Optical Coherence Tomography Angiography Images Research Articles

Differential Artery–Vein Analysis Improves OCTA Classification of Diabetic Retinopathy

This study investigates the impact of differential artery-vein (AV) analysis in optical coherence tomography angiography (OCTA) on machine learning classification of diabetic retinopathy (DR). Leveraging deep learning for arterial-venous area (AVA) segmentation, six quantitative features, including perfusion intensity density (PID), blood vessel density (BVD), vessel area flux (VAF), blood vessel caliber (BVC), blood vessel tortuosity (BVT), and vessel perimeter index (VPI) features, were derived from OCTA images before and after AV differentiation. A support vector machine (SVM) classifier was utilized to assess both binary and multiclass classifications of control, diabetic patients without DR (NoDR), mild DR, moderate DR, and severe DR groups. Initially, one-region features, i.e., quantitative features extracted from the entire OCTA, were evaluated for DR classification. Differential AV analysis improved classification accuracies from 78.86% to 87.63% and from 79.62% to 85.66% for binary and multiclass classifications, respectively. Additionally, three-region features derived from the entire image, parafovea, and perifovea, were incorporated for DR classification. Differential AV analysis further enhanced classification accuracies from 84.43% to 93.33% and from 83.40% to 89.25% for binary and multiclass classifications, respectively. These findings highlight the potential of differential AV analysis in augmenting disease diagnosis and treatment assessment using OCTA.

Deep learning-driven automated quality assessment of ultra-widefield optical coherence tomography angiography images for diabetic retinopathy

AbstractImage quality assessment (IQA) of fundus images constitutes a foundational step in automated disease analysis. This process is pivotal in supporting the automation of screening, diagnosis, follow-up, and related academic research for diabetic retinopathy (DR). This study introduced a deep learning-based approach for IQA of ultra-widefield optical coherence tomography angiography (UW-OCTA) images of patients with DR. Given the novelty of ultra-widefield technology, its limited prevalence, the high costs associated with equipment and operational training, and concerns regarding ethics and patient privacy, UW-OCTA datasets are notably scarce. To address this, we initially pre-train a vision transformer (ViT) model on a dataset comprising 6 mm × 6 mm OCTA images, enabling the model to acquire a fundamental understanding of OCTA image characteristics and quality indicators. Subsequent fine-tuning on 12 mm × 12 mm UW-OCTA images aims to enhance accuracy in quality assessment. This transfer learning strategy leverages the generic features learned during pre-training and adjusts the model to evaluate UW-OCTA image quality effectively. Experimental results demonstrate that our proposed method achieves superior performance compared to ResNet18, ResNet34, and ResNet50, with an AUC of 0.9026 and a Kappa value of 0.7310. Additionally, ablation studies, including the omission of pre-training on 6 mm × 6 mm OCTA images and the substitution of the backbone network with the ViT base version, resulted in varying degrees of decline in AUC and Kappa values, confirming the efficacy of each module within our methodology.

Impact of Solar Eclipses on Vision: Insights from Optical Coherence Tomography and Optical Coherence Tomography Angiography Analysis*

Purpose Solar retinopathy, resulting from solar eclipse exposure, poses risks to visual health. This study explores acute and chronic phase findings using clinical examinations and optical coherence tomography (OCT) and optical coherence tomography angiography (OCT-A) with a focus on longitudinal assessment. Methods Seven eyes with a history of unprotected solar eclipse exposure were included. Clinical examination, fundus photography, OCT, and OCT-A imaging were performed at initial assessment, as well as at one-month and six-month follow-up intervals. Data analysis included descriptive statistics. Results The cases, exposed without protection, underwent assessments, revealing variable visual acuity, outer retinal layer, and Henle fiber layer changes during follow-up. Regression of hyperreflectivity within the outer retinal and Henle fiber layers was observed over time in all eyes, although persistent microdefects within the outer retinal layer were noted in specific cases. OCT-A imaging revealed a larger foveal avascular zone, which persisted over a six-month period in select cases. Additionally, affected eyes exhibited a decrease in superficial vascular density, with subsequent improvement noted during the six-month period. Conclusion Solar retinopathy can result in visual impairment, accompanied by alterations observed in the Henle fiber layer using OCT. Additionally, OCT-A findings indicate possible vascular involvement. This study underscores the significance of adopting protective measures during solar eclipses and emphasizes the value of employing longitudinal multimodal imaging techniques to comprehend the pathophysiology of the condition.

A deep learning system for the detection of optic disc neovascularization in diabetic retinopathy using optical coherence tomography angiography images

A deep learning system for the detection of optic disc neovascularization in diabetic retinopathy using optical coherence tomography angiography images

Relationship between macular perfusion and lesion distribution in diabetic retinopathy.

To assess the relationship between macular vessel density metrics and foveal avascular zone (FAZ) characteristics on optical coherence tomography angiography (OCTA) and lesion distribution in eyes with diabetic retinopathy (DR). Patients with DR who underwent both Optos ultrawidefield (UWF) pseudocolor imaging and macular OCTA (Cirrus Angioplex, 6 × 6 mm) were included in this cross-sectional observational study. The distribution of DR lesions was assessed by comparing each of the peripheral ETDRS extended fields (3-7) against their corresponding ETDRS field, hence eyes were defined as either having predominantly peripheral lesions (PPL) or predominantly central lesions (PCL). En face OCTA images from the superficial and deep capillary plexuses (SCP and DCP) were then analysed using Image J software. Perfusion density (PD), vessel length density (VLD), and fractal dimensions (FD) were calculated following binarization and skeletonization of the images. Out of 344 eyes, 116 (33.72%) eyes had PPL and 228 (66.28%) eyes had PCL. For all DRSS levels, VLD, PD, and FD were not significantly different between eyes with PPL and PCL. The FAZ in eyes with PPL, however, was found to be more circular in shape compared to eyes with PCL (p = 0.037). Although the presence of PPL has been associated with a higher risk for diabetic retinopathy progression, the macular perfusion is similar in eyes with PPL and PCL. The FAZ is more circular in eyes with PPL, but the clinical relevance of this difference remains to be defined.

Glaucoma detection using non-perfused areas in OCTA

Multiple ophthalmic diseases lead to decreased capillary perfusion that can be visualized using optical coherence tomography angiography images. To quantify the decrease in perfusion, past studies have often used the vessel density, which is the percentage of vessel pixels in the image. However, this method is often not sensitive enough to detect subtle changes in early pathology. More recent methods are based on quantifying non-perfused or intercapillary areas between the vessels. These methods rely upon the accuracy of vessel segmentation, which is a challenging task and therefore a limiting factor for reliability. Intercapillary areas computed from perfusion-distance measures are less sensitive to errors in the vessel segmentation since the distance to the next vessel is only slightly changing if gaps are present in the segmentation. We present a novel method for distinguishing between glaucoma patients and healthy controls based on features computed from the probability density function of these perfusion-distance areas. The proposed approach is evaluated on different capillary plexuses and outperforms previously proposed methods that use handcrafted features for classification. Moreover the results of the proposed method are in the same range as the ones of convolutional neural networks trained on the raw input images and is therefore a computationally efficient, simple to implement and explainable alternative to deep learning-based approaches.

Reduction of Optic Disc Microvasculature and Retinal Nerve Fiber Layer Thinning in Patients with Glaucoma

PurposeTo assess the relationship between the change of optic disc vessel density (ODVD) and retinal nerve fiber layer (RNFL) thinning in primary open-angle glaucoma (POAG) patients. DesignRetrospective case series. MethodsFor 105 POAG patients, ≥ 5 consecutive optical coherence tomography (OCT) and OCT angiography (OCTA) images were obtained during ≥ 2 years of follow-up.Based on enface OCTA imaging, ODVD was calculated as the ratio of pixels occupied by vessels below the internal limiting membrane within the temporal area of the optic cup, and ODVD reduction was determined when there was a statistically significant negative slope (P < 0.05) for any of the global, superior, or inferior sectors. The association between the rates of ODVD change and RNFL thinning was assessed by multivariable longitudinal linear mixed-effects model vs. time. ResultsDuring 2.9 ± 0.3 years of follow-up on the 105 participants with visual field mean deviation at baseline of 5.7 ± 4.8 dB, 46 (43.8%) showed ODVD reduction. Faster global RNFL thinning was associated with the smaller Bruch's membrane opening area (ß = 0.381; 95% confidence interval (CI), 0.120 to 0.646; P = 0.006), optic disc hemorrhage (ß = -0.567; 95% CI, -0.909 to -0.228; P = 0.002) and faster rate of global ODVD change (ß = -0.090; 95% CI, -0.139 to -0.042; P = 0.001). ConclusionsReduction of optic disc microvasculature was associated with rapid RNFL thinning in POAG. This suggests a role for deep ONH circulation in the glaucoma pathogenesis.

Optical coherence tomography angiography changes in patients with hemoglobinopathy

Aim: The aim of this study is to evaluate retinal vascular changes in patients with sickle cell disease (SCD) and beta-thalassemia with optical coherence tomography angiography (OCT-A). Methods: For this purpose, 98 patients with SCD, 75 patients with beta-thalassemia, and 100 healthy controls in Mersin University Hospital between January 1, 2020, and November 1, 2021, were included in this study. OCT-A imaging was performed with ZEISS AngioPlex OCT angiography (Carl Zeiss Meditec, Dublin, CA, USA). Results: All OCT-A parameters (FAZ area, perimeter, circularity, vessel, and perfusion density) were found to be statistically significantly different in both patients with thalassemia and patients with sickle cell disease when compared to the controls. Conclusions: In conclusion, retinopathy related to both hemoglobinopathy subgroups can be diagnosed and followed up with OCT-A. It was also found that OCT-A parameters are affected before the development of clinically detectable retinopathy.

Parafoveal capillary density as a potential biomarker in the assessment of diabetic retinopathy progression

AbstractPurposeThe aim of this study was to assess longitudinal changes in retinal vessel density in diabetic patients using optical coherence tomography angiography (OCTA) to identify the most sensitive parameter for detecting retinopathy progression.MethodsPatients with diabetes mellitus were enrolled in this study. Each study subject underwent two imaging sessions, during which three OCTA images of the macular area and three images of the optic nerve head were obtained. The two sessions took place one year apart. The OCTA imaging was performed using an AngioVue device. Superficial vessel density was evaluated in the central 3 mm and parafoveal area, and the nonflow area was measured using the built-in automated AngioAnalytics software of the Optovue system.ResultsThis study included 78 eyes of 39 diabetic patients (age: 55.16 ± 13.73 years) with a mean of 7.70 ± 1.07 mmol L−1 HgA1c level at baseline. At the one-year visit, the eyes of the diabetic subjects had significantly lower superficial vessel density in the parafoveal macula compared to corresponding values at baseline (p &lt; 0.05). There was no statistically significant difference between the baseline and one-year results for the other vascular density parameters and the foveal avascular zone (p &gt; 0.05).ConclusionAt the one-year follow-up, we found that vessel density had decreased in the parafoveal ring in these patients, while there was no significant change in other vascular parameters. This result suggests that superficial parafoveal capillary density is the most sensitive OCTA parameter that can be used as a biomarker for diabetic retinopathy progression. None of the other vascular density parameters nor the foveal avascular zone were able to indicate the subtle changes in retinal microcirculation due to the progression of diabetic microvasculopathy.

Analysis of Choriocapillaris Reperfusion Topography following Faricimab Treatment for Neovascular Age-Related Macular Degeneration in Non-Treatment-Naïve Patients.

To assess changes in choriocapillaris (CC) vascular density surrounding macular neovascularization (MNV) in age-related macular degeneration (AMD) when transitioning from various anti-VEGF treatments to faricimab, using optical coherence tomography angiography (OCTA). 25 eyes of 22 individuals who underwent intravitreal faricimab injections for neovascular AMD with type 1 MNV were included. OCTA images were obtained prior to (T0), after one (T1), and after three faricimab injections (T2); Noteworthy changes occurred in the first ring at T2 in comparison to T0. The percentage of CC flow deficit (FD%), FD average area (FDa), and FD number (FDn) in 5 rings (R1-R5) surrounding the dark halo around the MNV were calculated. A reduction in FD% at T2 compared to T0 (50.5 ± 10.2% at T0, 46.4 ± 10.6% at T2; p = 0.020) was seen, indicating CC reperfusion. Additionally, we observed a reduction in the average FDa (140.2 ± 172.1% at T0, 93.7 ± 101.8% at T2; p = 0.029). Our study highlights an FD% after three consecutive faricimab injections. The most pronounced effect was observed in the first ring, directly adjacent to the dark halo, suggesting a partial CC reperfusion surrounding the MNV, potentially indicating disease regression.

Aflibercept combined with triamcinolone acetonide in the treatment of diabetic macular edema: optical coherence tomography and optical coherence tomography angiography.

To analyze the relationship between optical coherence tomography (OCT) and OCT angiography (OCTA) imaging in patients with diabetic macular edema (DME) who are treated with a combination of aflibercept and triamcinolone acetonide (TA). A total of 76 eyes newly diagnosed DME were included in this study. They were randomly assigned to receive either aflibercept or a combination of aflibercept and TA. Injections once a month for a total of three injections. Central macular thickness (CMT), number of hyperreflective foci (HRF), height of subretinal fluid (SRF), and area of foveal avascular zone (FAZ) were evaluated using OCT and OCTA at baseline and after each monthly treatment. Both groups showed improvement in best corrected visual acuity (BCVA) and reduction in macular edema after treatment, and the difference in BCVA between the two groups was statistically significant after each treatment (P<0.05). The difference in CMT between the two groups was statistically significant after the first two injections (P<0.01), but not after the third injection (P=0.875). The number of HRF (1mo: 7.41±8.25 vs 10.86±7.22, P=0.027; 2mo: 5.33±6.13 vs 9.12±8.61, P=0.034; 3mo: 3.58±3.00 vs 6.37±5.97, P=0.007) and height of SRF (1mo: 82.39±39.12 vs 105.77±42.26 µm, P=0.011; 2mo: 36.84±10.02 vs 83.59±37.78 µm, P<0.01; 3mo: 11.57±3.29 vs 45.43±12.60 µm, P<0.01) in combined group were statistically significant less than aflibercept group after each injection, while the area of FAZ showed no significant change before and after treatment in both groups. The combination therapy of aflibercept and TA shows more significant effects on DME eyes with decreased HRF and SRF. However, both aflibercept and combination therapy show no significant change in the area of FAZ.

Computational Retinal Microvascular Biomarkers from an OCTA Image in Clinical Investigation.

Retinal structural and functional changes in humans can be manifestations of different physiological or pathological conditions. Retinal imaging is the only way to directly inspect blood vessels and their pathological changes throughout the whole body non-invasively. Various quantitative analysis metrics have been used to measure the abnormalities of retinal microvasculature in the context of different retinal, cerebral and systemic disorders. Recently developed optical coherence tomography angiography (OCTA) is a non-invasive imaging tool that allows high-resolution three-dimensional mapping of the retinal microvasculature. The identification of retinal biomarkers from OCTA images could facilitate clinical investigation in various scenarios. We provide a framework for extracting computational retinal microvasculature biomarkers (CRMBs) from OCTA images through a knowledge-driven computerized automatic analytical system. Our method allows for improved identification of the foveal avascular zone (FAZ) and introduces a novel definition of vessel dispersion in the macular region. Furthermore, retinal large vessels and capillaries of the superficial and deep plexus can be differentiated, correlating with retinal pathology. The diagnostic value of OCTA CRMBs was demonstrated by a cross-sectional study with 30 healthy subjects and 43 retinal vein occlusion (RVO) patients, which identified strong correlations between OCTA CRMBs and retinal function in RVO patients. These OCTA CRMBs generated through this "all-in-one" pipeline may provide clinicians with insights about disease severity, treatment response and prognosis, aiding in the management and early detection of various disorders.

Quantitative evaluation of OCT angiography images in healthy and glaucomatous subjects through a novel approach: exploring inter-image variability.

This study aims to investigate inter-image intra-observer variability of macular, and optic disc (ONH) microvasculature measurements of glaucomatous and normal subjects using Swept-Source Optical Coherence Tomography Angiography (OCT-A) (OCT Topcon ImageNet 6; DRI OCT Triton, Topcon Corporation, JAPAN) - based imaging data analysis and processing with a newly made quantitative approach. A total of 20 glaucomatous and 20 healthy eyes underwent three OCT-A scanning of the ONH and macula. Macular and papillary and peripapillary vascular networks were calculated. For each eye, eighteen scans were analyzed using a novel approach: custom MATLAB 2021b scripts were employed for imaging analyses. Grayscale distribution was performed using the histcounts MATLAB function with 51 bins. For all layers, the vascular layer coefficient of variation (vl CoV) of the three measures were performed. The vl CoV difference between the two groups was analyzed by Student t-test. In glaucomatous eyes, the vl CoV ranged from 4.49% to 8.54%, while in the control group from 3.58% to 8.32%. Both groups exhibited higher CoVs when assessing the optic disc. The papillary and macular microvasculature reproducibility was comparable between groups. Utilizing Swept-Source OCT-A images our study has identified an easy and reproducible method that appears to be fast and can assist physicians in assessing macular and ONH perfusion with less inter-image variability, particularly in the 70 μm superficial area of the optic disc. The high reliability obtained suggested that this method could be useful as early clinical biomarker.

Multimodal deep transfer learning to predict retinal vein occlusion macular edema recurrence after anti-VEGF therapy

PurposeTo develop a multimodal deep transfer learning (DTL) fusion model using optical coherence tomography angiography (OCTA) images to predict the recurrence of retinal vein occlusion (RVO) and macular edema (ME) after three consecutive anti-VEGF therapies. MethodsThis retrospective cross-sectional study consisted of 2800 B-scan OCTA macular images collected from 140 patients with RVO-ME. The central macular thickness (CMT) > 250 μm was used as a criterion for recurrence in the three-month follow-up after three injections of anti-VEGF therapy. The qualified OCTA image preprocessing and the lesion area segmentation were performed by senior ophthalmologists. We developed and validated the clinical, DTL, and multimodal fusion models based on clinical and extracted OCTA imaging features. The performance of the models and experts predictions were evaluated using several performance metrics, including the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity. ResultsThe DTL models exhibited higher prediction efficacy than the clinical models and experts’ predictions. Among the DTL models, the Vgg19 performed better than that of the other models, with an AUC of 0.968 (95 % CI, 0.943–0.994), accuracy of 0.913, sensitivity of 0.922, and specificity of 0.902 in the validation cohort. Moreover, the fusion Vgg19 model showed the highest prediction efficacy among all the models, with an AUC of 0.972 (95 % CI, 0.946–0.997), accuracy of 0.935, sensitivity of 0.935, and specificity of 0.934 in the validation cohort. ConclusionsMultimodal fusion DTL models showed robust performance in predicting RVO-ME recurrence and may be applied to assist clinicians in determining patients’ follow-up time after anti-VEGF therapy.

The impact of serum BNP on retinal perfusion assessed by an AI-based denoising optical coherence tomography angiography in CHD patients

BackgroundTo investigate the correlation between retinal vessel density (VD) parameters with serum B-type natriuretic peptide (BNP) in patients with coronary heart disease (CHD) using novel optical coherence tomography angiography (OCTA) denoising images based on artificial intelligence (AI). MethodsOCTA images of the optic nerve and macular area were obtained using a Canon-HS100 OCT device in 176 patients with CHD. Baseline information and blood test results were recorded. ResultsRetinal VD parameters of the macular and optic nerves on OCTA were significantly decreased in patients with CHD after denoising. Retinal VD of the superficial capillary plexus (SCP), deep capillary plexus (DCP) and radial peripapillary capillary (RPC) was strongly correlated with serum BNP levels in patients with CHD. Significant differences were noted in retinal thickness and retinal VD (SCP, DCP and RPC) between the increased BNP and normal BNP groups in patients with CHD. ConclusionDeep learning denoising can remove background noise and smooth rough vessel surfaces. SCP,DCP and RPC may be potential clinical markers of cardiac function in patients with CHD. Denoising shows great potential for improving the sensitivity of OCTA images as a biomarker for CHD progression.

Imaging of human corneoscleral junction (Limbus) vasculature with 100 KHz A-scan rate swept-source optical coherence tomography/angiography (OCT/A) system

The study demonstrates in-vivo three-dimensional depth-resolved Optical Coherence Tomography Angiography (OCTA) and Doppler OCT imaging performed on the anterior segment of the human eye. A high-speed swept-source OCT system was developed utilizing a swept-source laser having a 100 kHz A-scan sweep rate and centre wavelength at 1060 nm. The proposed system has an axial resolution of 4.5 µm and a transverse imaging resolution of 14 µm. The Doppler OCT provides information regarding the direction of blood flow in corneoscleral junction vessels, which employs the Joint Spectral and Time domain OCT (STdOCT) method. Parameters influencing the imaging of various vessels, like scanning density, A-scan rate, B-scan averaging, and number of BM scanning repeats, were also studied. Slit lamp images and OCT angiography en-face projections produced from PVOCTA (Phase-variance OCTA) and STdOCTA from the same regular patients were compared in order to demonstrate the additional potential of the system. The vascular systems at the Limbus consisting of circulatory, lymphatic, and aqueous outflow systems were imaged and presented. The output clinical parameters of the system will contribute to understanding the blood circulation with directionality at the limbus location, which can be further utilized to identify glaucoma. In the author's opinion, this article represents the first attempt to provide Doppler OCTA data for the anterior chamber of the human eye.

The longitudinal follow-up of a newly proposed OCTA imaging finding (SSPiM) and the importance of it as a new biomarker for treatment response in diabetic macular edema.

This study aimed to evaluate the frequency of SSPiM (suspended scattering particles in motion), systemic risk factors, ocular findings, progression characteristics, and treatment response in diabetic retinopathy (DR) patients. In this prospective study, a total of 109 eyes of 109 patients with diabetic macular edema (DME) were included. Demographic characteristics and systemic data of the patients were recorded. In addition to a detailed ophthalmological examination, optical coherence tomography (OCT) and OCT angiography (OCTA) imaging were performed. According to the OCTA images, the patients were divided into two categories: SSPiM detected (SSPiM +) and undetected (SSPiM -). The patients were followed up at 0, 3, and 6months. Treatment responses at 6months in treatment-administered patients with and without SSPiM were examined. The frequency of SSPiM in DME cases was found to be 34.9%. No significant correlation was found between SSPiM and demographic characteristics, systemic, and biochemical parameters (p > 0.05). It was observed that SSPIM was most frequently localized in the outer nuclear layer adjacent to the outer plexiform (81.6%). SSPiM appearance disappeared in 7 (19.4%) of 36 patients with SSPiM who had regular follow-up for 6months. In 4 (11.1%) of these seven patients, hard exudate plaques developed in the areas where SSPiM disappeared. Regarding treatment response at 6months, the decrease in CMT was statistically significantly lower in the SSPiM group compared to cases without SSPiM. SSPiM is a finding seen in approximately one-third of DME patients and may adversely affect the response to the treatment.

Choriocapillaris flow features in children with myopic anisometropia

AimsTo examine differences between the eyes in choriocapillaris perfusion and choroidal thickness in children with myopic anisometropia.MethodsIn this observational and prospective study, 46 children with myopic anisometropia were enrolled. Choriocapillaris...

Retinal vasculature of different diameters and plexuses exhibit distinct vulnerability in varying severity of diabetic retinopathy.

To study the changes in vessel densities (VD) stratified by vessel diameter in the retinal superficial and deep vascular complexes (SVC/DVC) using optical coherence tomography angiography (OCTA) images obtained from people with diabetes and age-matched healthy controls. We quantified the VD based on vessel diameter categorized as <10, 10-20 and >20 μm in the SVC/DVC obtained on 3 × 3 mm2 OCTA scans using a deep learning-based segmentation and vascular graph extraction tool in people with diabetes and age-matched healthy controls. OCTA images obtained from 854 eyes of 854 subjects were divided into 5 groups: healthy controls (n = 555); people with diabetes with no diabetic retinopathy (DR, n = 90), mild and moderate non-proliferative DR (NPDR) (n = 96), severe NPDR (n = 42) and proliferative DR (PDR) (n = 71). Both SVC and DVC showed significant decrease in VD with increasing DR severity (p < 0.001). The largest difference was observed in the <10 μm vessels of the SVC between healthy controls and no DR (13.9% lower in no DR, p < 0.001). Progressive decrease in <10 μm vessels of the SVC and DVC was seen with increasing DR severity (p < 0.001). However, 10-20 μm vessels only showed decline in the DVC, but not the SVC (p < 0.001) and there was no change observed in the >20 μm vessels in either plexus. Our findings suggest that OCTA is able to demonstrate a distinct vulnerability of the smallest retinal vessels in both plexuses that worsens with increasing severity of DR.

Deep learning for automatic prediction of early activation of treatment naïve non-exudative MNVs in AMD.

Around 30% of non-exudative macular neovascularizations(NE-MNVs) exudate within 2 years from diagnosis in patients with age-related macular degeneration(AMD).The aim of the study is to develop a deep learning classifier based on optical coherence tomography(OCT) and OCT angiography(OCTA) to identify NE-MNVs at risk of exudation. AMD patients showing OCTA and fluorescein angiography (FA) documented NE-MNV with a 2-years minimum imaging follow-up were retrospectively selected. Patients showing OCT B-scan-documented MNV exudation within the first 2 years formed the EX-GROUP while the others formed QU-GROUP.ResNet-101, Inception-ResNet-v2 and DenseNet-201 were independently trained on OCTA and OCT B-scan images. Combinations of the 6 models were evaluated with major and soft voting techniques. Eighty-nine (89) eyes of 89 patients with a follow-up of 5.7 ± 1.5 years were recruited(35 EX GROUP and 54 QU GROUP). Inception-ResNet-v2 was the best performing among the 3 single convolutional neural networks(CNNs).The major voting model resulting from the association of the 3 different CNNs resulted in improvement of performance both for OCTA and OCT B-scan (both significantly higher than human graders' performance). Soft voting model resulting from the combination of OCTA and OCT B-scan based major voting models showed a testing accuracy of 94.4%. Peripheral arcades and large vessels on OCTA enface imaging were more prevalent in QU GROUP. Artificial intelligence shows high performances in identifications of NE-MNVs at risk for exudation within the first 2 years of follow up, allowing better customization of follow up timing and avoiding treatment delay. Better results are obtained with the combination of OCTA and OCT B-scan image analysis.