 and nested patch division using fundus images

Aim: Fundus images are very important to diagnose some ophthalmologic disorders. Hence, fundus images have become a very important data source for machine-learning society. Our primary goal is to propose a new automated disorder classification model for diabetic retinopathy (DR) using the strength of deep learning. In this model, our proposed model suggests a treatment technique using fundus images. Material and Method: In this research, a new dataset was acquired and this dataset contains 1365 Fundus Fluorescein Angiography images with five classes. To detect these disorders automatically, we proposed a transfer learning-based feature engineering model. This feature engineering model uses pretrained MobileNetv2 and nested patch division to extract deep and exemplar features. The neighborhood component analysis (NCA) feature selection function has been applied to choose the top features. k nearest neighbors (kNN) classification function has been used to get results and we used 10-fold cross-validation (CV) to validate the results. Results: The proposed MobileNetv2 and nested patch-based image classification model attained 87.40% classification accuracy on the collected dataset. Conclusions: The calculated 87.40% classification accuracy for five classes has been demonstrated high classification accuracy of the proposed deep feature engineering model

Characterization of the Retinal Microvasculature and FAZ Changes in Ischemic Stroke and Its Different Types

Purpose

This study aimed to assess morphological changes in the retinal microvasculature and foveal avascular zone (FAZ) in patients with ischemic stroke and its different subtypes.

Methods

Thirty-three patients with ischemic stroke (14 with nonlacunar infarction and 19 with lacunar infarction) and 27 control participants were enrolled in this study. Based on optical coherence tomography angiography (OCTA), three vascular parameters, including vascular area density, vascular fractal dimension (VFD), and vascular orientation distribution (VOD), and four FAZ-related parameters, including FAZ area, FAZ axis ratio (FAR), FAZ circularity (FC), and FAZ roundness, in the superficial capillary plexus (SCP) and deep capillary plexus (DCP) were extracted and analyzed.

Results

Logistic regression results showed that worse best-corrected visual acuity (odds ratio [OR], 0.21), higher FAR (OR, 2.77) and lower FC (OR, 0.36) of the DCP were associated with ischemic stroke. Furthermore, lower VOD of the SCP was significantly associated with lacunar infarction compared with nonlacunar infarction.

Conclusions

Our study shows that the FAR and FC of the DCP may be potential biomarkers of ischemic stroke. Moreover, we demonstrated that OCT showed specific damage patterns in retinal microvascular and macular morphology in different subtypes of ischemic stroke.

Translational Relevance

This work lays the foundation for the pathophysiological characteristics of cerebrovascular diseases assisted by retinal imaging and artificial intelligence.

AOSLO-net: A Deep Learning-Based Method for Automatic Segmentation of Retinal Microaneurysms From Adaptive Optics Scanning Laser Ophthalmoscopy Images

Purpose

Accurate segmentation of microaneurysms (MAs) from adaptive optics scanning laser ophthalmoscopy (AOSLO) images is crucial for identifying MA morphologies and assessing the hemodynamics inside the MAs. Herein, we introduce AOSLO-net to perform automatic MA segmentation from AOSLO images of diabetic retinas.

Method

AOSLO-net is composed of a deep neural network based on UNet with a pretrained EfficientNet as the encoder. We have designed customized preprocessing and postprocessing policies for AOSLO images, including generation of multichannel images, de-noising, contrast enhancement, ensemble and union of model predictions, to optimize the MA segmentation. AOSLO-net is trained and tested using 87 MAs imaged from 28 eyes of 20 subjects with varying severity of diabetic retinopathy (DR), which is the largest available AOSLO dataset for MA detection. To avoid the overfitting in the model training process, we augment the training data by flipping, rotating, scaling the original image to increase the diversity of data available for model training.

Results

The validity of the model is demonstrated by the good agreement between the predictions of AOSLO-net and the MA masks generated by ophthalmologists and skillful trainees on 87 patient-specific MA images. Our results show that AOSLO-net outperforms the state-of-the-art segmentation model (nnUNet) both in accuracy (e.g., intersection over union and Dice scores), as well as computational cost.

Conclusions

We demonstrate that AOSLO-net provides high-quality of MA segmentation from AOSLO images that enables correct MA morphological classification.

Translational Relevance

As the first attempt to automatically segment retinal MAs from AOSLO images, AOSLO-net could facilitate the pathological study of DR and help ophthalmologists make disease prognoses.

The Role of Intravitreal Corticosteroids in the Treatment of DME: Predictive OCT Biomarkers

This work aims to summarize predictive biomarkers to guide treatment choice in DME. Intravitreal anti-VEGF is considered the gold standard treatment for centers involving DME, while intravitreal steroid treatment has been established as a second-line treatment in DME. However, more than 1/3 of the patients do not adequately respond to anti-VEGF treatment despite up to 4-weekly injections. Not surprisingly, insufficient response to anti-VEGF therapy has been linked to low-normal VEGF levels in the serum and aqueous humor. These patients may well benefit from an early switch to intravitreal steroid treatment. In these patients, morphological biomarkers visible in OCT may predict treatment response and guide treatment decisions. Namely, the presence of a large amount of retinal and choroidal hyperreflective foci, disruption of the outer retinal layers and other signs of chronicity such as intraretinal cysts extending into the outer retina and a lower choroidal vascular index are all signs suggestive of a favorable treatment response of steroids compared to anti-VEGF. This paper summarizes predictive biomarkers in DME in order to assist individual treatment decisions in DME. These markers will help to identify DME patients who may benefit from primary dexamethasone treatment or an early switch.

Treatment responses for branch retinal vein occlusion predicted by semi-automated fluorescein angiography quantification

Backgrounds

Branch retinal vein occlusion (BRVO) is one of the most important causes of visual loss in retinal vascular diseases. The aim of this study is to predict the treatment response of anti-vascular endothelial growth factor (anti-VEGF) therapy in BRVO using semi-automated quantified fluorescein angiography (FA) features.

Methods

This retrospective case-control study enrolled patients with BRVO who are receiving anti-VEGF therapy and have been followed up for > 1 year. Those receiving < 5 anti-VEGF injections in the first year were classified as the responsive group, while those receiving ≥5 injections were the refractory group. The FA images were subjected to semi-automated pre-processing. Fluorescein leakages at the 5-min image were represented by mean gray value over parafoveal and perifoveal regions. FA leakages and central retinal thickness (CRT) on optical coherence tomography (OCT) were used for predicting the treatment response and compared using area under receiver operating characteristic curve (AUC).

Results

Eighty-nine patients (56 males, 33 females, mean age 62.5 ± 10.9 years) with BRVO were enrolled. Of the 89 eyes, 47 (53%) were in the responsive group and 42 (47%) were in the refractory group. The refractory group had a significantly higher number of anti-VEGF injections in the first year (5.9 ± 1.6 versus 2.4 ± 1.2, p < 0.001) when compared with that of the responsive group. It had thicker pre-treatment CRT (p = 0.011), post-treatment best CRT (p < 0.001) and CRT at 1-year (p < 0.001). It also had a higher mean gray value over the parafoveal (p < 0.001) and the perifoveal (p < 0.001) regions. The mean gray value over perifoveal (AUC 0.846) and parafovel (AUC 0.818) had significantly larger AUC than that of the pre-treatment OCT (AUC 0.653; p = 0.005 and p = 0.016, respectively) when predicting treatment response.

Conclusion

The refractory group had a more severe fluorescein leakage over the parafoveal and the perifoveal regions than the responsive group had. Semi-automated quantified FA leakage can be used as a biomarker for the prediction of anti-VEGF treatment response in macular edema due to BRVO.

Expression of Long Non-Coding RNA (lncRNA) SNHG5 in Patients with Refractory Diabetic Macular Edema and Its Regulatory Mechanism

BACKGROUND The aim of this study was to assess use of lncRNAs as biomarkers in serum and aqueous humor of patients with diabetic macular edema (DME). MATERIAL AND METHODS Optical coherence tomography and fundus photography were used to analyze the retinal features of the patients. RT-qPCR was used to analyze the differential expression of lncRNA snhg5 in patients who have idiopathic macular hole (MH), DME, or refractory DME. The relationship between SNHG5 and the clinical characteristics of the patients was analyzed. The effect of SNHG5 on the hyperplasia and apoptosis of human retino-microvascular endothelial cells (HRMECs) and its mechanism were analyzed in vitro. RESULTS Patients with idiopathic MH developed retinal nerve epithelium rupture and retinal fundus thickening, and patients with DME or refractory DME showed significant macular edema with hemorrhaging. The refractory DME patients improved after treatment but still showed significant macular edema and multiple laser scarring. SNHG5 expression was not only low in the atrial fluid and plasma in DME patients, but also lower in the refractory DME group compared to the idiopathic MH patients. SNHG5 expression in the aqueous humor and plasma was negatively correlated with disease duration, body mass index, and levels of fasting blood glucose, glycated hemoglobin, proteinuria, and glycosuria. In the in vitro experiments, SNHG5 expression was significantly downregulated in high glucose-induced HMECs. After SNHG5 overexpression, cell proliferation, angiogenesis, and VEGF-A protein levels were distinctly downregulated. CONCLUSIONS SNHG5 correlates with the development of DME and is a potential target for therapy.

A Contemporary Insight of Metabolomics Approach for Type 1 Diabetes: Potential for Novel Diagnostic Targets

High-throughput omics has been widely applied in metabolic disease, type 1 diabetes (T1D) was one of the most typical diseases. Effective prevention and early diagnosis are very important because of infancy and persistent characteristics of T1D. The occurrence and development of T1D is a chronic and continuous process, in which the production of autoantibodies (ie serum transformation) occupies the central position. Metabolomics can evaluate the metabolic characteristics of serum before seroconversion, the changes with age and T1D complications. And the addition of natural drug metabolomics is more conducive to the systematic and comprehensive diagnosis and treatment of T1D. This paper reviewed the metabolic changes and main pathogenesis from pre-diagnosis to treatment in T1D. The metabolic spectrum of significant abnormal energy and glucose-related metabolic pathway, down-regulation of lipid metabolism and up-regulation of some antioxidant pathways has appeared before seroconversion, indicating that the body has been in the dual state of disease progression and disease resistance before T1D onset. Some metabolites (such as methionine) are closely related to age, and the types of autoantibodies produced are age-specific. Some metabolites may jointly predict DN with eGFR, and metabolomics can further contribute to the pathogenesis based on the correlation between DN and DR. Many natural drug components have been proved to act on abnormal metabolic pathways of T1D and have a positive impact on some metabolic levels, which is very important for further finding therapeutic targets and developing new drugs with small side effects. Metabolomics can provide auxiliary value for the diagnosis of T1D and provide a new direction to reveal the pathogenesis of T1D and find new therapeutic targets. The development of T1D metabolomics shows that high-throughput research methods are expected to be introduced into clinical practice.

Changes in and the association of retinal blood perfusion and retinal nerves in diabetic patients without retinopathy

OBJECTIVE

To explore intraretinal blood flow perfusion and nerve changes, as well as the correlation between them, in diabetic patients without diabetic retinopathy (NDR).

METHOD

Eighty-six NDR patients (86 eyes) who attended the ophthalmology clinic between December 2019 and December 2021 were included. Sixty-four eyes of 64 healthy examined controls in the same period were selected as the control group. The patients underwent routine ophthalmological examination, optical coherence tomography (OCT) and OCT angiography.

RESULTS

The average thickness, minimum thickness and thickness of each quadrant except for the superior temporal quadrant of the ganglion cell-inner plexiform layer (GCIPL) in the macular area of the affected eyes in the NDR group were lower than that of the tested eyes in the control group (P < 0.05). The average retinal nerve fibre layer (RNFL) thickness of the NDR group and the superior, inferior and nasal quadrants around the optic disc of the affected eyes in the NDR group were lower compared with the tested eyes in the control group (P < 0.001, P = 0.003, P = 0.001, P = 0.009). The mean vessel length density in the parafoveal and perifoveal areas in the NDR group was positively associated with the mean GCIPL thickness in the macular area (ρ = 0.265, ρ = 0.257 and P < 0.001). No blood flow perfusion parameters in the NDR group were correlated with the RNFL thickness of the corresponding quadrant around the optic disc (P > 0.05).

CONCLUSION

In diabetic patients without diabetic retinopathy, the superficial retinal vessel density in the macular area positively correlated with GCIPL thickness, and the superficial retinal vessel density around the optic disc was not correlated with RNFL thickness.

Secreted Protein Acidic and Rich in Cysteine Mediates the Development and Progression of Diabetic Retinopathy

BACKGROUNDS

Diabetic retinopathy (DR) is one of the most severe microvascular complications of diabetes mellitus (DM). Secreted protein acidic and rich in cysteine (SPARC) has been found to play an important role in many diseases, but its role and mechanism in DR remain unknown.

METHODS

We studied the role of SPARC and integrin β1 in vascular pathophysiology and identified potential therapeutic translation. The SPARC levels were tested in human serum and vitreous by ELISA assay, and then the Gene Expression Omnibus (GEO) dataset was used to understand the key role of the target gene in DR. In human retinal capillary endothelial cells (HRCECs), we analyzed the mRNA and protein level by RT-PCR, immunohistochemistry, and Western blotting. The cell apoptosis, cell viability, and angiogenesis were analyzed by flow cytometry, CCK-8, and tube formation.

RESULTS

In this study, we investigated the role of SPARC in the development and progression of human DR and high glucose-induced HRCEC cells and found that the SPARC-ITGB1 signaling pathway mimics early molecular and advanced neurovascular pathophysiology complications of DR. The result revealed that DR patients have a high-level SPARC expression in serum and vitreous. Knockdown of SPARC could decrease the expressions of inflammatory factors and VEGFR, inhibit cell apoptosis and angiogenesis, and increase cell viability by regulating integrin β1 in HRCECs.

CONCLUSION

SPARC promotes diabetic retinopathy via the regulation of integrin β1. The results of this study can provide a potential therapeutic application for the treatment of DR.

Metabolomics of various samples advancing biomarker discovery and pathogenesis elucidation for diabetic retinopathy

Diabetic retinopathy (DR) is a universal microvascular complication of diabetes mellitus (DM), which is the main reason for global sight damage/loss in middle-aged and/or older people. Current clinical analyses, like hemoglobin A1c, possess some importance as prognostic indicators for DR severity, but no effective circulating biomarkers are used for DR in the clinic currently, and studies on the latent pathophysiology remain lacking. Recent developments in omics, especially metabolomics, continue to disclose novel potential biomarkers in several fields, including but not limited to DR. Therefore, based on the overview of metabolomics, we reviewed progress in analytical technology of metabolomics, the prominent roles and the current status of biomarkers in DR, and the update of potential biomarkers in various DR-related samples via metabolomics, including tear as well as vitreous humor, aqueous humor, retina, plasma, serum, cerebrospinal fluid, urine, and feces. In this review, we underscored the in-depth analysis and elucidation of the common biomarkers in different biological samples based on integrated results, namely, alanine, lactate, and glutamine. Alanine may participate in and regulate glucose metabolism through stimulating N-methyl-D-aspartate receptors and subsequently suppressing insulin secretion, which is the potential pathogenesis of DR. Abnormal lactate could cause extensive oxidative stress and neuroinflammation, eventually leading to retinal hypoxia and metabolic dysfunction; on the other hand, high-level lactate may damage the structure and function of the retinal endothelial cell barrier via the G protein-coupled receptor 81. Abnormal glutamine indicates a disturbance of glutamate recycling, which may affect the activation of Müller cells and proliferation via the PPP1CA-YAP-GS-Gln-mTORC1 pathway.

Effect of optical coherence tomography and angiography sampling rate towards diabetic retinopathy severity classification

Optical coherence tomography (OCT) and OCT angiography (OCT-A) may benefit the screening of diabetic retinopathy (DR). This study investigated the effect of laterally subsampling OCT/OCT-A en face scans by up to a factor of 8 when using deep neural networks for automated referable DR classification. There was no significant difference in the classification performance across all evaluation metrics when subsampling up to a factor of 3, and only minimal differences up to a factor of 8. Our findings suggest that OCT/OCT-A can reduce the number of samples (and hence the acquisition time) for a volume for a given field of view on the retina that is acquired for rDR classification.

[Diabetic Macular Edema]

Diabetic macular edema (DME) is a chronic retinal disease, which requires intensive clinical monitoring. Within the last ten years the intravitreal anti-VEGF (vascular endothelial growth factor) therapy has become the standard of care to improve and stabilize vision in patients with centre involving DME. Long-acting intravitreal corticosteroids can achieve similar visual results with fewer injection rates. Because of steroid-induced side effects (progression of cataract, glaucoma) these drugs are regarded as second-line medication. Since the introduction of anti-VEGF-medication the focal laser photocoagulation is no longer considered as first-line therapy for DME. However, a focal laser treatment can sometimes be a possible alternative in specific situations. In patients with proliferative diabetic retinopathy and DME, the intravitreal anti-VEGF therapy is approved for both conditions. In ischemic maculopathy the functional outcome is restricted. For the indication of anti-VEGF-treatment for DME with accompanying central ischemia not only visual acuity and optical coherence tomography parameters should be considered, the amount of ischemia seen on fluorescein angiography should also be taken into account. In tractional macular edema due to epiretinal membranes and vitreomacular adhesions a pars-plana vitrectomy with membrane peeling is indicated.

Diabetisches Makulaödem

Das diabetische Makulaödem (DMÖ) ist für den Großteil der Visusminderungen bei Diabetikern verantwortlich. Eine Reduktion der Sehschärfe durch ein DMÖ kann schon im milden nicht-proliferativen Stadium der diabetischen Retinopathien (DR) auftreten 1. Der chronische Verlauf der diabetischen Makulopathie erfordert ein intensives klinisches Monitoring über einen langen Zeitraum. Dieser Beitrag gibt einen Überblick über das diagnostische und therapeutische Management beim Vorliegen eines DMÖ.

Importance of Identifying Novel Biomarkers of Microvascular Damage in Type 1 Diabetes

Microvascular complications of type 1 diabetes, which primarily include diabetic kidney disease, retinopathy, and neuropathy, are characterized by damage to the microvasculature of the kidney, retina, and neurons. The pathogenesis of these complications is multifactorial, and several pathways are implicated. These complications are often silent during their early stages, and once symptoms develop, there might be little to be done to cure them. Thus, there is a strong need for novel biomarkers to identify individuals at risk of microvascular complications at an early stage and guide the implementation of new therapeutic options for preventing their development and progression. Recent advancements in proteomics, metabolomics, and other 'omics' have led to the identification of several potential biomarkers of microvascular complications. However, biomarker discovery has met several challenges and, up to now, there are no new biomarkers that have been implemented into clinical practice. This highlights the need for further work in this area to move towards better diagnostic and prognostic approaches.

Metabolomics in Diabetic Retinopathy: A Systematic Review

Purpose

Diabetic retinopathy (DR), a common microvascular complication of diabetes, is the leading cause of acquired blindness in the working-age population. Individuals with diabetes still develop DR despite appropriate glycemic and blood pressure control, highlighting the pressing need to identify useful biomarkers for risk stratification. The purpose of this review is to systematically summarize potential metabolic biomarkers and pathways of DR, which could facilitate developing an understanding of the disease mechanisms, as well as new therapeutic measures.

Methods

We searched PubMed and Web of Science for relevant metabolomics studies on humans published before September 30, 2020. Information regarding authors, title, publication date, study subjects, analytical platforms, methods of statistical analysis, biological samples, directions of change of potential metabolic biomarkers, and predictive values of metabolic biomarker panels was extracted, and the quality of the studies was assessed. Pathway analysis, including enrichment analysis and topology analysis, was derived from integrating differential metabolites using MetaboAnalyst 3.0, based on the Kyoto Encyclopedia of Genes and Genomes and Human Metabolome Database.

Results

We found nine studies focused on the identification of potential biomarkers. Repeatedly identified metabolites including l-glutamine, l-lactic acid, pyruvic acid, acetic acid, l-glutamic acid, d-glucose, l-alanine, l-threonine, citrulline, l-lysine, and succinic acid were found to be potential biomarkers of DR. It was observed that l-glutamine and citrulline changed in all biological samples. Dysregulation of metabolic pathways involved amino acid and energy metabolism.

Conclusions

This review summarizes potential biomarkers and metabolic pathways, providing insights into new pathogenic pathways for this microvascular complication of diabetes.

Foveal Microvascular Integrity Association With Anti-VEGF Treatment Response for Diabetic Macular Edema

Purpose

To investigate the association between foveal microvascular integrity and anti-vascular endothelial growth factor (VEGF) treatment response for diabetic macular edema (DME).

Methods

This retrospective study enrolled 58 eyes (from 45 patients) with DME. Treatment strategy was three to five monthly anti-VEGF injections followed by a PRN protocol. Treatment with an intravitreal corticosteroid would be considered for persistent DME after five consecutive anti-VEGF injections. Eyes achieving a treatment-free interval ≥ four months within two years were classified into the good clinical course group (group 1). Eyes with frequent recurrent edema (treatment-free interval < four months) or requiring an intravitreal corticosteroid within two years were classified into the suboptimal clinical course group (group 2). Foveal microvascular integrity was evaluated by two continuous variables, that is, vessel density (%) within a width of 300 µm around the foveal avascular zone (FD-300) on optical coherence tomography angiography (OCTA) and perifoveal leakage (area %) on fluorescein angiography (FA).

Results

There were 37 eyes in group 1 and 21 eyes in group 2. FD-300 (odds ratio 0.733, 95% CI 0.620–0.867, P < 0.001) and perifoveal leakage (odds ratio 1.064, 95% CI 1.007–1.124, P = 0.027) were significantly associated with suboptimal clinical course. Area under curve (AUC) was 0.820 for FD-300 and 0.723 for perifoveal leakage in predicting clinical course. FD-300 was negatively correlated with perifoveal leakage (coefficient = −0.325, P = 0.014).

Conclusions

Compromised foveal microvascular integrity, represented by lower FD-300 and more severe perifoveal fluorescein leakage, was associated with suboptimal clinical course in anti-VEGF treatment for DME. A negative correlation between FD-300 and perifoveal leakage existed.

Challenges in Diabetic Macular Edema Management: An Expert Consensus Report

Purpose

This paper aimed to present daily-practice recommendations for the management of diabetic macular edema (DME) patients based on available scientific evidence and the clinical experience of the consensus panel.

Methods

A group of Spanish retina experts agreed to discuss different aspects related with the clinical management of DME patients.

Results

Panel was mainly focused on therapeutic objectives in DME management; definition terms; and role of biomarkers as prognostic and predictive factors to intravitreal treatment response. The panel recommends to start DME treatment as soon as possible in those eyes with a visual acuity less than 20/25 (always according to the retina unit capacity). Naïve patient was defined, in a strict manner, as a patient who, up to that moment, had never received any treatment. A refractory DME patient may be defined as the one who did not achieve a complete resolution of the disease, regardless of the treatment administered. Different optical coherence tomography biomarkers, such as disorganization of the retinal inner layers, hyperreflective dots, and cysts, have been identified as prognostic factors.

Conclusion

This document has sought to lay down a set of recommendations and to identify key issues that may be useful for the daily management of DME patients.

Deep learning for diabetic retinopathy detection and classification based on fundus images: A review

Diabetic Retinopathy is a retina disease caused by diabetes mellitus and it is the leading cause of blindness globally. Early detection and treatment are necessary in order to delay or avoid vision deterioration and vision loss. To that end, many artificial-intelligence-powered methods have been proposed by the research community for the detection and classification of diabetic retinopathy on fundus retina images. This review article provides a thorough analysis of the use of deep learning methods at the various steps of the diabetic retinopathy detection pipeline based on fundus images. We discuss several aspects of that pipeline, ranging from the datasets that are widely used by the research community, the preprocessing techniques employed and how these accelerate and improve the models' performance, to the development of such deep learning models for the diagnosis and grading of the disease as well as the localization of the disease's lesions. We also discuss certain models that have been applied in real clinical settings. Finally, we conclude with some important insights and provide future research directions.

A Multimodal Imaging-Based Deep Learning Model for Detecting Treatment-Requiring Retinal Vascular Diseases: Model Development and Validation Study

BACKGROUND

Retinal vascular diseases, including diabetic macular edema (DME), neovascular age-related macular degeneration (nAMD), myopic choroidal neovascularization (mCNV), and branch and central retinal vein occlusion (BRVO/CRVO), are considered vision-threatening eye diseases. However, accurate diagnosis depends on multimodal imaging and the expertise of retinal ophthalmologists.

OBJECTIVE

The aim of this study was to develop a deep learning model to detect treatment-requiring retinal vascular diseases using multimodal imaging.

METHODS

This retrospective study enrolled participants with multimodal ophthalmic imaging data from 3 hospitals in Taiwan from 2013 to 2019. Eye-related images were used, including those obtained through retinal fundus photography, optical coherence tomography (OCT), and fluorescein angiography with or without indocyanine green angiography (FA/ICGA). A deep learning model was constructed for detecting DME, nAMD, mCNV, BRVO, and CRVO and identifying treatment-requiring diseases. Model performance was evaluated and is presented as the area under the curve (AUC) for each receiver operating characteristic curve.

RESULTS

A total of 2992 eyes of 2185 patients were studied, with 239, 1209, 1008, 211, 189, and 136 eyes in the control, DME, nAMD, mCNV, BRVO, and CRVO groups, respectively. Among them, 1898 eyes required treatment. The eyes were divided into training, validation, and testing groups in a 5:1:1 ratio. In total, 5117 retinal fundus photos, 9316 OCT images, and 20,922 FA/ICGA images were used. The AUCs for detecting mCNV, DME, nAMD, BRVO, and CRVO were 0.996, 0.995, 0.990, 0.959, and 0.988, respectively. The AUC for detecting treatment-requiring diseases was 0.969. From the heat maps, we observed that the model could identify retinal vascular diseases.

CONCLUSIONS

Our study developed a deep learning model to detect retinal diseases using multimodal ophthalmic imaging. Furthermore, the model demonstrated good performance in detecting treatment-requiring retinal diseases.

Vitreoretinal Society of India practice pattern survey 2020: Medical retina

Purpose:

The aim of this study was to present the outcomes of the Vitreo-retinal Society of India (VRSI) Practice Pattern Survey 2020 in medical retina.

Methods:

An online survey of members of VRSI was conducted in April 2020 regarding their practice-patterns on varied medical and surgical retina topics concerning imaging and management approach. The results were evaluated by two independent experts in this field and compared with the evidence and other practice patterns in the world.

Results:

A total of 107 VRSI members participated in the online survey. Responses were obtained on management of wide-ranging chorioretinal disorders such as Central Serous Chorioretinopathy (CSCR), Polypoidal Choroidal Vasculopathy (PCV), Neovascular age related macular degeneration (n-AMD), Retinal Vein Occlusions (RVO), and Diabetic Retinopathy (DR). Participants were also surveyed regarding their attitudes and perceptions about anti-VEGF practice patterns and role of imaging in their current practice. Each of the survey question responses were then compared to contemporary literature, including evidence-based guidelines, randomized controlled trials (RCTs), real-world evidence and analogous international surveys. Comprehensive analysis related to this has been put forward in the article.

Conclusion:

This survey represents the contemporary practice patterns amongst vitreoretinal specialists in India. The survey results are vital for fellow practitioners to understand the 'standard of care’ practice in medical retina. This will guide them to devise the best possible individualized treatment strategy for most favorable clinical outcomes.

Artificial intelligence velocimetry and microaneurysm-on-a-chip for three-dimensional analysis of blood flow in physiology and disease

Understanding the mechanics of blood flow is necessary for developing insights into mechanisms of physiology and vascular diseases in microcirculation. Given the limitations of technologies available for assessing in vivo flow fields, in vitro methods based on traditional microfluidic platforms have been developed to mimic physiological conditions. However, existing methods lack the capability to provide accurate assessment of these flow fields, particularly in vessels with complex geometries. Conventional approaches to quantify flow fields rely either on analyzing only visual images or on enforcing underlying physics without considering visualization data, which could compromise accuracy of predictions. Here, we present artificial-intelligence velocimetry (AIV) to quantify velocity and stress fields of blood flow by integrating the imaging data with underlying physics using physics-informed neural networks. We demonstrate the capability of AIV by quantifying hemodynamics in microchannels designed to mimic saccular-shaped microaneurysms (microaneurysm-on-a-chip, or MAOAC), which signify common manifestations of diabetic retinopathy, a leading cause of vision loss from blood-vessel damage in the retina in diabetic patients. We show that AIV can, without any a priori knowledge of the inlet and outlet boundary conditions, infer the two-dimensional (2D) flow fields from a sequence of 2D images of blood flow in MAOAC, but also can infer three-dimensional (3D) flow fields using only 2D images, thanks to the encoded physics laws. AIV provides a unique paradigm that seamlessly integrates images, experimental data, and underlying physics using neural networks to automatically analyze experimental data and infer key hemodynamic indicators that assess vascular injury.

Large-cube 30° × 25° optical coherence tomography in diabetic macular edema

Background

To evaluate the contribution of large-cube 30° × 25° optical coherence tomography (OCT) in the characterization of diabetic macular edema (DME) by assessing its extent and the presence of additional retinal edemas and to evaluate the factors that influenced their occurrence.

Methods

This retrospective study enrolled patients with diabetes who presented with retinal edema detected by horizontal large-cube 30° × 25° (8.7 × 7.3 mm) OCT. Two individualized areas were selected from the thickness map: the area within the 6-mm Early Treatment of Diabetic Retinopathy Study (ETDRS) grid, and that outside the ETDRS grid. Retinal edemas located within the ETDRS grid were designated as “main DME” and those located outside the ETDRS grid were designated as “peripheral retinal edemas.” For each area, OCT features were assessed while the extent of the main DME and the presence of peripheral retinal oedema were analysed in the area outside the ETDRS grid. Finally, part of included eyes was followed by the same protocol, of which a part benefited from intravitreal injections.

Results

Peripheral events were detected outside the ETDRS area in 279 eyes (74.4%) of the 375 eyes of the 218 patients included in this study: an extension of the main DME outside ETDRS grid in 177 eyes (47.2%) and/or the presence of peripheral retinal edemas in 207 eyes (55.2%). The analysis of associations between main DME and peripheral retinal edemas patterns did not find an association for retinal cyst localization (P = 0.42) while a week association was found fort cyst size (Cramer’s V = 0.188, p = 0.028). Nevertheless, a moderate association was found for the presence of microaneurysms (Cramer’s V = 0.247, p < 0.001) and strong association for hard exudates (Cramer’s V = 0.386, p < 0.001), The binary logistic regression analysis retained the following influencing factors of the occurrence of peripheral events: advanced DR stage (Odds ratio OR = 2.19, p = 0.03), diffuse DME (OR = 7.76, p < 0.001) and its location in outer fields (OR = 7.09, p = 0.006). Likewise, the extension of the main DME outside the ETDRS area in was influenced by the same factors in addition to CMT (OR = 0.98, p = 0.004) while the presence of peripheral retinal edema was influenced by the same factors except the outer location of the Main DME. Finally, from the 94 eyes treated by intravitreal injections, extension of the main DME outside the ETDRS grid was detected in 54 eyes (56.44%) at baseline visit and still remained detectable in 37 eyes (39.36%) after treatment initiation.

Conclusions

Large-cube 30° × 25° OCT allowed for more precise assessment of DME extension and better detection of retinal thickening mainly in the advanced stages of diabetic retinopathy with significant DME whether at the baseline visit or during follow-up. The combination of this protocol with a wider ETDRS grid would enhance DME detection and topography.

Hyperreflective foci in predicting the treatment outcomes of diabetic macular oedema after anti-vascular endothelial growth factor therapy

This retrospective study evaluated the association of hyperreflective foci (HRF) with treatment response in diabetic macular oedema (DME) after anti-vascular endothelial growth factor (VEGF) therapy. The medical records, including of ophthalmologic examinations and optical coherence tomography (OCT) images, of 106 patients with DME treated with either intravitreal ranibizumab or aflibercept were reviewed. The correlations between best-corrected visual acuity (BCVA) changes and HRF along with other OCT biomarkers were analysed. The mean logMAR BCVA improved from 0.696 to 0.461 after an average of 6.2 injections in 1 year under real-world conditions. Greater visual-acuity gain was noted in patients with a greater number of HRF in the outer retina at baseline (p = 0.037), along with other factors such as poor baseline vision (p < 0.001), absence of epiretinal membrane (p = 0.048), and presence of subretinal fluid at baseline (p = 0.001). The number of HRF after treatment was correlated with the presence of hard exudate (p < 0.001) and baseline haemoglobin A1C (p = 0.001). Patients with proliferative diabetic retinopathy had greater HRF reduction after treatment (p = 0.018). The number of HRF in the outer retina, in addition to other baseline OCT biomarkers, could be used to predict the treatment response in DME after anti-VEGF treatment.

Early Visual Functional Outcomes and Morphological Responses to Anti-Vascular Growth Factor Therapy in Diabetic Macular Oedema Using Optical Coherence Tomography Angiography

Purpose

Diabetic macular oedema (DME) is a vision-threatening complication of diabetic retinopathy. Spectral domain optical coherence tomography angiography (OCTA) is useful for assessing DME. We performed serial OCTA measurements before and after 3 monthly intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections to understand its relationship with best corrected visual acuity (BCVA), central macular thickness (CMT) and vascular indices.

Methods

OCTA assessment consisting of 6 × 6 mm scans centred on fovea and en-face retinal angiograms were obtained. Scans were done at baseline and at 1 month following each of the 3 intravitreal injections of anti-VEGF (ranibizumab, Lucentis^®).

Results

Our prospective study included 24 eyes of 24 subjects, aged 63.0 ± 5.0 years. There was a mean gain of 0.07 LogMAR (3 letters) and reduction of CMT, vessel density (VD) and perfusion density (PD) by end of 3rd month. CMT was independently associated with improvement in BCVA. There were significant reductions in FAZ area, VD and PD over 3 months from baseline.

Conclusion

At three months, intravitreal anti-VEGF therapy is associated with visual function recovery in DME, with reduction in CMT being the most reliable predictor of response. OCTA shows a reduction in VD and PD respectively suggestive of reduced macular perfusion in the short term. Simultaneous reduction in FAZ size can be due to reversal of manual displacement of the retinal capillaries resulting from resolution of intraretinal oedema.

Retinal disease in the neurology clinic

PURPOSE OF REVIEW

Retinal disease can manifest with visual symptoms similar to those which result from central nervous system disorders. We provide a framework for considering retinal causes of common visual complaints presenting to a neurology clinic.

RECENT FINDINGS

Technological advances have afforded quicker detection and a more thorough understanding of these retinal entities and are crucial to consider when evaluating visual complaints in the neurology clinic.

SUMMARY

It is essential to maintain a working knowledge of common retinal conditions that symptomatically overlap with common neurologic conditions. Furthermore, the ophthalmoscopic exam and retinal imaging modalities can both aid in the diagnosis and workup of visual complaints and neurologic disease.