Ophthalmic Surgery, Lasers and Imaging Retina

Clinical Science 

Macular Capillary Perfusion in Chinese Patients With Diabetic Retinopathy Obtained With Optical Coherence Tomography Angiography

Dawei Yang, MD; Dan Cao, MD; Zhongning Huang, MD; Jianteng Xie, MD; Qianli Meng, MD; Xinran Dong, MD; Yunyan Hu, MD; Yunkao Zeng, MD; Liang Zhang, MD

Abstract

BACKGROUND AND OBJECTIVE:

To compare the macular perfusion in the retina and choroidal layer between control subjects and Chinese patients with diabetic retinopathy (DR) using optical coherence tomography angiography (OCTA) and to evaluate the association of OCTA characteristics with the stage of DR.

PATIENTS AND METHODS:

A total of 200 eyes (normal controls = 40; mild non-proliferative diabetic retinopathy [NPDR] = 40; moderate NPDR = 40; severe NPDR = 40; and PDR [proliferative diabetic retinopathy] = 40) underwent OCTA imaging. OCTA parameters were vessel densities in the superficial capillary plexus (SCP), deep capillary plexus (DCP), and choriocapillaris, as well as foveal avascular zone (FAZ) area (mm2) in the SCP.

RESULTS:

The reduction of macular perfusion in the SCP, DCP, and choriocapillaris was correlated with increasing severity of DR. Vessel density in the SCP, DCP, and choriocapillaris was 55.31% ± 2.56%, 62.40% ± 2.46%, and 66.87% ± 1.30%, respectively, in control subjects; 50.58% ± 3.14%, 56.31% ± 4.24%, and 66.20% ± 1.69%, respectively, in mild NPDR; 46.46% ± 3.09%, 49.40% ± 5.68%, and 64.39% ± 1.94%, respectively, in moderate NPDR; 45.61% ± 3.81%, 49.33% ± 6.14%, and 63.75% ± 2.21%, respectively, in severe NPDR; and 43.78% ± 3.71%, 44.78% ± 6.36%, and 61.32% ± 6.29%, respectively, in PDR. Vessel density in DR groups decreased compared with normal controls (P < .001). FAZ area in the SCP was 0.34 ± 0.09 mm2 in control subjects compared with 0.48 ± 0.17 mm2 (mild NPDR), 0.52 ± 0.13 mm2 (moderate NPDR), 0.62 ± 0.24 mm2 (severe NPDR), and 0.75 ± 0.30 mm2 (PDR). FAZ in the SCP of patients with DR was greater than that in control subjects (P < .001). Vessel density in the DCP shows better ability to identify the severity of DR (area under the curve, sensitivity, and specificity of 0.967, 92.5%, and 93.1%, respectively) than vessel density in the SCP and choriocapillaris.

CONCLUSION:

OCTA might be clinically useful to evaluate different stages of DR in a noninvasive manner. Vessel density in DCP could be an objective and reliable indicator for monitoring progression of DR.

[Ophthalmic Surg Lasers Imaging Retina. 2019;50:e88–e95.]

Abstract

BACKGROUND AND OBJECTIVE:

To compare the macular perfusion in the retina and choroidal layer between control subjects and Chinese patients with diabetic retinopathy (DR) using optical coherence tomography angiography (OCTA) and to evaluate the association of OCTA characteristics with the stage of DR.

PATIENTS AND METHODS:

A total of 200 eyes (normal controls = 40; mild non-proliferative diabetic retinopathy [NPDR] = 40; moderate NPDR = 40; severe NPDR = 40; and PDR [proliferative diabetic retinopathy] = 40) underwent OCTA imaging. OCTA parameters were vessel densities in the superficial capillary plexus (SCP), deep capillary plexus (DCP), and choriocapillaris, as well as foveal avascular zone (FAZ) area (mm2) in the SCP.

RESULTS:

The reduction of macular perfusion in the SCP, DCP, and choriocapillaris was correlated with increasing severity of DR. Vessel density in the SCP, DCP, and choriocapillaris was 55.31% ± 2.56%, 62.40% ± 2.46%, and 66.87% ± 1.30%, respectively, in control subjects; 50.58% ± 3.14%, 56.31% ± 4.24%, and 66.20% ± 1.69%, respectively, in mild NPDR; 46.46% ± 3.09%, 49.40% ± 5.68%, and 64.39% ± 1.94%, respectively, in moderate NPDR; 45.61% ± 3.81%, 49.33% ± 6.14%, and 63.75% ± 2.21%, respectively, in severe NPDR; and 43.78% ± 3.71%, 44.78% ± 6.36%, and 61.32% ± 6.29%, respectively, in PDR. Vessel density in DR groups decreased compared with normal controls (P < .001). FAZ area in the SCP was 0.34 ± 0.09 mm2 in control subjects compared with 0.48 ± 0.17 mm2 (mild NPDR), 0.52 ± 0.13 mm2 (moderate NPDR), 0.62 ± 0.24 mm2 (severe NPDR), and 0.75 ± 0.30 mm2 (PDR). FAZ in the SCP of patients with DR was greater than that in control subjects (P < .001). Vessel density in the DCP shows better ability to identify the severity of DR (area under the curve, sensitivity, and specificity of 0.967, 92.5%, and 93.1%, respectively) than vessel density in the SCP and choriocapillaris.

CONCLUSION:

OCTA might be clinically useful to evaluate different stages of DR in a noninvasive manner. Vessel density in DCP could be an objective and reliable indicator for monitoring progression of DR.

[Ophthalmic Surg Lasers Imaging Retina. 2019;50:e88–e95.]

Introduction

The number of people with diabetes aged 20 to 79 years is predicted to rise to 642 million by 2040.1 China has the largest population of diabetes in the world. It is estimated that more than 92 million people aged older than 20 years developed diabetes in China in 2010.2,3 With the increasing prevalence of diabetes in the world, diabetic retinopathy (DR) has become the leading cause of blindness in the working-age population.4 Vitreous hemorrhage, tractional retinal detachment, and neovascular glaucoma are common complications of DR causing vision loss.

Accurate staging and classification of DR are crucial for guiding treatment and determining prognosis. Despite the diagnostic value of fluorescein angiography (FA), it has a number of potential side effects and is cautiously used on patients with renal failure.5,6

Optical coherence tomography angiography (OCTA) is an alternative noninvasive angiographic technique that can examine the retinal vasculature and choriocapillaris without any contrast agent injection and demonstrate vessel densities and shape of the vascular arcades of the foveal avascular zone (FAZ) in patients with DR. In this research, we aim to compare macular perfusion and FAZ area of Chinese patients with different stages of DR, as well as to evaluate the association between OCTA characteristics and the severity of DR.

Patients and Methods

Subjects

Patients were recruited from Guangdong General Hospital who presented between January 2017 and July 2017. Eyes of control subjects and patients with type 2 diabetes mellitus were randomly selected in the study. All included subjects were Chinese. Exclusion criteria were any other ocular disease that may affect ocular circulation (eg, glaucoma, age-related macular degeneration, retinal vascular occlusion, refractive error greater than 3 diopters [D]), intraocular surgery, panretinal photocoagulation, hypertension exceeding 150/100 mm Hg, and intraocular pressure (IOP) greater than 21 mm Hg.

Study Setup

Subjects were tested for best-corrected visual acuity (BCVA), IOP, and refractive error (autorefractometry). Slit-lamp and fundus examinations using direct and / or indirect ophthalmoscope were performed. Early Treatment Diabetic Retinopathy Study (ETDRS) 35° seven standard field color retinal photographs (Topcon TRC; Topcon, Tokyo, Japan) were obtained from each participant. DR was graded according to the International Diabetic Retinopathy Severity Scale.7 The study was masked: Two graders evaluated the fundus photographs for absence of changes related to DR or other disease that may affect retinal blood flow. In case of inconsistent opinions between graders, a third grader determined the status of the participant. Macular capillary perfusion parameters were obtained after pupillary dilation in a dark room by using AngioVue OCTA system (RTVue-XR Avanti; Optovue, Fremont, CA, version 2016.2.035). Split-spectrum amplitude decorrelation angiography software algorithm was used for evaluation of vessel density and FAZ area.

Two sets of imaging were performed at one assessment. Each image set comprised two raster volumetric patterns (one vertical priority and one horizontal priority) covering high-definition 6 mm × 6 mm. An orthogonal registration algorithm (built-in software, which has the ability to correct some of the motion artifacts) was used to merge three-dimensional OCT angiograms. Each volume was composed of 400 scan lines. The parameters were evaluated as vessel density (%) of the superficial capillary plexus (SCP), superficial FAZ area (mm2), vessel density (%) of the deep capillary plexus (DCP), and choriocapillaris. The vessel density was the percentage of signal positive pixels per total pixels in an area of interest. FAZ area (mm2) was evaluated in the superficial FAZ area by using the nonflow area tool of the software that delineated it automatically after selecting a segment of the FAZ. The superficial retinal, deep retinal, and choroidal vascular networks were generated by using an automated software algorithm.

The boundaries for each layer were a slab extending from 3 μm to 15 μm from the internal limiting membrane (ILM) for detecting the superficial vascular layer, a slab extending from 15 μm to 70 μm below the ILM for the deep retinal vascular layer, and a slab extending from 30 μm to 60 μm below the retinal pigment epithelium reference for choriocapillaris.

Image quality was considered by including images having signal strength (SS) of at least 40. In patients with poor images, we repeated the scans until an image with at least fair quality could be obtained. Registered image sets with residual artifacts (substantial discontinuous vessel pattern or hazy images) were excluded from the analysis. Intraoperator reproducibility was checked in five control participants for whom five consecutive measurements were taken by two technicians who took the OCTA measurements.

Statistical Analysis

Statistical analysis was performed with SPSS 19.0 software (SPSS Inc., Chicago, IL). Qualitative variables were presented as number and percentage. Quantitative variables were presented as means and standard deviations. One-way analysis of variance (ANOVA) test was used to compare the mean values of vessel density of the SCP, DCP, and choriocapillaris as well as the FAZ area of patients with DR (graded according to their severity) and normal subjects. A P value of less than .05 was considered statistically significant. Receiver operator characteristic (ROC) curve was used to assess the relationship between the severity of DR and macular perfusion indices.

The study was performed in accordance with the Declaration of Helsinki and approved by the Research Ethics Committee of Guangdong General Hospital (registration number: gdrec2016232A).

Results

A total of 200 eyes (40 eyes of 40 normal controls and 160 eyes of 160 diabetic patients with DR) were assessed initially for image quality after they met inclusion and exclusion criteria. Baseline demographics of the entire cohort were comparable in all five groups. Of the 200 patients, 107 (53.5%) were male and 93 (46.5%) were female. Gender and age showed no significant differences in between-group comparisons. Of the 200 selected eyes, 97 (48.5%) were right eyes, and 103 (51.5%) were left eyes. Patients were classified into a normal group and four DR (mild non-proliferative diabetic retinopathy [NPDR], moderate NPDR, severe NPDR, and proliferative diabetic retinopathy [PDR]) groups.

Perfusion Index and Severity of DR

Table 1 shows vessel density (mean ± SD) and mean FAZ area (mm2) in different groups. There was a significant decrease in vessel density and increase in FAZ area as the severity of DR progressed.

Vessel Density and FAZ Area of Different Layers of DR Subgroups and Normal Controlsin Relation to DR Severity

Table 1:

Vessel Density and FAZ Area of Different Layers of DR Subgroups and Normal Controlsin Relation to DR Severity

Table 2 demonstrates the outcome of one-way ANOVA comparing DR subgroups and normal controls in relation to DR severity. Macular perfusion of SCP and DCP were noted and inversely correlated with the degree of DR severity. A similar trend was also found in choriocapillaris. However, it was also noted that there was no significant difference when moderate NPDR was compared with severe NPDR in different plexus.

Outcome of One-Way ANOVA Comparing DR Subgroups and Normal Controlsin Relation to DR Severity

Table 2:

Outcome of One-Way ANOVA Comparing DR Subgroups and Normal Controlsin Relation to DR Severity

We also compared vessel density from four quadrants (temporal, nasal, inferior, superior) according to the ETDRS standard classification (Figure 1). Vessel density from different quadrants all decreased as DR progressed in SCP and DCP. Table 3 demonstrates perfusion index of four different quadrants in five groups.

A high-definition 6 mm × 6 mm of the macular area in a normal individual in optical coherence tomography angiography. (1) Superior area in the parafoveal area. (2) Temporal area in the parafoveal area. (3) Inferior area in parafoveal area. (4) Nasal area in parafoveal area.

Figure 1.

A high-definition 6 mm × 6 mm of the macular area in a normal individual in optical coherence tomography angiography. (1) Superior area in the parafoveal area. (2) Temporal area in the parafoveal area. (3) Inferior area in parafoveal area. (4) Nasal area in parafoveal area.

Performance Parameters for Vessel Density From Different Quadrants in Different Groups

Table 3:

Performance Parameters for Vessel Density From Different Quadrants in Different Groups

Foveal Avascular Zone and Severity of DR

FAZ area enlarged significantly with the worsening of DR. The increase of FAZ area agreed strongly with the increase in severity of DR; however, there was no significant difference between mild NPDR and moderate NPDR in the FAZ area (Table 2).

AUROC Analysis

Table 4 details the performance of the vessel density of different layers for identifying severity of DR based on area under the receiver operating characteristic (AUROC) curve. Area under the curve (AUC) in the SCP, DCP, and choriocapillaris was 95.8%, 96.7%, and 84.5%, respectively. Vessel density of DCP shows better ability to identify the severity of DR (sensitivity and specificity of 92.5%, and 93.1%, respectively) than in SCP and choriocapillaris (Figure 2). The AUC in the superior quadrant was 97.8% in the SCP and 95.3% in the DCP. The AUC of the superior quadrant was higher than that of other quadrants of parafovea (Table 5; Figures 3 and 4).

Performance Parameters for Vessel Density From Different Layers for IdentifyingSeverity of DR

Table 4:

Performance Parameters for Vessel Density From Different Layers for IdentifyingSeverity of DR

(A) Receiver operating characteristic (ROC) curves for vessel density of superficial capillary plexus (SCP), deep capillary plexus (DCP) and choriocapillaris for detecting stage of diabetic retinopathy (DR). (B) ROC curves for vessel density of temporal, nasal, inferior and superior in SCP for detecting different stage of DR. (C) ROC curves for vessel density of temporal, nasal, inferior and superior in DCP for detecting stage of DR.

Figure 2.

(A) Receiver operating characteristic (ROC) curves for vessel density of superficial capillary plexus (SCP), deep capillary plexus (DCP) and choriocapillaris for detecting stage of diabetic retinopathy (DR). (B) ROC curves for vessel density of temporal, nasal, inferior and superior in SCP for detecting different stage of DR. (C) ROC curves for vessel density of temporal, nasal, inferior and superior in DCP for detecting stage of DR.

Performance Parameters for Vessel Density From Different Quadrants for Identifying Severity of DR

Table 5:

Performance Parameters for Vessel Density From Different Quadrants for Identifying Severity of DR

Discussion

In our research, the reduction in the macular perfusion of the SCP, DCP, and choriocapillaris was significantly related to the degree of DR scale, whereas FAZ area increased significantly with the progression of DR. Vessel density from different quadrants all decreased as DR progressed in different plexus. AUROC for the superior quadrant was significantly higher than other quadrants of parafoveal area. In terms of vessel densities in different layers, the DCP showed better ability to identify the severity of DR than in the SCP and choriocapillaris.

Previous studies reported that macular perfusion density agreed closely with grading based on clinical features and may offer an objective method for monitoring disease progression in DR.8–10 Agemy et al. found that diabetic patients had a significantly lower capillary perfusion density compared with normal controls.8 In the study of Lee et al., vessel densities were significantly lower in different layers of DR groups compared with normals. Patients with PDR had statistically significant lower vessel density values compared with patients with NPDR.11 In our study, a significant decrease in vessel density of SCP, DCP, and choriocapillaris was also noted as the severity of DR progressed, which was in accordance to previous reports showing that quantitative information of OCTA could be clinically useful to evaluate the stage of DR.

Sambhav et al. showed that as DR progressed, the decrease in perfusion index was more pronounced in the DCP than in SCP.12 For the first time, we used AUROC analysis to demonstrate that vessel density of DCP shows better ability to identify the severity of DR than the other two layers, with sensitivity and specificity of 92.5% and 93.1%, respectively. DR is a microangiopathy causing compromise of deep and superficial retinal plexus. Hyperglycemia induced the thickening of the basement as well as the death of pericyte,13,14 which resulted in the incompetence of the vascular walls. Anatomically, superficial vascular plexus consisted of most of the arteries and veins, of which the precapillary arterioles were surrounded by a single layer of smooth muscle. However, the deep vascular plexus contained the majority of capillaries, which were enveloped by pericytes;15,16 therefore, DCP may be more sensitive to hypoxia. In addition, reports suggested that microaneurysms, one of the hallmark microvascular features of DR,17 were found in the DCP easier than in the SCP.18,19

We also found that the enlargement of FAZ area in the SCP was correlated with increasing severity of DR. FAZ area in the SCP was 0.34 ± 0.09 mm2 in control subjects compared with 0.48 ± 0.17 mm2 in mild NPDR, 0.52 ± 0.13 mm2 in moderate NPDR, 0.62 ± 0.24 mm2 in severe NPDR, and 0.75 ± 0.30 mm2 in PDR. The increased FAZ area in diabetic eyes is consistent with previous studies.20–22 The correlation analysis of FAZ area and the stages of DR indicate FAZ area has a significant relation to the progression of DR. Takase et al.23 evaluated the FAZ area detected by OCTA in diabetic eyes and showed that there was a statistically significant enlargement of the FAZ compared with nondiabetic eyes, regardless of the presence of diabetic retinopathy. Studies24–29 suggested that the mechanism of the enlargement of FAZ area was related with capillary closure. Miyamoto et al.27 found the expression of intercellular adhesion molecule 1 was elevated, and leukocyte aggregation resulted in obstructed capillary at the early stage of DR. Other studies30 also implied that capillary dropout could lead to the increase of FAZ area. However, FAZ area in healthy individuals has high variability. In our research, AUC of the FAZ area was 0.126; therefore, FAZ area may not be a sensitive biomarker to the stage of DR.

Research by Kern et al.31 showed that vascular lesions in diabetes were distributed non-uniformly within the retina. Microaneurysms and acellular capillaries were not uniformly distributed across the retina in diabetes, both lesions being significantly more prevalent in the superior temporal retina than in the inferior nasal retina. Another study also revealed that lesions such as microaneurysms and pericyte ghosts were most numerous in the temporal area of the retina, followed closely by the superior retina in diabetic patients.32 The findings might indicate that ischemia in diabetes developed in a non-uniform pattern as non-perfusion area were uniformly distributed across the retina.33 Our results indicated that vessel density from the superior quadrant, comparing with other quadrants, showed better correlation with the severity of DR, with AUC of 97.8% in the SCP and 95.3% in the DCP. Interestingly, clinical observation also suggested that neovascularization also appeared to be more common in the superior and temporal portion of the retina.34 Whether a certain sector is more susceptible to DR progression and better associated with the severity of DR has not been clear yet. It is of great interest to analyze the relationship between the stage of DR and the decrease of vessel density from different sectors.

Our study has several limitations. First, this study only included a limited number of patients. More patients need to be evaluated with OCTA before our findings can be fully validated. Second, vessel projection artifact from the SCP on DCP and choriocapillaris should be reduced so that vessel density in different layers could be analyzed more precisely.35 Third, OCTA can only detect certain area of the posterior pole. Peripheral retina is also required for the diagnosis of DR. A larger scanning area of OCTA should be considered in its upgraded version in the future.

In conclusion, OCTA might be a useful method for screening diabetic eyes. As DR progressed, the reduction of vessel density in capillary plexus and choriocapillaris was noted, and vessel density of the DCP showed good ability to identify the severity of DR. Although it is unlikely to replace conventional FA in its current form, it has the potential to alter clinical practice and reduce the need for FA in diabetic patients.

References

  1. Ogurtsova K, da Rocha Fernandes JD, Huang Y, et al. IDF Diabetes Atlas: Global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes Res Clin Pract. 2017;128:40–50. doi:10.1016/j.diabres.2017.03.024 [CrossRef]
  2. Yang SH, Dou KF, Song WJ. Prevalence of diabetes among men and women in China. N Engl J Med. 2010;362(25):2425–2426; author reply 2426. doi:10.1056/NEJMoa0908292 [CrossRef]
  3. Guariguata L, Whiting DR, Hambleton I, Beagley J, Linnenkamp U, Shaw JE. Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res Clin Pract. 2014;103(2):137–149. doi:10.1016/j.diabres.2013.11.002 [CrossRef]
  4. Yang L, Shao J, Bian Y, et al. Prevalence of type 2 diabetes mellitus among inland residents in China (2000–2014): A meta-analysis. J Diabetes Investig. 2016;7(6):845–852. doi:10.1111/jdi.12514 [CrossRef]
  5. Alemzadeh-Ansari MJ, Beladi-Mousavi SS, Feghhei M. Effect of fluorescein on renal function among diabetic patients. Nefrologia. 2011;31(5):612–613.
  6. Almalki WH, Abdalla AN, Elkeraie AF, Abdelhadi AM, Elrggal M, Elrggal ME. Effect of fluorescein angiography on renal functions in type 2 diabetes patients: A pilot study. Saudi J Kidney Dis Transpl. 2017;28(3):491–498. doi:10.4103/1319-2442.206444 [CrossRef]
  7. Polak BC, Hartstra WW, Ringens PJ, Scholten RJ. Revised guideline ‘Diabetic retinopathy: Screening, diagnosis and treatment’. Ned Tijdschr Geneeskd. 2008;152(44):2406–2413.
  8. Agemy SA, Scripsema NK, Shah CM, et al. (2015) Retinal vascular perfusion density mapping using optical coherence tomography angiography in normals and diabetic retinopathy patients. Retina. 2015;35(11):2353–2363. doi:10.1097/IAE.0000000000000862 [CrossRef]
  9. Bonnin S, Mané V, Couturier A, et al. New Insight into the macular deep vascular plexus imaged by optical coherence tomography angiography. Retina. 2015;35(11):2347–2352. doi:10.1097/IAE.0000000000000839 [CrossRef]
  10. Hwang TS, Jia Y, Gao SS, et al. Optical coherence tomography angiography features of diabetic retinopathy. Retina. 2015;35(11):2371–2376. doi:10.1097/IAE.0000000000000716 [CrossRef]
  11. Lee J, Rosen R. Optical coherence tomography angiography in diabetes. Curr Diab Rep. 2016;16(12):123. doi:10.1007/s11892-016-0811-x [CrossRef]
  12. Sambhav K, Abu-Amero KK, Chalam KV. Deep capillary macular perfusion indices obtained with OCT angiography correlate with degree of nonproliferative diabetic retinopathy. Eur J Ophthalmol. 2017;27(6):716–729. doi:10.5301/ejo.5000948 [CrossRef]
  13. Iwasaki M, Inomata H. Relation between superficial capillaries and foveal structures in the human retina. Invest Ophthalmol Vis Sci. 1986;27(12):1698–1705.
  14. Garner A. Histopathology of diabetic retinopathy in man. Eye (Lond). 1993;7(Pt 2):250–253. doi:10.1038/eye.1993.58 [CrossRef]
  15. Tenth Cambridge Ophthalmological Symposium. Perrers Taylor Memorial September 1980. The ocular vasculature. Trans Ophthalmol Soc UK. 1980;100(3):327–447.
  16. Flower RW. Physiology of the developing ocular vasculature. Birth Defects Orig Artic Ser. 1988;24(1):129–146.
  17. Duh EJ, Sun JK, Stitt AW. (2017) Diabetic retinopathy: Current understanding, mechanisms, and treatment strategies. JCI Insight. 2017;2(14). pii: 93751. doi:10.1172/jci.insight.93751 [CrossRef]
  18. Hasegawa N, Nozaki M, Takase N, Yoshida M, Ogura Y. New insights into microaneurysms in the deep capillary plexus detected by optical coherence tomography angiography in diabetic macular edema. Invest Ophthalmol Vis Sci. 2016;57(9):OCT348–355. doi:10.1167/iovs.15-18782 [CrossRef]
  19. Querques G, Bandello F, Souied EH. Abnormal deep retinal capillary networking and microaneurysms in the outer nuclear layer of diabetic eyes. Ophthalmology. 2014;121(3):803–804.e1. doi:10.1016/j.ophtha.2013.10.034 [CrossRef]
  20. Takusagawa HL, Liu L, Ma KN, et al. Projection-resolved optical coherence tomography angiography of macular retinal circulation in glaucoma. Ophthalmology. 2017;124(11):1589–1599. doi:10.1016/j.ophtha.2017.06.002 [CrossRef]
  21. Lupidi M, Coscas G, Coscas F, et al. Retinal microvasculature in nonproliferative diabetic retinopathy: Automated quantitative optical coherence tomography angiography assessment. Ophthalmic Res. 2017;58(3):131–141. doi:10.1159/000471885 [CrossRef]
  22. de Carlo TE, Chin AT, Bonini Filho MA, et al. Detection of microvascular changes in eyes of patients with diabetes but not clinical diabetic retinopathy using optical coherence tomography angiography. Retina. 2015;35(11):2364–2370. doi:10.1097/IAE.0000000000000882 [CrossRef]
  23. Takase N, Nozaki M, Kato A, Ozeki H, Yoshida M, Ogura Y. Enlargement of foveal avascular zone in diabetic eyes evaluated by en face optical coherence tomography angiography. Retina. 2015;35(11):2377–2383. doi:10.1097/IAE.0000000000000849 [CrossRef]
  24. Chibber R, Ben-Mahmud BM, Chibber S, Kohner EM. Leukocytes in diabetic retinopathy. Curr Diabetes Rev. 2007;3(1):3–14. doi:10.2174/157339907779802139 [CrossRef]
  25. Yamana Y, Oka Y, Ohnishi Y, Ishibashi T, Inoguchi T. Reflow of obstructed capillaries in the maculae of humans with diabetic retinopathy, observed by fluorescein angiography. Br J Ophthalmol. 1988;72(9):660–665. doi:10.1136/bjo.72.9.660 [CrossRef]
  26. McLeod DS, Lefer DJ, Merges C, Lutty GA. Enhanced expression of intracellular adhesion molecule-1 and P-selectin in the diabetic human retina and choroid. Am J Pathol. 1995;147(3):642–653.
  27. Miyamoto K, Khosrof S, Bursell SE, et al. Prevention of leukostasis and vascular leakage in streptozotocin-induced diabetic retinopathy via intercellular adhesion molecule-1 inhibition. Proc Natl Acad Sci USA. 1999;96(19):10836–10841. doi:10.1073/pnas.96.19.10836 [CrossRef]
  28. Miyamoto K, Ogura Y. Pathogenetic potential of leukocytes in diabetic retinopathy. Semin Ophthalmol. 1999;14(4):233–239. doi:10.3109/08820539909069542 [CrossRef]
  29. Kim SY, Johnson MA, McLeod DS, Alexander T, Hansen BC, Lutty GA. Neutrophils are associated with capillary closure in spontaneously diabetic monkey retinas. Diabetes. 2005;54(5):1534–1542. doi:10.2337/diabetes.54.5.1534 [CrossRef]
  30. Tam J, Dhamdhere KP, Tiruveedhula P, et al. Subclinical capillary changes in non-proliferative diabetic retinopathy. Optom Vis Sci. 2012;89(5):E692–703. doi:10.1097/OPX.0b013e3182548b07 [CrossRef]
  31. Kern TS, Engerman RL. Vascular lesions in diabetes are distributed non-uniformly within the retina. Exp Eye Res. 1995;60(5):545–549. doi:10.1016/S0014-4835(05)80069-7 [CrossRef]
  32. Tang J, Mohr S, Du YD, Kern TS. Non-uniform distribution of lesions and biochemical abnormalities within the retina of diabetic humans. Curr Eye Res. 2003;27(1):7–13. doi:10.1076/ceyr.27.2.7.15455 [CrossRef]
  33. Borgmann H. Diabetic retinopathy, diagnosis and treatment. Hippokrates. 1969;40(7):251–254.
  34. Taylor E, Dobree JH. Proliferative diabetic retinopathy. Site and size of initial lesions. Br J Ophthalmol. 1970;54(1):11–18. doi:10.1136/bjo.54.1.11 [CrossRef]
  35. Tang F, Cheung CY. Quantitative retinal optical coherence tomography angiography in patients with diabetes without diabetic retinopathy. Invest Ophthalmol Vis Sci. 2017;58(3):1766. doi:10.1167/iovs.17-21552 [CrossRef]

Vessel Density and FAZ Area of Different Layers of DR Subgroups and Normal Controlsin Relation to DR Severity

NormalMild NPDRModerate NPDRSevere NPDRPDR
SCP (%)55.31 ± 2.5650.58 ± 3.1446.46 ± 3.0945.61 ± 3.8143.78 ± 3.71
DCP (%)62.40 ± 2.4656.31 ± 4.2449.40 ± 5.2849.33 ± 6.1444.78 ± 6.36
Choriocapillaris (%)66.87 ± 1.3066.20 ± 1.6964.39 ± 1.9463.75 ± 2.2161.32 ± 6.29
FAZ (mm2)0.34 ± 0.0930.48 ± 0.1720.52 ± 0.1350.62 ± 0.2430.75 ± 0.308

Outcome of One-Way ANOVA Comparing DR Subgroups and Normal Controlsin Relation to DR Severity

Statistical Results (ANOVA)
ParametersSCP (P Value)DCP (P Value)Choriocapillaris (P Value)
Comparison Between Vessel Density of Five GroupsNormal vs. Mild NPDR< .001< .001.034
Normal vs. Moderate NPDR< .001< .001< .001
Normal vs. Severe NPDR< .001< .001< .001
Normal vs. Severe PDR< .001< .001< .001
Mild NPDR vs. Moderate NPDR< .001< .001< .001
Mild NPDR vs. Severe NPDR< .001< .001< .001
Mild NPDR vs. PDR< .001< .001< .001
Moderate NPDR vs. Severe NPDR.252.952.380
Moderate NPDR vs. PDR< .001< .001< .001
Severe NPDR vs. PDR.013< .001< .001
Comparison Between FAZ Area of Five GroupsNormal vs. Mild NPDR.003
Normal vs. Moderate NPDR< .001
Normal vs. Severe NPDR< .001
Normal vs. Severe PDR< .001
Mild NPDR vs. Moderate NPDR.342
Mild NPDR vs. Severe NPDR.003
Mild NPDR vs. PDR< .001
Moderate NPDR vs. Severe NPDR.041
Moderate NPDR vs. PDR< .001
Severe NPDR vs. PDR.008

Performance Parameters for Vessel Density From Different Quadrants in Different Groups

NormalMild NPDRModerate NPDRSevere NPDRPDR
SCPTemporal59.29 ± 1.9352.85 ± 4.4949.37 ± 4.4047.60 ± 6.2844.64 ± 4.50
Superior59.95 ± 2.6952.29 ± 4.6748.32 ± 5.3545.11 ± 5.8443.27 ± 5.85
Nasal58.49 ± 1.7651.90 ± 4.7847.56 ± 5.7446.59 ± 4.9543.02 ± 6.05
Inferior59.04 ± 2.3752.49 ± 5.4747.24 ± 5.7344.93 ± 6.2542.18 ± 4.85
DCPTemporal64.86 ± 2.2659.96 ± 4.3455.22 ± 6.8053.19 ± 6.9848.61 ± 7.94
Superior66.27 ± 2.2659.87 ± 5.5055.50 ± 6.2852.94 ± 6.8048.70 ± 8.65
Nasal64.48 ± 2.4059.50 ± 4.5154.36 ± 7.6453.83 ± 5.6847.69 ± 9.76
Inferior65.73 ± 2.9360.42 ± 5.2956.23 ± 5.1252.18 ± 7.7047.78 ± 7.51

Performance Parameters for Vessel Density From Different Layers for IdentifyingSeverity of DR

AUROC (95% CI)Sensitivity (%)Specificity (%)
AUC for SCP0.958 (0.930–0.987)95.085.6
AUC for DCP0.967 (0.942–0.992)92.593.1
AUC for CL0.845 (0.782–0.909)80.080.0
AUC for FAZ0.126 (0.073–0.159)97.51.3

Performance Parameters for Vessel Density From Different Quadrants for Identifying Severity of DR

AUROC (95% CI) in SCPAUROC (95% CI) in DCP
Temporal0.972 (0.953–0.991)0.942 (0.912–0.973)
Superior0.978 (0.957–0.999)0.953 (0.925–0.982)
Nasal0.977 (0.959–0.995)0.935 (0.902–0.969)
Inferior0.952 (0.924–0.980)0.931 (0.897–0.965)
Authors

From the Department of Ophthalmology, Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences, Guangzhou, Guangdong Province, China (DY, DC, ZH, XD, QM, YH, YZ, LZ); Shantou University Medical College, Guangdong Province, China (DY, YZ); and the Department of Nephrology, Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences, Guangzhou, Guangdong Province, China (JX).

Dr. Yang and Dr. Cao contributed equally to this paper as co-first authors.

The authors report no relevant financial disclosures.

Address correspondence to Liang Zhang, MD, Department of Ophthalmology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, No. 106, Zhongshan Er Road, Yuexiu District, Guangzhou, Guangdong, China; email: zhangliang5413@163.com.

Received: January 29, 2018
Accepted: May 09, 2018

10.3928/23258160-20190401-12

Sign up to receive

Journal E-contents