Automated FAZ segmentation and diabetic retinopathy classification using OCTA images

Aims/Purpose: To reduce the variability when performing manual segmentation of the foveal avascular zone (FAZ) on retinal optical coherence tomography angiography (OCTA) images, and to find a new accurate FAZ delimitation criterion.Methods: Seventy‐three eyes from healthy subjects, 40 eyes from Type 1 diabetes mellitus (DM1) patients without diabetic retinopathy (DR) and 54 eyes from Type 2 diabetes mellitus (DM2) patients with moderate DR and without diabetic macular oedema (DME) were evaluated using OCTA. A manual FAZ segmentation was performed by different observers of each OCTA image of the superficial (SCP) and deep capillary plexuses (DCP). After comparing the results, a new criterion was established to reduce the variability in the segmentations. The FAZ area and circularity were also studied.Results: The novel criterion produces smaller areas (closer to the real FAZ), as well as a diminution of the standard deviation values, with a factor of 0.5, with lower variability than the different criteria of the explorers in both plexuses for the three groups. This was particularly notorious for the DM2 group with damaged retinas, where it was very difficult to define the FAZ limits. The acircularity values were also slightly reduced with the final criterion in all groups. In the DM2 group a significant reduction in the acircularity mean value of both plexuses was obtained. Looking for a relationship between the FAZ area and the acircularity, we found slightly higher values of acircularity for smaller FAZ areas. The number of extremal values was much higher for the initial method, and when the DCP was evaluated.Conclusions: The implementation of a new criterion for FAZ segmentation produced more consistent results with less variability among observers.

Background: Diabetes mellitus (DM) is a multi-system disease with vision-threatening complications such as diabetic retinopathy (DR) and diabetic macular edema (DME), requiring multi-specialist care. In Egypt and the Middle East, DM prevalence and its complications vary widely, with DR being a leading cause of visual impairment. Imaging tools like optical coherence tomography (OCT) and OCT angiography (OCTA) have become essential in detecting and monitoring DME and DR, offering non-invasive, high-resolution evaluation of retinal microvasculature. Aim: This study aimed to detect changes in the superficial and deep macular plexus in DME before and after injection of anti-vascular endothelial growth factor (anti-VEGF) using OCTA. Patients and methods: This prospective interventional trial included 50 eyes of 39 patients with DME who were divided into non-proliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR) groups. Patients with a history of retinal procedures or non-diabetic causes of macular edema were excluded. All participants underwent standardized ophthalmologic examination and data analysis. Results: The study did not reveal differences in demographics, glycated hemoglobin (HbA1c), intraocular pressure (IOP), DM duration, and the side of the lesion between NPDR and PDR groups. Significant differences in pre- and post-operative best-corrected visual acuity (BCVA) and central macular thickness (CMT) were noted. Vascular density (VD) of the superficial capillary plexus (SCP) and deep capillary plexus (DCP) in parafoveal areas post-operatively was significantly different between the groups, with each group having distinct intra-group changes in visual and anatomical parameters. Conclusion: The VD of the SCP and DCP of the fovea and parafoveal area, as well as foveal avascular zone (FAZ) size at baseline were good predictor factors for both anatomical and visual improvement after intravitreal ranibizumab injection in eyes with DME. OCTA offers measurements for VD and FAZ size at the macula and could be used to predict the visual prognosis of anti-VEGF treatment in DME.

To investigate the effect of image averaging on qualitative and quantitative assessments of optical coherence tomography angiography (OCTA) images from eyes of patients with branch retinal vein occlusion (BRVO). Macular OCTA images of 33 eyes of 33 patients with BRVO were obtained using the HS100 HR-SD-OCT system (Canon, Inc.). For each eye, five OCTA cube scans were obtained with a 3 × 3mm scan protocol, and the data were averaged and compounded into a single high image quality cube data using built-in software. Pre- and post-averaging images were compared qualitatively and quantitatively in superficial capillary plexus (SCP) and deep capillary plexus (DCP) OCTA image slabs. After averaging, all OCTA images showed marked improvement in image quality with less background noise and better vessel continuity. The number of microaneurysms in both the SCP and DCP was larger in single images than in averaged images. A significant increase in the detection rate of capillary telangiectasia in the DCP was noted after image averaging. The number of eyes with disrupted foveal avascular zone (FAZ) decreased significantly after averaging (P = .0253). Five eyes (15.2%) with a disrupted FAZ on the single image showed an intact FAZ after averaging. Vessel length density (VLD) and fractal dimension (FD) significantly decreased and vessel diameter index (VDI) increased after averaging, while significant changes were not observed in vessel density (VD) in both the SCP and DCP. In the SCP, lower VD, VLD, and fractal dimension were significantly correlated with worse visual acuity. OCTA averaging has a significant effect on qualitative and quantitative assessments in eyes with BRVO.

Diabetic retinopathy (DR) is the most severe and frequent retinal vascular disease that causes significant visual loss on a global scale. The purpose of our study was to evaluate retinal vascularization in the superficial capillary plexus (SCP), the deep capillary plexus (DCP) and the choriocapillaris (CC) and changes in the foveal avascular zone (FAZ) by optical tomography angiography (OCTA) in patients with type 2 diabetes mellitus (DM2) with moderate DR but without diabetic macular oedema (DME). Fifty-four eyes of DM2 with moderate DR (level 43 in the ETDRS scale) and without DME and 73 age-matched healthy eyes were evaluated using OCTA with swept-source (SS)-OCT to measure microvascularization changes in SCP, DCP, CC and the FAZ. The mean ages were 64.06 ± 11.98 and 60.79 ± 8.62 years in the DM2 and control groups, respectively. Visual acuity (VA) was lower in the DM2 patients (p = 0.001), OCTA showed changes in the SCP with a significant diminution in the vascular density and the FAZ area was significantly higher compared to healthy controls, with p < 0.001 at the SCP level. The most prevalent anatomical alterations were peripheral disruption in the SCP (83.3%), microaneurysms (MA) in the SCP and in the DCP (79.6% and 79.6%, respectively) and flow changes in the DCP (81.5%). A significant positive correlation was observed between the DM2 duration and the FAZ area in the SCP (0.304 with p = 0.025). A significant negative correlation was also found between age and CC central perfusion (p < 0.001). In summary, a decrease in the vascular density in DM2 patients with moderate DR without DME was observed, especially at the retinal SPC level. Furthermore, it was found that the FAZ was increased in the DM2 group in both retinal plexuses and was greater in the SCP group.

Perifoveal capillary changes in diabetic patients and association between severity and type of diabetes, visual acuity, and enlargement of non-flow area in the retinal capillary plexuses

Purpose To evaluate the retinal vascular response to hyperoxia in patients with diabetes at the preclinical stage of diabetic retinopathy (DR) and to quantify the changes in comparison with normal subjects using optical coherence tomography angiography (OCTA). Methods In this prospective study, 40 eyes of 20 participants comprising 10 diabetic patients with no diabetic retinopathy (NDR) and 10 normal subjects were recruited. OCTA images were acquired in the resting position and were repeated after a hyperoxic challenge using a nasal mask connected to a reservoir bag supplying 100% oxygen at the rate of 15 L per minute for 5 minutes. The changes of mean parafoveal vessel density (VD) in the superficial capillary plexus (SCP) and deep capillary plexus (DCP), the foveal avascular zone (FAZ) size, and the outer retina flow index were compared between two conditions in each group and between the two study groups. The statistical significance of differences in the means was evaluated using Student's t-test for unpaired samples with consideration of the generalized estimating equations (GEE) for intereye correlation. Results At baseline, the mean parafoveal VD of SCP and DCP were significantly lower in the NDR participants compared to the healthy subjects (P < 0.001 and P = 0.006, respectively). After induction of the hyperoxic challenge in healthy participants, mean parafoveal VD reduced at both the SCP and DCP, but reached a statistical significance only in DCP (P = 0.006). However, following induction of hyperoxic challenge in patients with NDR, no significant decline was noticed in mean parafoveal VD of SCP and DCP. The degree of change in mean parafoveal VD of DCP was statistically significantly more pronounced in healthy subjects compared to the NDR group (P = 0.034). The change in FAZ size and the outer retina flow index were comparable between the two study groups. Conclusion Retinal capillary layers responded differently to hyperoxia-induced challenge, and in normal subjects, the autoregulatory mechanism was mostly effective in the parafoveal DCP. Retinal vascular reactivity was impaired in SCP and DCP at the preclinical stage of DR. OCTA as a noninvasive modality was able to quantify the retinal vascular response to the hyperoxic challenge.

PurposeThe purpose of this study was to evaluate specifically in type 1 diabetes mellitus (DM) individuals the relationship between perifoveal superficial capillary plexus (SCP) parameters assessed by optical coherence tomography angiography (OCTA) and diabetic retinopathy (DR) grade.MethodsCross-sectional analysis of a large scale prospective OCTA trial cohort (ClinicalTrials.gov NCT03422965). A total of 1186 eyes (593 individuals), 956 type 1 DM eyes (478 patients), and 230 control eyes (115 healthy volunteers) were included in this study. DR stage was graded according to the International Classification. OCTA imaging was performed with a commercially available device (Cirrus HD-OCT). Vessel density (VD), perfusion density (PD), and foveal avascular zone (FAZ) area, perimeter and circularity measurements were quantified in the SCP and receiver operating characteristic (ROC) curves were constructed for each OCTA parameter.ResultsVD and PD (in both 3 × 3 and 6 × 6 mm captures) were inversely associated with DR stage (P < 0.001 in all cases) in a multiple regression analysis after controlling by age, gender, signal strength index, axial length, and DM duration. Greater FAZ area and perimeter and conversely lower circularity measurements were observed as DR severity increased in both scanning protocols (P < 0.05 in all cases).ConclusionsIn type 1 DM individuals, OCTA provides an objective, continuous, and reliable method for accurate quantification of VD, PD, and FAZ parameters in the SCP, which ultimately correlate with DR stages.Translational RelevanceObjective OCTA measurements of the retinal microvasculature could substitute the clinical DR classification in patients with type 1 DM, identify patients at risk of DR progression, and inform treatment decisions to modify the evolution of the disease.

The Foveal Avascular Zone (FAZ) is of clinical importance since the retinal vascular arrangement around the fovea changes with retinal vascular diseases and in high myopic eyes. Therefore, it is important to segment and quantify the FAZs accurately. Using a novel location-aware deep learning method the FAZ boundary was segmented in en-face optical coherence tomography angiography (OCTA) images. The FAZ dimensions were compared the parameters determined using four methods: (1) device in-built software (Cirrus 5000 Angioplex), (2) manual segmentation using Image J software by an experienced clinician, and (4) the new method (new location-aware deep-learning method). The parameters were measured from OCTA data from healthy subjects (n=34) and myopic patients (n=66). For this purpose, FAZ location was manually delineated in en-face OCTA images of dimensions 420x420 pixels corresponding to 6mm x 6mm. A modified UNet segmentation with an additional channel from a Gaussian distribution around the likely location of the FAZ was designed and trained using 100 manually segmented OCTA images. The predicted FAZ and the related parameters were then obtained using a test dataset consisting of 100 images. For analysis, two strategies were applied. The segmentation of FAZ was compared using the Dice coefficient and Structural Similarity Index (SSIM) to determine the effectiveness of the proposed deep learning method when compared to the other three methods. Furthermore, to provide deeper insight, a set of FAZ dimensions namely area, perimeter, circularity index, eccentricity, perimeter, major axis, minor axis, inner circle radius, circumcircle radius, the maximum and minimum boundary dimensions, and orientation of major axis were compared between the 3 methods. Finally, vessel-related parameters including tortuosity, vessel diameter index (VDI) and vessel avascular density (VAD) were calculated and compared. The high myopic eyes exhibited a narrowing the FAZ area and perimeter. The currently developed algorithm does not correct for axial length variations. This analysis should be extended with a larger number of images in each group of myopia as well as correcting for axial length variations.

This cross-sectional observational case–control clinical study investigated microvascular retinal changes in type 1 diabetes mellitus (DM1) patients with no signs of diabetic retinopathy (DR) in comparison with healthy controls using optical coherence tomography angiography (OCTA). The local institutional review board approved this study. Enrolment criteria included a diagnosis of DM1 made at least 1 year prior to study enrolment (age ≥ 18 years). Only data from the right eye were analysed. Exclusion criteria included bad metabolic control (HbA1c > 9%) and history of any other retinal disease. Two experienced examiners (MP and MV) classified the eyes as no DR and analysed the optical coherence tomography (OCT) images. We used the AngioVue device (RTVue XR, Optovue, Inc., Fremont, CA), to quantify parafoveal vessel density and foveal avascular zone (FAZ) area in the superficial capillary plexus (SCP) and deep capillary plexus (DCP) (Simonett et al. 2017). The split-spectrum amplitude-decorrelation angiography algorithm (Jia et al. 2012) was used to evaluate the blood flow in a 3 × 3 mm scanning area centred on the fovea. En face OCT angiograms were segmented to define the SCP and DCP, using the built-in software segmentation algorithm. Parafoveal vessel density and FAZ were calculated using the automated AngioVue Analytics tool of the OCTA device as shown in details in Fig. 1. This study showed for the first time that DM1 patients with no sign of DR have reduced parafoveal vessel density limited to the DCP when compared to non-diabetic controls. Any difference was found in FAZ area between cohorts. Demographic and clinical data are shown in Fig. 1. In the past, fluorescein angiography was considered a useful tool for revealing incipient retinopathy; indeed, the earlier vascular changes including microaneurysms and leakage are detectable using fluorescein angiography before these latter may seen by means ophthalmoscopy (Dorchy et al. 1979). Today, using OCTA, it is possible to study the complex vascular anatomical architecture of the DCP in DR, which might be more susceptible to diabetic damage because of its complexity. In a previous study using OCTA (Simonett et al. 2017), we found that DM1 patients with no and mild signs of non-proliferative DR showed a reduction in DCP parafoveal vessel density. In this study, by analysing a subgroup of DM1 patients, we demonstrated that patients with no signs of DR also present DCP microvascular impairment. Local retinal acidosis, found in the outer nuclear layer at the early stages of a DR (Dmitriev et al. 2016), could increase leucostasis in small retinal capillaries. Also, diabetic maculopathy seems to be a result of disturbances in retinal vasomotion. (Bek 1999) Therefore, our findings might corroborate the hypothesis that the increased recruitment of microvascular units in the macular area can lead to increased hydrostatic pressure in small capillaries (Bek 1999) with a consequent severely diminished blood flow velocity/rate below the threshold necessary to register as flow in the OCTA system. Finally, the diabetic metabolic disturbances, particularly changes in blood glucose, can influence retinal perfusion and retinal autoregulation, implying that fluctuations in the arterial feeding pressure may be transmitted to the smaller vessels. Limitations are the small sample size and a limited field of view of 3 × 3 mm2 area. Additionally, we cannot exclude that projection from more superficial vasculature may be affecting these findings, particularly in the DCP, for either DM1 patients or control. (Shahlaee et al. 2016) In conclusion, DCP impairment might be harbinger of retinal vascular damage before any sign of DR can be detected by means of ophthalmoscopy, and this finding demonstrated that a decrease in DCP parafoveal capillary density represents an early process in the disease.

PurposeTo describe an automated algorithm to quantify the foveal avascular zone (FAZ), using optical coherence tomography angiography (OCTA), and to compare its performance for diagnosis of diabetic retinopathy (DR) and association with best-corrected visual acuity (BCVA) to that of extrafoveal avascular area (EAA).MethodsWe obtained 3 × 3-mm macular OCTA scans in diabetic patients with various levels of DR and healthy controls. An algorithm based on a generalized gradient vector flow (GGVF) snake model detected the FAZ, and metrics assessing FAZ size and irregularity were calculated. We compared the automated FAZ segmentation to manual delineation and tested the within-visit repeatability of FAZ metrics. The correlations of two conventional FAZ metrics, two novel FAZ metrics, and EAA with DR severity and BCVA, as determined by Early Treatment Diabetic Retinopathy Study (ETDRS) charts, were assessed.ResultsSixty-six eyes from 66 diabetic patients and 19 control eyes from 19 healthy participants were included. The agreement between manual and automated FAZ delineation had a Jaccard index > 0.82, and the repeatability of automated FAZ detection was excellent in eyes at all levels of DR severity. FAZ metrics that incorporated both FAZ size and shape irregularity had the strongest correlation with clinical DR grade and BCVA. Of all the tested OCTA metrics, EAA had the greatest sensitivity in differentiating diabetic eyes without clinical evidence of retinopathy, mild to moderate nonproliferative DR (NPDR), and severe NPDR to proliferative DR from healthy controls.ConclusionsThe GGVF snake algorithm tested in this study can accurately and reliably detect the FAZ, using OCTA data at all DR severity grades, and may be used to obtain clinically useful information from OCTA data regarding macular ischemia in patients with diabetes. While FAZ metrics can provide clinically useful information regarding macular ischemia, and possibly visual acuity potential, EAA measurements may be a better biomarker for DR.

Epidemiologically, men have a higher incidence, severity, and progression of diabetic retinopathy (DR) than women. We investigated microvascular differences between men and women with diabetes on optical coherence tomography angiography (OCTA). Three × 3 mm OCTA macula scans of non-diabetic and patients with diabetes were obtained. Vascular parameters included parafoveal vessel density (VD), vessel length density (VLD), and flow index (FI) of the superficial capillary plexus (SCP) and deep capillary plexus (DCP) as well as foveal avascular zone (FAZ) area and perimeter. Multivariable linear regression was used for statistical analysis. There were 1809 patients with diabetes and 217 non-diabetic participants that were included in this study. Diabetic individuals included those with no DR (n = 1356), mild non-proliferative DR (NPDR; n = 286), moderate NPDR (n = 126), and severe NPDR/proliferative DR (PDR; n = 41). Male sex was significantly associated with smaller FAZ area/perimeter and lower DCP VLD in both non-diabetic subjects and patients with diabetes. Male sex in the diabetic group was additionally associated with lower SCP VD/VLD and DCP VD. Addition of an interaction between male sex and diabetes status in the interaction analysis showed that being male and diabetic conferred increased reduction in DCP VD and VLD compared to sex-based changes in non-diabetics. Larger FAZ perimeter, lower SCP VD/VLD, and lower DCP VLD were associated with poorer visual acuity in diabetics. On OCTA, male patients with diabetes may have more severe microvascular disease especially in the DCP compared to women. Sex-based alterations in diabetic microvascular disease has the potential to influence future basic and clinical studies as well as the implementation of OCTA disease markers.

The retinal and choroidal parameters were analyzed to understand the impairment of microcirculation of both retina and choroid in patients with diabetic retinopathy (DR). Fifty-five treatment-naive non-proliferative diabetic retinopathy (NPDR) patients (75 eyes) with type 2 diabetes mellitus (T2DM), including 28 patients (36 eyes) with diabetic macular edema (DME) and 27 patients (39 eyes) without DME, and 25 healthy subjects (47 eyes) were enrolled in this study. The following parameters of DR patients with and without DME were evaluated: the foveal avascular zone area (FAZ-a), FAZ perimeter (FAZ-p), FAZ circularity index (FAZ-CI), total subfoveal choroidal area (TCA), luminal area (LA), stromal area (SA), choroidal vascularity index (CVI), choriocapillaris flow area percentage, superficial capillary plexus (SCP), and deep capillary plexus (DCP). SCP, DCP, and the percentage of choriocapillaris flow area were significantly different between DR patients with and without DME. The DR patients presented lower LA, CVI, and FAZ-CI compared to those of healthy controls (all p < 0.05). The percentage of choriocapillaris flow area in DR patients with and without DME was significantly lower than that in healthy controls (p < 0.05). SCP and DCP were significantly correlated with FAZ-a and FAZ-p but presented insignificant associations with FAZ-CI. Optical coherence tomography (OCT) and OCT angiography (OCTA) parameters, such as LA, CVI, FAZ-CI, and the percentage of choriocapillaris flow area, were reduced compared to those in controls, indicating that the microcirculations of the retina and choroid in the macular area were impaired in DR patients with DME and without DME.

BackgroundTo analyze the changes in foveal avascular zone (FAZ) area, perimeter, and circularity in the superficial (SCP) and deep (DCP) capillary plexuses in eyes with diabetic macular edema (DME) treated with intravitreal anti-VEGF using optical coherence tomography angiography (OCTA).MethodsThis prospective observational study included 56 eyes from 32 patients with DME that received intravitreal anti-VEGF. OCTA images were obtained at baseline and 1, 3, and 6 months of follow-up. The outcome measures were FAZ area, perimeter, and circularity in both the SCP and DCP, as well as central subfield thickness (CST) and best-corrected visual acuity (BCVA).ResultsThe mean number of intravitreal anti-VEGF injections received during the observation period was 4.60 ± 0.82 (range: 3–6). The FAZ area, perimeter, and circularity were statistically unchanged at all observation points in both the SCP (p = 0.772, p = 0.405, p = 0.157, respectively) and the DCP (p = 0.620, p = 0.769, p = 0.481, respectively). Despite having no change in the FAZ parameters, there was still a statistically significant decrease in CST (p < 0.001) as well as a statistically significant increase in BCVA (p = 0.004) during the observation period.ConclusionsThe FAZ area, perimeter, and circularity in the SCP and DCP as measured by OCTA remained stable during the first 6 months of intravitreal anti-VEGF therapy in eyes with DME. While there were no significant changes in the FAZ, treatment with intravitreal anti-VEGF still resulted in decreased CST and improved BCVA.

Purpose To evaluate the effectiveness of intravitreal bevacizumab treatment in patients with diabetic macular edema (DME) by assessing retinal changes using optical coherence tomography angiography (OCT-A). Methods This prospective study was performed in patients with treatment-naïve DME. The eyes of patients were imaged using a swept-source OCT system with a scan area of 6 × 6 mm. The DME patients with a central macular thickness (CMT) of ≥300 μm received nine bevacizumab injections within 12 months. The demographic, systemic, and ocular parameters, including the best-corrected visual acuity (BCVA), CMT, microaneurysm (MA) count, and foveal avascular zone (FAZ) area in both superficial capillary plexus (SCP) and deep capillary plexus (DCP), as well as vessel density in SCP, were assessed in the patients. In addition, the response (good or poor) of the DME eyes to bevacizumab treatment and the final visual acuity (BCVA of 75 letters) were analyzed. Results Seventy-seven eyes of DME patients were subjected to the final analysis. Bevacizumab treatment reduced CMT from 425.06 μm (±77.15) to 350.25 μm (±82.04) and improved BCVA by about 8.61 letters (from 64.73 to 73.34) in the patients. The mean number of MAs in SCP decreased from 3.51 ± 2.07 to 2.31 ± 1.15 (p < 0.001) and in DCP from 17.12 ± 11.56 to 12.21 ± 6.99 (p < 0.001), whereas the area of FAZ increased in SCP from 328.22 ± 131.38 to 399.70 ± 156.98 (p < 0.001) and in DCP from 571.13 ± 396.01 to 665.89 ± 412.77 (p = 0.001). The final BCVA letter score and CMT were statistically significant in both poor and good responders, as well as in BCVA < 75 and BCVA ≥ 75 groups. Conclusion The fixed-regimen intravitreal bevacizumab therapy was effective in treating DME. Apart from noninvasive visualization of microvascular damage, OCT-A showed limited usefulness in predicting treatment response. Although the study showed that the number of MAs was significantly reduced during treatment, which is an OCT-A predictor of a good response to bevacizumab treatment at a 12-month visit, commonly observed artifacts may reduce the usefulness of OCT-A.

The decorrelation signals in optical coherence tomography angiography (OCTA) are derived from the flow of erythrocytes and concomitantly delineate the retinal vasculature. We compared the structural and functional characteristics of vascular lesions visualized in fluorescein angiography (FA), OCTA, and en-face OCT images in 53 eyes (28 patients) with diabetic retinopathy (DR). The foveal avascular zone (FAZ) areas in OCTA images in the superficial layer almost corresponded to those in FA images. The FAZ areas in the en-face OCT images in the superficial layer were smaller than those in the FA images and correlated with each other, which agreed with the finding that en-face OCT images often delineated the vascular structure in the nonperfused areas in FA images. Microaneurysms appeared as fusiform, saccular, or coiled capillaries in OCTA images and ringed, round, or oval hyperreflective lesions in en-face OCT images. OCTA and en-face OCT images detected 41.0 ± 16.1% and 40.1 ± 18.6%, respectively, of microaneurysms in FA images, although both depicted only 13.9 ± 16.4%. The number of microaneurysms in FA images was correlated with that in OCTA and en-face OCT images. Comparisons of these modalities showed the associations and dissociations between blood flow and vascular structures, which improves the understanding of the pathogenesis of DR.