Automated diabetic retinopathy detection using optical coherence tomography angiography: a pilot study

2018 ◽  
Vol 102 (11) ◽  
pp. 1564-1569 ◽  
Author(s):  
Harpal Singh Sandhu ◽  
Nabila Eladawi ◽  
Mohammed Elmogy ◽  
Robert Keynton ◽  
Omar Helmy ◽  
...  

BackgroundOptical coherence tomography angiography (OCTA) is increasingly being used to evaluate diabetic retinopathy, but the interpretation of OCTA remains largely subjective. The purpose of this study was to design a computer-aided diagnostic (CAD) system to diagnose non-proliferative diabetic retinopathy (NPDR) in an automated fashion using OCTA images.MethodsThis was a two-centre, cross-sectional study. Adults with type II diabetes mellitus (DMII) were eligible for inclusion. OCTA scans of the macula were taken, and the five vascular maps generated per eye were analysed by a novel CAD system. For the purpose of classification/diagnosis, three different local features—blood vessel density, blood vessel calibre and the size of the foveal avascular zone (FAZ)—were segmented from these images and used to train a new, automated classifier.ResultsOne hundred and six patients with DMII were included in the study, 23 with no DR and 83 with mild NPDR. When using features of the superficial retinal map alone, the system demonstrated an accuracy of 80.0% and area under the curve (AUC) of 76.2%. Using the features of the deep retinal map alone, accuracy was 91.4% and AUC 89.2%. When data from both maps were combined, the presented CAD system demonstrated overall accuracy of 94.3%, sensitivity of 97.9%, specificity of 87.0%, area under curve (AUC) of 92.4% and dice similarity coefficient of 95.8%.ConclusionAutomated diagnosis of NPDR using OCTA images is feasible and accurate. Combining this system with OCT data is a plausible next step that would likely improve its robustness.

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Reza Mirshahi ◽  
Pasha Anvari ◽  
Hamid Riazi-Esfahani ◽  
Mahsa Sardarinia ◽  
Masood Naseripour ◽  
...  

AbstractThe purpose of this study was to introduce a new deep learning (DL) model for segmentation of the fovea avascular zone (FAZ) in en face optical coherence tomography angiography (OCTA) and compare the results with those of the device’s built-in software and manual measurements in healthy subjects and diabetic patients. In this retrospective study, FAZ borders were delineated in the inner retinal slab of 3 × 3 enface OCTA images of 131 eyes of 88 diabetic patients and 32 eyes of 18 healthy subjects. To train a deep convolutional neural network (CNN) model, 126 enface OCTA images (104 eyes with diabetic retinopathy and 22 normal eyes) were used as training/validation dataset. Then, the accuracy of the model was evaluated using a dataset consisting of OCTA images of 10 normal eyes and 27 eyes with diabetic retinopathy. The CNN model was based on Detectron2, an open-source modular object detection library. In addition, automated FAZ measurements were conducted using the device’s built-in commercial software, and manual FAZ delineation was performed using ImageJ software. Bland–Altman analysis was used to show 95% limit of agreement (95% LoA) between different methods. The mean dice similarity coefficient of the DL model was 0.94 ± 0.04 in the testing dataset. There was excellent agreement between automated, DL model and manual measurements of FAZ in healthy subjects (95% LoA of − 0.005 to 0.026 mm2 between automated and manual measurement and 0.000 to 0.009 mm2 between DL and manual FAZ area). In diabetic eyes, the agreement between DL and manual measurements was excellent (95% LoA of − 0.063 to 0.095), however, there was a poor agreement between the automated and manual method (95% LoA of − 0.186 to 0.331). The presence of diabetic macular edema and intraretinal cysts at the fovea were associated with erroneous FAZ measurements by the device’s built-in software. In conclusion, the DL model showed an excellent accuracy in detection of FAZ border in enfaces OCTA images of both diabetic patients and healthy subjects. The DL and manual measurements outperformed the automated measurements of the built-in software.


2019 ◽  
Vol 30 (6) ◽  
pp. 1418-1423 ◽  
Author(s):  
Claudio Furino ◽  
Grazia Montrone ◽  
Maria Vittoria Cicinelli ◽  
Stefania Balestra ◽  
Maria Oliva Grassi ◽  
...  

Purpose: To investigate a subset of diabetic patients without diabetic retinopathy with optical coherence tomography angiography, assessing the differences in macular perfusion between diseased eyes and healthy controls. Methods: Monocentric cross-sectional study, including 86 eyes from 43 diabetic patients with no clinical signs of diabetic retinopathy and 78 eyes from 39 controls. Patients underwent 3.0 × 3.0 mm and 4.5 × 4.5 mm swept-source optical coherence tomography angiography. Vessel density (%), foveal avascular zone area (mm2), and avascular density (%) were provided for the superficial capillary plexus and the deep capillary plexus. Results: The foveal avascular zone area at the superficial capillary plexus was larger in the study group compared to controls, irrespective of the area of the slab considered. A meaningful difference was found in the vessel density at the deep capillary plexus of the 3.0 × 3.0 mm slab (p = 0.03). Almost all the variables considered in the study showed a significant within-subject effect. Age significantly correlated with vessel density of superficial capillary plexus on 4.5 × 4.5 mm in both control and diabetic eyes. Conclusion: Diabetic patients with subclinical diabetic retinopathy feature a larger foveal avascular zone at the superficial capillary plexus compared with controls, as well as relative reduction of the vessel density at the deep capillary plexus. These findings might serve as the basis for screening between normal and diabetic subjects.


2020 ◽  
pp. bjophthalmol-2020-317983
Author(s):  
Edward S Lu ◽  
Ying Cui ◽  
Rongrong Le ◽  
Ying Zhu ◽  
Jay C Wang ◽  
...  

AimsTo compare the efficacy of diabetic retinal neovascularisation (NV) detection using the widefield swept-source optical coherence tomography angiography (WF SS-OCTA) vitreoretinal interface (VRI) Angio slab and SS-OCT VRI Structure slab.MethodsA prospective, observational study was performed at Massachusetts Eye and Ear from January 2019 to June 2020. Patients with proliferative diabetic retinopathy (PDR), patients with non-proliferative diabetic retinopathy and patients with diabetes but without diabetic retinopathy were included. All patients were imaged with WF SS-OCTA using the 12×12 mm Angio scan protocol centred on the fovea and optic disc. The en-face SS-OCTA VRI Angio slab and SS-OCT VRI Structure slab were evaluated for the presence or absence of NV. SS-OCTA B-scan was used to classify NV according to cross-sectional morphology (forward, tabletop or flat). All statistical analyses were performed using SPSS V.26.0.ResultsOne hundred and forty-two eyes of 89 participants were included in the study. VRI Angio detected NV at higher rates compared with VRI Structure (p<0.05). Combining VRI Angio and Structure improved detection rates compared with VRI Angio alone (p<0.05). Due to segmentation errors of the internal limiting membrane, NV with flat morphological classification had lower rates of detection on VRI Angio compared with NV with forward and tabletop morphology (p<0.05).ConclusionsWF SS-OCTA 12×12 mm VRI Angio and SS-OCT VRI Structure imaging centred on the fovea and optic disc detected NV with high sensitivity and low false positives. The VRI slab may be useful to diagnose and monitor PDR in clinical practice.


2020 ◽  
Vol 12 ◽  
pp. 251584141989745 ◽  
Author(s):  
Hamid Safi ◽  
Pasha Anvari ◽  
Dariush Naseri ◽  
Hamideh Shenazandi ◽  
Pegah Kazemi ◽  
...  

Purpose: To evaluate the correlation of the visual acuity and diabetic retinopathy stage using optical coherence tomography and optical coherence tomography angiography metrics. Methods: In this prospective study, optical coherence tomography and optical coherence tomography angiography images of patients with different stages of diabetic retinopathy were obtained. In optical coherence tomography angiography images, the size of foveal avascular zone, central macular thickness, and vessel density at superficial and deep capillary layers of the macula were measured. In optical coherence tomography images, the presence of intraretinal cyst, disorganization of retinal inner layer, and ellipsoid zone and external limiting membrane disruption were evaluated. The associations between the variables with visual acuity and diabetic retinopathy stage were analyzed. Results: In total, 68 eyes of 38 patients with a mean age of 58.96 ± 10.59 years were included. In total, 34 eyes were categorized as non-proliferative diabetic retinopathy, 14 as active, and 20 as regressed proliferative diabetic retinopathy. Univariate analysis showed deep parafoveal vessel density, central macular thickness, ellipsoid zone disruption, disorganization of retinal inner layer, and external limiting membrane disruption had a significant relationship with visual acuity. However, in multivariate analysis, only central macular thickness and ellipsoid zone disruption had significant association with visual acuity ( p = 0.02 and p = 0.01, respectively). There was a significant difference in deep parafoveal vessel density ( p = 0.04), but not in foveal avascular zone area, between different stages of diabetic retinopathy. Conclusion: In this study, the foveal avascular zone area did not correlate with visual acuity and different stages of diabetic retinopathy. Structural abnormalities on optical coherence tomography images with especial focus on outer retinal disruption provided more reliable predictors for visual acuity outcomes in patients with diabetic retinopathy.


Author(s):  
Fariba Ghassemi ◽  
Sahar Berijani ◽  
Ramak Roohipoor ◽  
Masoumeh Mohebbi ◽  
Ameneh Babeli ◽  
...  

Abstract Purpose To measure optic nerve head (ONH) blood perfusion using optical coherence tomography angiography (OCTA) at various stages of diabetic retinopathy (DR). Methods One hundred seventy six eyes of 94 patients included in this retrospective single-centre cross-sectional study. The subjects were studied in normal, no diabetic retinopathy (NDR), non-proliferative diabetic retinopathy (NPDR) and proliferative retinopathy (PDR) groups. The eyes were subjected to AngioDisc ONH imaging using OCTA for papillary (Disc) and peripapillary (RPC) vascular density (VD) evaluation. Results The mean age of the participants was 56.08 ± 8.87 years and 34 (36.2 percent) were male. With increased DR severity, a statistically significant decrease in peripapillary VD was found. The study showed that only VD of the whole RPC (W-RPC) could be a valid biomarker in the staging assessment. VD of RPC, in all subsections, was considerably different from normal cases in the PDR group. Visual acuity was correlated with whole image ONH VD. The duration of DM, FBS, hyperlipidemia and DME had no effect on the ONH perfusion. Conclusions The study showed that only the W-RPC VD could be a reasonable marker in the staging assessment. VDs assessed by OCTA can be useful for assessing and tracking early ONH changes in DR patients.


2018 ◽  
Vol 45 (10) ◽  
pp. 4582-4599 ◽  
Author(s):  
Nabila Eladawi ◽  
Mohammed Elmogy ◽  
Fahmi Khalifa ◽  
Mohammed Ghazal ◽  
Nicola Ghazi ◽  
...  

2015 ◽  
Vol 254 (6) ◽  
pp. 1051-1058 ◽  
Author(s):  
Florentina J. Freiberg ◽  
Maximilian Pfau ◽  
Juliana Wons ◽  
Magdalena A. Wirth ◽  
Matthias D. Becker ◽  
...  

2019 ◽  
Vol 6 (1) ◽  
Author(s):  
Kai Yuan Tey ◽  
Kelvin Teo ◽  
Anna C. S. Tan ◽  
Kavya Devarajan ◽  
Bingyao Tan ◽  
...  

Abstract Background Diabetic retinopathy (DR) is a leading cause of vision loss in adults. Currently, the standard imaging technique to monitor and prognosticate DR and diabetic maculopathy is dye-based angiography. With the introduction of optical coherence tomography angiography (OCTA), it may serve as a potential rapid, non-invasive imaging modality as an adjunct. Main text Recent studies on the role of OCTA in DR include the use of vascular parameters e.g., vessel density, intercapillary spacing, vessel diameter index, length of vessels based on skeletonised OCTA, the total length of vessels, vascular architecture and area of the foveal avascular zone. These quantitative measures may be able to detect changes with the severity and progress of DR for clinical research. OCTA may also serve as a non-invasive imaging method to detect diabetic macula ischemia, which may help predict visual prognosis. However, there are many limitations of OCTA in DR, such as difficulty in segmentation between superficial and deep capillary plexus; and its use in diabetic macula edema where the presence of cystic spaces may affect image results. Future applications of OCTA in the anterior segment include detection of anterior segment ischemia and iris neovascularisation associated with proliferative DR and risk of neovascular glaucoma. Conclusion OCTA may potentially serve as a useful non-invasive imaging tool in the diagnosis and monitoring of diabetic retinopathy and maculopathy in the future. Future studies may demonstrate how quantitative OCTA measures may have a role in detecting early retinal changes in patients with diabetes.


2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Jing Zhai ◽  
Wei Fang ◽  
Xueting Yu ◽  
Xinjie Ye ◽  
Lijie Hou

Abstract Background To quantify and compare retinal microvascular features using optical coherence tomography angiography (OCTA) in constant (XT) and intermittent exotropia (IXT). Methods An observational cross-sectional study was conducted from September 2020 to November 2020 at the Affiliated Eye Hospital of Wenzhou Medical University. OCTA examination was performed to evaluate the macular perfusion density of the deep capillary plexus (DCP), superficial capillary plexus (SCP), and foveal avascular zone (FAZ) parameters in XT and IXT patients, and in age-matched controls. The study parameters were analyzed. Results A total of 76 participants, including 16 (21%) XT patients, 24 (32%) IXT patients, and 36 (47%) controls, were recruited. The exodeviation angle was 39.06 ± 10.38 prism degrees (PD) at near and 43.00 ± 9.74 PD at distance in the XT group, and 27.13 ± 18.28 PD at near and 31.04 ± 18.82 PD at distance in the IXT group. The macular perfusion density of the DCP in 6 × 6-mm scans and the SCP in 3 × 3-mm scans were lower in the deviating eyes than in the fixating eyes of XT patients (p < 0.001, p = 0.032, respectively). The macular perfusion density of the DCP in the deviating eyes of XT and IXT patients was significantly lower than that of the controls. There was no significant difference in the FAZ parameters among the groups. Conclusions In XT patients, OCTA revealed lower macular perfusion density in deviating eyes than in fixating eyes and control eyes. IXT patients showed no difference between the two eyes, but the macular perfusion density of the DCP was lower than that of the controls.


Sign in / Sign up

Export Citation Format

Share Document