Diabetic retinopathy (DR) is a leading cause of blindness worldwide,1–3 and an increase in the number of diabetic patients is anticipated in the future. The increase in the incidence of DR is not only in the developed but also in the underdeveloped countries. The increase is a consequence of urbanization, population aging, and changing lifestyle.4–7 Accordingly, DR and its vision-threatening complications such as macular edema, proliferative diabetic retinopathy (PDR), and neovascular glaucoma are emerging problems in eye care throughout the world.8,9
The lesions in eyes with DR are most likely caused by chronic retinal ischemia in the non-perfused retinal areas (NPAs). The ischemia also causes the release of vascular endothelial growth factor (VEGF) from the cells surrounding the NPAs, and the VEGF induces a progression of the DR to PDR.10,11 Once the NPAs are identified, they are treated by photocoagulation to prevent the development of neovascularization and progression to nonproliferative diabetic retinopathy (NPDR) and to PDR.
Fluorescein angiography (FA) has been the gold standard technique to detect NPAs, but the recent introduction of widefield FA has been shown to be more helpful for detecting the NPAs that are widely distributed throughout the retina.12,13 However, FA is an invasive technique because it requires an intravenous injection of fluorescein, and it cannot be performed on all patients.
Optical coherence tomography angiography (OCTA) is a relatively new technique that can detect retinal blood vessels including capillaries without requiring an intravenous dye injection.14–16 The advent of widefield OCTA has allowed clinicians to examine wider areas of the fundus.17 However, obtaining a clear OCTA image of the periphery by conventional optical coherence tomography (OCT) devices is difficult especially in eyes with opacities in the media.
The advantages of widefield image compared to conventional devices are not only its ease in obtaining images of wide areas of the fundus but it also allows the acquisition of seamless images avoiding the absence of some areas of the far periphery. This is especially critical in finding all of the retinal ischemic areas in eyes with DR or retinal vein occlusion (RVO). At present, widefield imaging and multicolor scanning laser ophthalmoscopy (SLO) imaging can be performed by the Optos and Mirante devices. The Mirante device is a multimodal imaging ophthalmic instrument whose image covers a field of view of 163 degrees. The color images are obtained by red, green, and blue confocal lasers. Although the field of view of this instrument is narrower than Optos, which has a 200-degree field of view, the Mirante has an additional blue laser with the green and red lasers. Thus, the characteristics of the images obtained by SLO are different. When a widefield color image is recorded, the original single blue, green, and red SLO images are obtained simultaneously. In the multicolor SLO images, the appearance of the lesion is different for each color because of the differential absorption of the different colors of the laser.18,19
Shin et al20 reported that the hyporeflective areas in the red-free blue SLO images corresponded with the NPAs in eyes with DR and RVO. They suggested that a reduced amount of hemoglobin secondary to a decrease in blood flow in the ischemic areas results in the hyporeflection. However, they used an SLO with the conventional field of view, and the smaller visual field of view was a limitation because NPAs in DR or RVO are frequently seen in the mid-periphery of the fundus. The availability of widefield blue SLO images has allowed clinicians to examine a wider area of the retina. For retinal diseases such as DR or RVO that affect a wide area of the fundus, it would be better if the entire area of the NPAs could be examined non-invasively in a single shot by widefield fundus imaging. This would allow clinicians to investigate the NPAs outside the conventional posterior fundus.
Thus, the purpose of this study was to determine the concordance of widefield blue SLO and FA findings throughout fundus and to evaluate the usefulness of a new method in detecting the NPAs of DR in this relatively easy method.
Study Design and Patients
This study was a retrospective observational case series. We examined the medical records of patients with diabetes mellitus (DM) who had been examined in the outpatient clinic of the Department of Ophthalmology and Visual Science at Tokyo Medical and Dental University (TMDU). All of the subjects had undergone multicolor widefield SLO imaging with the Mirante device (NIDEK, Aichi, Japan) between February 2020 and July 2020. The DM eyes were classified as PDR, moderate to severe NPDR, or less than mild NPDR according to the classification by the global diabetic retinopathy project group.21 Patients excluded were those under 20-years-of-age and those whose SLO images were not clear due to media opacities.
The Ethics Committee of Tokyo Medical and Dental University approved the procedures and waivers of informed consent from all subjects in this data-based retrospective observational study. The research procedures conformed to the tenets of the Declaration of Helsinki.
Widefield Multicolor Scanning Laser Ophthalmoscopy (SLO) and Multimodal Imaging
All patients underwent multicolor widefield SLO fundus photography with the Mirante device. With this device, blue (480 nm), green (532 nm), and red (670 nm) wavelength images of the fundus were recorded, and synthesized color images were created from individual color images (Supplementary Digital Content, Figure 1, http://links.lww.com/APJO/A103). The widefield SLO images were taken with an additional attachment to the camera to increase the field of view to 163 degrees. Both the blue and synthesized color SLO images were analyzed. Fluorescein angiograms taken with a conventional fundus camera (KOWA Vx10-i, Tokyo, Japan) or with a SLO camera (HRA Spectralis, Heidelberg, Germany or NIDEK Mirante, Aichi, Japan) were also evaluated when available. In some cases, the OCT images were obtained with an ultra-widefield OCT device (Canon Xephilio, Kyoto, Japan). The identification and grading of the images were made by two of the authors (SH, KOM) independently. The graders were masked to the background information of the patients of each image. Only images with consistent results from the two graders of the existence and the presence or absence of hyporeflective areas were used for the statistical analyses.
Unpaired, nonparametric Kruskal-Wallis tests were used to determine the significance of the differences in the age, refractive error, and best-corrected visual acuity (BCVA) among the eyes with PDR, moderate to severe NPDR, and less than mild NPDR. Chi-square tests were used to determine the significance of the differences in the sex distribution. The concordance between the hyporeflective areas in the blue widefield SLO images and the NPAs observed in the FA images was determined by calculating Cohen's kappa coefficient (κ). We considered a κ value >0.8 as almost perfect concordance, and between 0.6 and 0.8 as substantial concordance. A P value of <0.05 was considered statistically significant. Comparison between those images was not quantitative. The cases analyzed were not consecutive, and the sample size was not optimized. Statistical analyses were performed using GraphPad Prism (GraphPad Software. Inc. Ver. 6.0).
Demographics of Patients (Table 1)
The medical records of 177 eyes of 90 DM patients were reviewed; 5 patients had type 1 DM and 85 patients had type 2 DM. In the DM group, 94 eyes of 51 patients (53.1%) had PDR, 53 eyes of 32 patients (29.9%) had moderate to severe NPDR, and 30 eyes of 16 patients (16.9%) had less than mild NPDR. Nine patients had signs of a different stage in one eye from that of the other one. Ninety-eight of the 177 eyes (55.4%) had received panretinal retinal photocoagulation (PRP) or local photocoagulation; 63 eyes with PDR, 35 eyes with moderate to severe NPDR, and none with less than mild NPDR. Fifty-three of the 177 eyes (29.9%) had received anti-VEGF therapies, and 33 of the 94 PDR eyes (35.1%) had undergone pars plana vitrectomy.
Table 1 -
Basic Characteristics of the Diabetic Patients and Non-diabetic Controls
||Moderate to severe NPDR
||Less than mild NPDR
P value (all groups)
|Number of eyes (patients)
|Age (mean ± years, range)
||63.0 ± 12.3 (35–85)
||59.5 ± 11.6 (35–79)
||67.3 ± 10.8 (47–84)
||67.3 ± 13.6 (38–85)
|Spherical equivalent (D, mean ± SD)
||-1.6 ± 2.6
||-1.7 ± 2.3
||-1.1 ± 3.0
||-2.0 ± 3.1
|BCVA (logMAR, mean ± SD)
||0.29 ± 0.40
||0.32 ± 0.43
||0.29 ± 0.33
||0.02 ± 0.17
BCVA indicates best-corrected visual acuity; D, diopter; DM, diabetes mellitus; logMAR, logarithm of the minimum angle of resolution; NPDR, nonproliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy.
†Kruskal-Wallis test; N.S., no significance.
Comparisons of the demographics of the DM patients with PDR, moderate to severe NPDR, and less than mild NPDR are shown in Table 1. The differences in the sex distributions and the mean refractive errors were not significant (Chi-square test for the sex, Kruskal-Wallis test for the mean refractive error). The differences in the age and mean BCVA in each DM groups were significant (P < 0.05 and P < 0.001 in Kruskal-Wallis test), and PDR patients were younger and had poorer BCVA. The average level of serum HbA1c in the DM group was 7.5 ± 1.5% (normal level, 4–5.6%) in 76 cases.
Presence of Hyporeflective Areas in Blue Widefield SLO Images in Eyes with Advanced Diabetic Retinopathy (DR)
In patients with no fundus lesions, a uniform background was observed in the blue widefield blue images (Supplementary Digital Content, Figure 1, http://links.lww.com/APJO/A103). On the other hand, hyporeflective areas were present in the blue widefield SLO images in 98 of the 177 eyes of the DM patients (55.4%; Table 2, Figs. 1–5). A hyporeflective area was recognized as an area of darker appearance than the surrounding areas. In some areas, the retinal vessels crossing the hyporeflective area were seen as white lines (Figs. 1–5). Hyporeflective areas were found in 72 of 94 eyes with PDR (76.6%), in 26 of the 53 eyes (49.1%) with moderate to severe NPDR, and none of the 30 eyes (0%) with less than mild NPDR. The hyporeflective areas were either localized or diffused. Among the 72 eyes with PDR that had hyporeflective areas, 63 eyes (87.5%) had a history of PRP. In these 72 eyes, the hyporeflective areas were observed both in eyes with and without prior PRP (Figs. 1–5).
Table 2 -
Frequency of Hyporeflective Area in Wide Field Blue SLO
||Numbers of eyes with hyporeflective area in SLO/total numbers of SLO-examined eyes (%)
| Moderate to severe NPDR
| Less than mild NPDR
| DM total
DM indicates diabetes mellitus; NPDR, nonproliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; SLO, scanning laser ophthalmoscopy.
Concordance between of Hyporeflective Areas in Blue Widefield SLO Images and Non-Perfused Areas (NPAs) in Fluorescein Angiograms (FA)
FA images were available for 126 eyes in the DM patients. In the DM patients, 70 of the 94 eyes with PDR had FA images, 46 of the 53 eyes with moderate to severe NPDR had FA images, and 10 of the 30 eyes with less than mild NPDR had FA images.
Among the eyes examined by FA, NPAs were found in 65 eyes with PDR, in 27 eyes with moderate to severe NPDR. In these cases, a concordance in the identification rate and location between the hyporeflective area in the blue SLO images and the NPAs in the FA images was evaluated (Table 3). Among the 65 eyes with PDR in which NPAs were found in the FA images, hyporeflective areas in the blue SLO images were observed in 52 eyes (52/65, 80.0% sensitivity). Similarly, in 27 eyes with moderate to severe NPDR and with NPAs in the FA images, the hyporeflective areas in the blue SLO images were found in 21 eyes (21/27, 77.8% sensitivity). Furthermore, in the cases without previous retinal photocoagulation, all 12 DR eyes with NPAs in the FA had corresponding hyporeflective areas in blue SLO. The locations of the hyporeflective areas were found within the location of the NPAs in the FA images (Figs. 1–5). The shape of the hyporeflective areas found in blue SLO were not identical to that of the NPAs in the FA but appeared relatively smaller than the NPAs (Figs. 1–3). In contrast, 13 of the 65 eyes (19.7%) with PDR and with NPAs in the FA images and 6 of the 27 eyes (22.2%) with moderate to severe NPDR and with NPAs did not have hyporeflective areas in the blue SLO images. In these eyes, the SLO images were obtained after photocoagulation, while the FA images were obtained before the treatment. Thus, the chorioretinal scars of photocoagulations appeared to interfere with finding the dark areas, with multiple scars indicating hyperreflection.
Table 3 -
Concordance Between the Numbers of Hyporeflective Area in Blue SLO and NPAs in FA
||SLO-Hypo (+) eyes / FA-NPA (+) eyes (%)
||SLO-Hypo (−) eyes / FA-NPA (−) eyes
|Moderate to severe NPDR
|Less than mild NPDR
Number of observed agreements: 107 (84.92% of the observations). Number of agreements expected by chance: 67.6 (53.65% of the observations). Kappa = 0.675. 95% confidence interval: 0.547–0.802. The strength of agreement is substantial.NPA indicates non-perfused areas; NPDR, nonproliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; SLO, scanning laser ophthalmoscopy.
The degree of concordance between the identification rate of hyporeflective areas in the widefield blue SLO and the NPAs in the FA images was determined by calculating Cohen's kappa coefficient (κ). The mean κ value was 0.675 with a range of 0.6 and 0.8. We conclude that there was good concordance (Table 3). Furthermore, hyporeflective areas were not found in the blue SLO images, when NPAs in the FA images were not found, in all 34 eyes for a specificity of 100%. Thus, the concordance relationship between the widefield blue SLO images and the NPAs in the FA images was substantial.
Thinning and Disorganization of Retina in Hyporeflective Areas in Blue Widefield SLO Images
The ultra-widefield OCT images were examined to determine the morphological changes of the retina in the hyporeflective areas in the blue SLO images. There were images from both widefield SLO images and corresponding OCT images of 3 eyes in 2 cases. In these 3 eyes, the area of dark findings in the blue SLO images corresponded to the NPAs detected in the FA images. In the ultra-widefield OCT images, a thinning of the entire retina with a partial disorganization of the inner retinal layer was observed corresponding to the hyporeflective areas in the SLO blue images (Figs. 4 and 5 and Supplementary Digital Content, Fig. 2, http://links.lww.com/APJO/A103).
The blue widefield images in a high percentage of patients with PDR had hyporefletive areas. Among the 92 eyes with NPAs in the FA images due to PDR or moderate to severe NPDR, 52 eyes with PDR and 21 eyes with moderate to severe NPDR had hyporeflective areas in the blue SLO images (79.3%). A good concordance of hyporeflective areas in the blue SLO and the NPAs in the FA images was confirmed by a kappa of 0.675. In addition, a direct comparison between the FA and blue SLO images showed that the location of the hyporeflective areas seen in the blue SLO images were also found within the location of the NPAs identified in the FA images in patients with PDR and moderate to severe NPDR.
The concordance between the hyporeflective areas found in the blue SLO images and NPAs in the FA was comparable to the findings of Shin et al.20 However, our findings were made in almost the entire fundus while the study of Shin et al a relatively narrower field. Our findings should be especially helpful in the management of DR because the ischemic retinal areas in DR are not limited to the posterior pole but throughout fundus including the periphery. Additionally, we examined the retinal structure in the ultra-widefield OCT images corresponding to the hyporeflective areas in blue SLO. We found morphological change including a thinning and disorganization of the retina indicating ischemic degeneration of the retina.
The widefield images obtained by multicolor SLO (Mirante) extend to the equator (163 degrees field of view), which allowed us to determine the presence of retinal ischemia in a single image. Compared to the widefield OCTA images, the blue SLO images can be easily obtained with a single shot as in the conventional fundus photographs. Blue, green, and red images are instantly recorded separately. Thus, once the multicolor images are obtained, the detection of hyporeflective areas in the blue images is possible.
Despite the high concordance between the number and locations of the hyporeflective areas and the NPAs, approximately 20% of DR eyes with NPAs did not have hyporeflective areas in the blue SLO images. One of the reasons for this was that the SLO images were not taken on or around the same day as the FA examinations in some eyes. In most cases, the SLO images were taken after targeted photocoagulation of the NPAs or after PRP, while the FA images were taken before the photocoagulation. The chorioretinal scars of the photocoagulations were occasionally seen as hyperreflective areas in the blue SLO images, and this brightness often makes it difficult to detect the hyporeflective areas in the blue SLO images especially in the areas with large number of photocoagulation scars. Our findings showed that all 12 DR eyes without previous retinal photocoagulation, and with NPAs in FA had corresponding hyporeflective areas in the blue SLO images. In some cases, media opacities such as mild vitreous hemorrhages that developed after the FA examination reduced the resolution of the SLO images.
Recent advances in OCT technology have allowed recordings that can reveal a disorganization of the inner retinal layer in the areas where the NPAs are located in eyes with DR.22,23 In addition, a thinning of the inner retina in the areas of the NPAs has been reported.23 Despite the limited number of cases examined, the ultra-widefield OCT images showed the thinning and disorganization of the inner retina in the areas where the NPAs were located. The shapes and size of the hyporeflective areas found in the blue SLO images were not identical to the corresponding NPAs in the FA, but corresponded in location (Figs. 1–3). Thus, it is highly likely that the hyporeflective areas in the blue SLO images represent morphological changes of the retina caused by non-perfusion of the retinal capillaries.
The exact pathology causing the hyporeflective blue SLO images was not definitively determined. An earlier study reported that the blue wavelength SLO images can detect nerve fiber layer defects.18 However, the hyporeflective areas in the blue SLO images did not have the typical form of a nerve fiber defect, but they were observed as block figures divided by retinal vessels. Shin et al20 hypothesized that a reduced amount of hemoglobin secondary to a decrease in blood flow in the ischemic areas and relative high absorption of the blue wavelength by the retinal pigment epithelium explained the hyporeflection. Our ultra-widefield OCT results showed retinal thinning and disorganization in the ischemic area. Thus, we suggest that the blue wavelength possibly passed more easily through the retinal layers, which were lacking normal amount of neural or glial components secondary to retinal ischemia. It is also interesting if other blue filter or red-free images by conventional fundus cameras or ophthalmoscopes can detect similar findings and further investigations are anticipated.
This study has several limitations. This was not a consecutive case series and the number of cases was not optimized for the statistics used. The study was conducted at a single tertiary eye center thus the results may not be applied to the general DR population. Another limitation was that the SLO and FA images were not always taken on the same day and different types of FA images, ie, widefield and conventional images, were mixed. This made it difficult to perform the quantitative area analysis and the comparison between both images by Bland-Altman Plots, although this would be expected in future analysis. Cases with dense photocoagulation made it difficult to detect the hyporeflective area in blue SLO images, which possibly resulted in a relatively lower detection rate than in the FA images, which were taken before photocoagulation. It would have been better if the cases without prior laser were used but we included eyes with laser scars to obtain sufficient numbers of patients. Furthermore, the number of PDR eyes was significantly greater than the number of NPDR eyes, which may have affected the results of Cohen's kappa.
In conclusion, these results indicate that widefield blue SLO can be used as a non-invasive and easy method to detect NPAs, and thus determine the severity of DR. This technique of detecting NPAs can be used as a practical clinical management and screenings for DR.24,25
The authors thank Professor Emeritus Duco Hamasaki of the Bascom Palmer Eye Institute for discussions and corrections of this manuscript. SH and TY belong to the department of Advanced Ophthalmic Imaging in Tokyo Medical and Dental University, which is funded by NIDEK Corporation. There is no other conflict of interest. The other authors: NK, KK, TIY, KOM have no competing interest related to this submission.
1. Yau JWY, Rogers SL, Kawasaki R, et al. Global prevalence and major risk factors of diabetic retinopathy. Diabetes Care
2. Flaxman SR, Bourne RRA, Resnikoff S, et al. Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis. Lancet Glob Health
3. Zhang X, Saaddine JB, Chou C-F, et al. Prevalence of diabetic retinopathy in the United States, 2005-2008. JAMA
4. Sabanayagam C, Yip W, Ting DSW, et al. Ten emerging trends in the epidemiology of diabetic retinopathy. Ophthalmic Epidemiol
5. Wang LZ, Cheung CY, Tapp RJ, et al. Availability and variability in guidelines on diabetic retinopathy screening in Asian countries. Br J Ophthalmol
6. Ruta LM, Magliano DJ, Lemesurier R, et al. Prevalence of diabetic retinopathy in Type 2 diabetes in developing and developed countries. Diabet Med
7. Guariguata L, Whiting DR, Hambleton I, et al. Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res Clin Pract
8. Kume A, Kashiwagi K. Recent epidemiological status of ocular and other major complications related to diabetes mellitus in Japan. Ophthalmologica
9. Miyazaki M, Kubo M, Kiyohara Y, et al. Comparison of diagnostic methods for diabetes mellitus based on prevalence of retinopathy in a Japanese population: the Hisayama Study. Diabetologia
10. Adamiec-Mroczek J, Oficjalska-Młyńczak J. Assessment of selected adhesion molecule and proinflammatory cytokine levels in the vitreous body of patients with type 2 diabetes--role of the inflammatory-immune process in the pathogenesis of proliferative diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol
11. Nawaz IM, Rezzola S, Cancarini A, et al. Human vitreous in proliferative diabetic retinopathy: characterization and translational implications. Prog Retin Eye Res
12. Mustafi D, Saraf SS, Shang Q, et al. New developments in angiography for the diagnosis and management of diabetic retinopathy. Diabetes Res Clin Pract
13. Nicholson L, Ramu J, Chan EW, et al. Retinal nonperfusion characteristics on ultra-widefield angiography in eyes with severe nonproliferative diabetic retinopathy and proliferative diabetic retinopathy. JAMA Ophthalmol
14. Tey KY, Teo K, Tan ACS, et al. Optical coherence tomography angiography in diabetic retinopathy: a review of current applications. Eye Vis (Lond)
15. Ashraf M, Sampani K, Clermont A, et al. Vascular density of deep, intermediate and superficial vascular plexuses are differentially affected by diabetic retinopathy severity. Invest Ophthalmol Vis Sci
16. Samara WA, Shahlaee A, Adam MK, et al. Quantification of diabetic macular ischemia using optical coherence tomography angiography and its relationship with visual acuity. Ophthalmology
17. Sawada O, Ichiyama Y, Obata S, et al. Comparison between wide-angle OCT angiography and ultra-wide field fluorescein angiography for detecting non-perfusion areas and retinal neovascularization in eyes with diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol
18. Terasaki H, Sonoda S, Kakiuchi N, et al. Ability of MultiColor scanning laser ophthalmoscope to detect non-glaucomatous retinal nerve fiber layer defects in eyes with retinal diseases. BMC Ophthalmol
19. Terasaki H, Sonoda S, Shiihara H, et al. More effective screening for epiretinal membranes with multicolor scanning laser ophthalmoscope than with color fundus photographs. Retina (Philadelphia, Pa)
20. Shin YU, Lee BR, Kim S, et al. A novel noninvasive detection method for retinal nonperfusion using confocal red-free imaging. Ophthalmology
21. Wilkinson CP, Ferris FL, Klein RE, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales
22. Nicholson L, Ramu J, Triantafyllopoulou I, et al. Diagnostic accuracy of disorganization of the retinal inner layers in detecting macular capillary non-perfusion in diabetic retinopathy. Clin Exp Ophthalmol
23. Moein H-R, Novais EA, Rebhun CB, et al. Optical coherence tomography angiography to detect macular capillary ischemia in patients with inner retinal changes after resolved diabetic macular edema. Retina (Philadelphia, Pa)
24. Rajalakshmi R, Subashini R, Anjana RM, et al. Automated diabetic retinopathy detection in smartphone-based fundus photography using artificial intelligence. Eye (Lond)
25. Rajalakshmi R, Arulmalar S, Usha M, et al. Validation of Smartphone Based Retinal Photography for Diabetic Retinopathy Screening. PLoS One