Abstract
Purpose
To examine the relationship between ischemia and disorganization of the retinal inner layers (DRIL).
Methods
Cross-sectional retrospective study of 20 patients (22 eyes) with diabetic retinopathy presenting to a tertiary academic referral center, who had DRIL on structural optical coherence tomography (OCT) using Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany) and OCT angiography (OCTA) with XR Avanti (Optovue Inc., Fremont, California, USA) on the same day. OCTA images were further processed to remove flow signal projection artifacts using a software algorithm adapted from recent studies. Retinal capillary perfusion in the superficial (SCP), middle (MCP), and deep capillary plexuses (DCP), as well as integrity of the photoreceptor lines on OCT were compared in areas with DRIL to control areas without DRIL in the same eye.
Results
Qualitative assessment of projection-resolved OCTA of eyes with DRIL on structural OCT demonstrated significant perfusion deficits compared to adjacent control areas (p<0.001). The majority of lesions (85.7%) showed superimposed SCP and/or MCP non-perfusion in addition to DCP non-flow. Areas of DRIL were significantly associated with photoreceptor disruption (p=0.035), compared to adjacent DRIL-free areas.
Conclusion
We found that DRIL is associated with multi-level retinal capillary non-perfusion, suggesting an important role for ischemia in this OCT phenotype.
Keywords: Projection-resolved optical coherence tomography angiography, Diabetic retinopathy, Disorganization of the retinal inner layers, DRIL, OCT, Retina
Introduction
Disorganization of the retinal inner layers (DRIL) was first described by Sun et al. as an optical coherence tomographic (OCT) biomarker and important predictor of vision in eyes with diabetic macular edema (DME).1 DRIL is seen on OCT as the absence of discernable boundaries between the ganglion cell-inner plexiform layer (GC-IPL) complex, inner nuclear layer (INL), and outer plexiform layer (OPL). Several studies have demonstrated that DRIL is a dynamic phenomenon.2,3 In addition, the extent of DRIL has been shown to be strongly predictive of future visual acuity outcomes, with each 300 μm increase in the lateral extent of DRIL during a four-month follow-up period predicting a one-line decrease in vision at eight months.1 DRIL due to conditions other than DME has also been shown to be a strong predictor of visual acuity.4,5
The exact pathogenesis of DRIL is still debated. In their cohort of eyes with DRIL and DME, Sun et al. proposed that the pathophysiology of DRIL could be related to mechanical stretching of neuronal elements in the setting of edema. Nicholson et al. reported that DRIL was significantly associated with capillary non-perfusion on fluorescein angiography (FA).6 While the findings of Nicholson et al. suggest that DRIL could be related to ischemia, their study design was based on qualitative assessment of FA, first identifying areas of capillary non-perfusion on FA and then examining the prevalence of DRIL on OCT in these areas, with the potential for missing areas of DRIL that were not associated with non-perfusion. In addition, while FA only reveals the superficial capillaries, the introduction of OCT angiography (OCTA) has allowed researchers to study the three-dimensional structure of the retinal vascular plexuses. Moein et al.7 have recently utilized OCTA to identify an increased foveal avascular zone (FAZ) as well as significant loss of the retinal capillary plexuses in the superficial, deep, and full retina in eyes with DRIL compared to healthy controls.
While the evidence in support of the role of ischemia in DRIL is mounting, the exact capillary level(s) responsible deserve further study. OCTA imaging is now able to distinguish not just two, but three distinct retinal vascular plexuses in the macula; the superficial (SCP), middle (MCP), and deep capillary plexuses (DCP).8 Moreover, recent software advances have provided a solution for one of the inherent limitations of OCTA software, projection artifact, which is cast by flow in the superficial retinal circulation onto the deeper layers. Projection artifact can confound capillary measurements at the deeper vasculature. Commercial OCTA software has typically addressed projection artifact with a slab-subtraction algorithm, which replaces flow projection artifact with an equally undesirable shadow (negative) artifact.9 Recently, Zhang et al. introduced the projection-resolved (PR)-OCTA algorithm, which overcomes many of the drawbacks of the subtraction algorithm. This algorithm preserves the integrity of the deeper capillary layers and removes projection artifact, allowing improved visualization of the three capillary plexuses.9,10 The goals of the current study were twofold; to capitalize on the PR-OCTA algorithm to better define the relationship between ischemia and DRIL in eyes with diabetic retinopathy, and to further characterize the specific macular capillary plexus (SCP, MCP, and/or DCP) associated with this OCT biomarker.
Methods
This study was a cross-sectional analysis of patients with diabetic retinopathy who underwent structural OCT and OCTA imaging in the Department of Ophthalmology at Northwestern University in Chicago, Illinois between June 2015 and December 2016. The study was approved by the Institutional Review Board of Northwestern University, followed the tenets of the Declaration of Helsinki, and was performed in accordance with Health Insurance Portability and Accountability Act regulations. Written informed consent was obtained from all participants.
Inclusion criteria for this study included eyes with diabetic retinopathy and evidence of DRIL on OCT. DRIL was defined as the presence of poorly defined or absent boundaries between the GC-IPL complex, INL, and OPL on OCT.1 We excluded eyes with other retinal diseases that might contribute to retinal non-perfusion including arterial or venous occlusions. In addition, we excluded eyes where the OCTA images had movement/shadow artifacts in the area of interest and those with OCTA signal strength score lower than 50.
Image acquisition
To identify areas of DRIL, we reviewed OCT images obtained on the Spectralis HRA+OCT system (Heidelberg Engineering, Heidelberg, Germany) with an average automatic real time (ART) of 16. We analyzed B-scans passing through the foveal center, as well as 5 - 7 total scans within a 1 mm-wide area centered on the foveal depression to identify areas of DRIL, similar to the method used by Sun et al.1 Areas of DRIL had to be present on one or more B-scans to be included in the DRIL group. Two readers, masked to the OCTA findings, independently graded the presence of DRIL (A.C.O. and A.A.F.). Following independent assessment, disagreements were discussed until consensus was reached. For each eye, the readers identified the control area, as an equivalent retinal region free of DRIL in the same 1×1mm zone.
To study retinal capillary perfusion status, we used the RTVue-XR Avanti OCTA system (Optovue Inc., Fremont, California, USA) with split-spectrum amplitude-decorrelation angiography (SSADA) software.11 This instrument has an A-scan rate of 70,000 scans per second and uses a light source centered on 840 nm and a FWHM bandwidth of 45 nm. A 3 × 3-mm2 scanning area, centered on the fovea was obtained. Two consecutive B-scans (M-B frames), each containing 304 A-scans were captured at each sampling location and SSADA was used to extract OCTA information.
Projection-resolved OCTA
OCTA angiograms were imported into a custom MATLAB program (MathWorks, Inc, Natick, Massachusetts, USA) to remove flow projection artifact. We implemented a version of the PR-OCTA algorithm developed by Zhang et al.9 This algorithm has been used to study the three capillary plexuses of the inner macular retina.12 The authors noted that the OCTA projection tail artifacts have lower decorrelation values than real vessels. Taking advantage of this observation, the algorithm searches each A-scan for successive high-valued peaks, which correspond to real vessels. The OCTA values at the peak positions are kept, while the remaining pixels in the A-scan are set to zero, resulting in the removal of projection artifact. An empirically determined parameter, α, is employed to control the search of the peak positions. When a larger a value is used, the peak positions must also be larger, thereby removing more of the projection artifact. In this study, we used an α value of 0.9 for all study eyes (Figure 1).
Figure 1.
Representative optical coherence tomography angiography (OCTA) B-scan with red flow signal overlay for blood flow (A). Note the high intensity of the signal as well as the significant tail artifact, where blood vessels appear vertically elongated and occupy multiple retinal layers. Projection-resolved OCTA B-scan (B) with white lines approximating boundaries between the superficial (SCP), middle (MCP), and deep capillary plexuses (DCP). Flow is present in all three layers with improvement in most, but not all, of the flow artifact present in (A).
Image Analysis
In general, control areas were chosen from the same OCT B-scans showing DRIL. Control areas were equivalent in size to the corresponding areas of DRIL, but on the other side of, and equidistant from the fovea, as the DRIL lesions (Figure 2). This equidistant area selection was emphasized as control areas more peripherally located relative to areas of DRIL could potentially skew study findings given that the FAZ is contained within the 1×1 mm zone of interest. For 3 out of 28 areas of DRIL that were present on both sides of the fovea, control areas were selected in B-scans located an equal number of scans inferior or superior from the horizontal meridian opposite the DRIL area. Since the control areas were selected on structural OCT scans, the reader was effectively agnostic to their perfusion status on PR-OCTA.
Figure 2.
Representative spectral-domain optical coherence tomography (SD-OCT) image displaying disorganization of the retinal inner layers (DRIL; blue bracket in B) at the left-hand side of the fovea (A). Retinal layers are labeled adjacent to the DRIL (B). Control area (white bracket) is selected on the opposite side of the fovea to DRIL as an equal-sized zone with intact inner retinal layer boundaries. Projection-resolved OCT angiography B-scan (C) at the location of the structural B-scan in (A). The red signal overlay denotes blood flow. Flow deficits are observed throughout the superficial (star), middle (asterisk), and deep capillary plexuses (triangle) in the area of DRIL, whereas the perfusion is relatively preserved in all three retinal capillary plexuses in the control area (D). Similar findings are observed in OCT angiography without projection-resolution (E and F).
Abbreviations: Ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL).
Areas of DRIL and their corresponding control areas seen on structural B-scans were identified on the corresponding B-scans on the OCTA system using unique structural features (e.g. foveal pit contour, mild macular edema). We assessed the flow signal at the three different capillary levels in areas of DRIL and compared them to the adjacent DRIL-free control areas.
Five of twenty-two eyes (22.7%) had more than one distinct area of DRIL present on -OCT. In total, there were twenty-eight DRIL lesions that were individually evaluated on PR-OCTA for capillary involvement. Due to errors in automated segmentation as a result of disrupted boundaries between the retinal layers in DRIL, the qualitative presence or absence of flow in the various capillary plexuses was assessed on the OCTA B-scans. The SCP flow was defined as the flow within the inner 80% of “the GC-IPL complex” which includes the nerve fiber (NFL), ganglion cell layer (GCL), and IPL. Flow signal present at the outer 20% of the GC-IPL complex, as well as the inner 50% of the INL was interpreted as the MCP. Finally, DCP flow was defined as flow at the outer 50% of the INL to the outer edge of the OPL (Figure 1). For those areas of DRIL present on more than one B-scan (17 of 28 areas of DRIL), we used the OCTA B-scans with non-perfusion in the maximum number of capillary plexuses to evaluate the entire lesion.
Photoreceptors
We assessed photoreceptor integrity through qualitative evaluation of the photoreceptor layers: the inner segment/outer segment (IS/OS) junction and cone outer segment tip (COST) lines in areas underlying DRIL as well as the corresponding control areas.
Statistics
We compared the presence of ischemia at the SCP, MCP, or DCP in areas with and without DRIL using Pearson’s chi-squared test in Microsoft Excel (Microsoft Excel 15.32, Microsoft Corporation, Redmond, WA, USA). Chi-squared analysis was also used to compare areas of photoreceptor disruption between DRIL and control areas. In this study, p < 0.05 was deemed statistically significant.
Results
We identified 107 patients (176 eyes) with diabetic retinopathy, of whom 20 (22 eyes) had evidence of DRIL on OCT. There were eight men and twelve women, with an average age of 55.4 ± 14.2 years. Of these eyes, the majority had proliferative diabetic retinopathy (14 of 22; 64%) with the remainder (8 of 22; 36%) having non-proliferative retinopathy. Two patients had bilateral DRIL on OCT. Nine eyes had either a previous history of treatment for diabetic macular edema or concurrent macular edema present at the time of imaging (Table 1).
Table 1.
Demographic Characteristics of Eyes with Disorganization of the Retinal Inner Layers (DRIL)
| Demographics | Patients/Eyes |
|---|---|
| Mean Age (±SD) | 55.4 ± 14.2 |
| Sex (M:F) | 8:12 |
| Eyes with NPDR:PDR | 8:14 |
| Eyes with previous panretinal photocoagulation | 8/22 |
| Eyes with previous anti-VEGF therapy | 4/22 |
| Eyes with previous focal macular laser | 2/22 |
| Eyes with current/history of macular edema | 9/22 |
Abbreviations: NPDR = non-proliferative diabetic retinopathy, PDR = proliferative diabetic retinopathy, VEGF = vascular endothelial growth factor
A total of 28 areas of DRIL and 28 control areas were identified on structural OCT (Figure 2, Table 2). Following registration of structural B-scans with corresponding PR-OCTA B-scans, we analyzed the flow signals in the SCP, MCP, and DCP and compared perfusion within DRIL regions to perfusion in the adjacent control areas (Figure 2).
Table 2.
Comparison of Qualitative Capillary Plexus Perfusion in Disorganization of the Retinal Inner Layers (DRIL) and Control Areas
| Layers of Non-Perfusion | DRIL Lesions (n=28) | Control Areas (n=28) |
|---|---|---|
| No Defect | 0 | 21 |
| Isolated SCP | 1 | 1 |
| Isolated MCP | 0 | 1 |
| Isolated DCP | 3 | 5 |
| SCP + MCP | 0 | 0 |
| SCP + DCP | 8 | 0 |
| MCP + DCP | 6 | 0 |
| All Layers Affected | 10 | 0 |
|
| ||
| Chi-square test: p < 0.001 | ||
Abbreviations: SCP = superficial capillary plexus, MCP = middle capillary plexus, DCP = deep capillary plexus
Overall, all 28 DRIL lesions had perfusion defects evident on PR-OCTA, while the majority of control areas (21 of 28; 75.0%) showed no evidence of perfusion deficits (Table 2). Of the 7 control areas with impaired flow, 5 showed isolated DCP defects (71.4%). Only one DRIL lesion had pure SCP deficit, while 27 of the 28 DRIL lesions (96.4%) had a component of DCP flow deficit. In 24 of these 27 lesions (88.9%), there was additional flow impairment in the MCP and/or SCP (Figures 2, 3). A chi-square test showed a statistically significant overall p value < 0.001 for the comparison between DRIL and control perfusion (Table 2).
Figure 3.
Spectral-domain optical coherence tomography (SD-OCT) in (A-C) displaying disorganization of the retinal inner layers (DRIL; blue bracket) associated with non-perfusion at various capillaries plexus levels on projection-resolved OCTA angiography (D-F). In (D), the control area (white bracket) is associated with apparent middle capillary plexus non-perfusion (asterisk), while the area of DRIL is associated with non-perfusion at the level of the superficial (star) and deep capillary plexus (triangle). OCT angiography without projection-resolution in (G, H, and I) with apparent discrepancies in (H) compared to (E) but also significant tail artifact in general. Perfusion findings in (G) and (I) are grossly consistent with those in (D) and (F) respectively.
Note the disruption in the inner segment/outer segment and cone outer segment tip lines under the control areas in (B, E, and H) and DRIL areas in (C, F, and I) compared to the intact photoreceptor lines in (A, D, and G).
Integrity of the photoreceptors was assessed at each control and DRIL area (Figure 3, Table 3). Eleven of twenty-eight DRIL lesions (39.3%) had disruption of the photoreceptors compared to four of twenty-eight control areas (14.3%), which was significant (p = 0.035). Three of the eleven DRIL areas with photoreceptor abnormalities had isolated DCP flow involvement, with the remaining eight lesions having a combination of DCP and MCP involvement or gross non-perfusion of all three capillary plexuses. Of the four control areas with photoreceptor disruption, two had isolated non-perfusion of the DCP, whereas the other two had no obvious flow abnormalities (Figure 3).
Table 3.
Photoreceptor Disruption in Areas of Disorganization of the Retinal Inner Layers (DRIL)
| Photoreceptor Status | DRIL Lesions | Control Areas |
|---|---|---|
| Disruption | 11 | 4 |
| Intact | 17 | 24 |
|
| ||
| Chi-square test: p = 0.035 | ||
Abbreviations: SCP = superficial capillary plexus, MCP = middle capillary plexus, DCP = deep capillary plexus
Discussion
In this study, we used PR-OCTA to examine the role of capillary non-perfusion in the pathogenesis of DRIL in patients with DR. All twenty-eight DRIL lesions were associated with flow defects evident on imaging; the majority involved the DCP with superimposed involvement of either the MCP, SCP, or both (24/28; 85.7%).
Nicholson et al. previously examined the question of ischemia in DRIL using FA.6 These researchers studied eyes undergoing treatment for DME and showed that macular non-perfusion on FA was a reliable predictor for areas of DRIL on OCT (84.4%).6 Conversely, Moein et al. used OCTA to show that eyes with DRIL had a significantly larger FAZ at all retinal sublayers compared to healthy eyes.7 We used a different methodology compared to both of these studies. Nicholson et al. first identified areas of capillary non-perfusion on FA and then examined those areas for evidence of DRIL on OCT, an approach that has the potential to overlook areas of DRIL with normal perfusion. Their control regions were chosen based on having normal perfusion on FA, which also may have biased their results. Furthermore, FA is not able to identify non-perfusion at the deeper capillary plexuses, which is an advantage of OCTA. In contrast, Moein et al. examined a single, highly-averaged foveal structural OCT scan to identify eyes with DRIL before correlating them with FAZ areas derived on OCTA, and compared these to healthy control eyes. Contrary to these authors’ methods, our approach involved examining the structural OCT within a 1×1 mm area in order to identify all the areas of DRIL within this volume, which increased the likelihood that we would capture all potential areas of DRIL in each eye. The control retinal areas in our study were then chosen in the same eyes on the structural OCT by the same reader (masked to the OCTA and perfusion status), based on the absence of DRIL and the distance from the fovea. Only 7 of 28 (25.0%) control areas compared to 28 of 28 (100.0%) areas of DRIL showed evidence of perfusion impairment on OCTA. We therefore believe that our results provide more direct evidence of the association between DRIL and capillary ischemia and further support the role of capillary non-perfusion in the pathogenesis of DRIL.
Our goal was to capitalize on PR-OCTA, and use this approach to identify the specific retinal capillary layer(s) that contribute to the pathogenesis of DRIL. Due to loss of the inner retinal structural boundaries in areas of DRIL, we were unable to use automated segmentation algorithms or en face slabs, and instead qualitatively assessed blood flow in the respective capillary plexuses on cross-sectional OCTA. Previous studies from our group in eyes with diabetic retinopathy have shown that blood flow in the deeper capillary layers may be more profoundly decreased relative to the superficial retinal vasculature.13 In fact, our recent studies have revealed that flow index (a surrogate marker for blood flow) shows no significant decrease in the SCP with increasing severity of retinopathy compared to a significant decrease at the deeper capillaries.14 In this current study, 27 of 28 areas of DRIL displayed involvement of the DCP. However, only 3 of these 27 areas (11.1%) had isolated involvement of the DCP, as did 5 of 28 control areas (Table 2). The remaining DRIL lesions displayed additional involvement of either the SCP, MCP, or both capillary plexuses. Given these findings, we believe that DRIL may be a marker of extreme multi-layered capillary ischemic changes in the setting of DR. While cause-and-effect cannot be inferred from the current cross-sectional study design, we hypothesize that superimposed SCP and/or MCP ischemia, as well as potentially worsening baseline DCP non-perfusion, could predispose these eyes to inner retinal structural disorganization.
Relevant to this study, earlier reports from our group have shown that isolated DCP non-flow is associated with loss of integrity of the photoreceptor layer.15,16 Although the photoreceptor layers were grossly intact in a majority of areas of DRIL, 11 of 28 areas of DRIL (39.3%) showed disruption, which was a significantly higher proportion than in control areas (p = 0.035). It is possible that the observed non-perfusion in the deeper capillary layers may in part contribute to the previously observed relationship between DRIL and visual acuity through photoreceptor disruption.1 Interestingly, in their recent study, Moein et al.7 found significant correlations between lower visual acuity and larger FAZ areas in both superficial and full, but not deep, retinal slabs. This finding is difficult to explain and deserves further investigation, but could be due to a confounding effect from partial inclusion of the MCP within the DCP slab due to the segmentation protocol.
While the present study supports the hypothesis that ischemia plays an important role in the development of DRIL, there could be other pathophysiological factors. Sun et al. originally suggested a structural hypothesis for the pathogenesis of DRIL in which edema causes disruption of bipolar axons once their elasticity limit has been exceeded.3 It is possible that both mechanical and vascular factors could be contributing to the pathogenesis of DRIL in eyes with edema, although the relative importance of these factors is yet to be determined. Six of the twenty-two eyes used in this study did have mild macular edema on imaging. However, given the limited number of affected eyes, our reliance on cross-sectional rather than en face OCTA, and the mild nature of the edema in this cohort of eyes (Figure 3), we believe that our study results were not confounded by cystoid macular edema. An additional concern is that the apparent loss of perfusion observed in this study could be due to collapse of the capillary plexuses into one another. Given the retinal thinning that occurs in some of these eyes, we believe that this could be a reasonable possibility. However, it would, by default, suggest that there is relative loss of blood flow in all the relevant capillary layers.
Strengths of our study include the masked independent OCT grading, use of highly-averaged OCT scans to detect DRIL, and the PR-OCTA approach to remove flow artifact. In some cases, however, it may be difficult to discern whether the projection-resolution algorithm is not only removing projection artifact but also real flow. Other limitations of the current study include its cross-sectional nature, as well as the small number of eyes. DRIL is an OCT biomarker whose presence can be difficult to assess even on highly-averaged OCT B-scans as used in the current study. Our readers used a masked approach to identify and adjudicate the presence of DRIL to improve reliability. In addition, DRIL, by definition, obfuscates the retinal boundaries, making retinal layer segmentation using currently available software technology difficult. As a result, we were limited to qualitative assessment of flow in each capillary sublayer on cross-sectional scans as opposed to quantitative analysis, and we may have missed subtle defects in blood flow. Furthermore, the cross-sectional nature of the study does not allow us to examine the dynamic nature of DRIL and the potential contribution of capillary reperfusion to improvements in DRIL.
In conclusion, our results provide evidence that ischemia plays an important role in the pathogenesis of DRIL. In addition, we identified multi-level capillary non-perfusion in DRIL in a majority of eyes, with superimposed flow impairment of the inner retinal capillaries on top of DCP ischemia. Given that FA may not reveal the deeper capillaries,17 it is important that future prospective studies use OCTA in larger cohorts to help confirm our findings and further elucidate the pathogenesis of the dynamic remodeling that has been reported in DRIL.
Summary Statement.
Using projection-resolved optical coherence tomography angiography (PR-OCTA) in eyes with disorganization of the retinal inner layers (DRIL) on structural OCT, we show that these areas of DRIL are correlated with multi-level capillary ischemia.
Acknowledgments
Financial Support: This work was funded in part by NIH 1DP3DK108248 (AAF) and research instrument support by OptoVue, Inc. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Footnotes
Conflict of Interest: No conflict of interest exists for any author.
Proprietary Interest: The authors have no proprietary interest in the subject of this manuscript.
References
- 1.Sun JK, Lin MM, Lammer J, et al. Disorganization of the retinal inner layers as a predictor of visual acuity in eyes with center-involved diabetic macular edema. JAMA Ophthalmol. 2014;132(11):1309–1316. doi: 10.1001/jamaophthalmol.2014.2350. [DOI] [PubMed] [Google Scholar]
- 2.Radwan SH, Soliman AZ, Tokarev J, et al. Association of Disorganization of Retinal Inner Layers With Vision After Resolution of Center-Involved Diabetic Macular Edema. JAMA Ophthalmol. 2015;133(7):820–825. doi: 10.1001/jamaophthalmol.2015.0972. [DOI] [PubMed] [Google Scholar]
- 3.Sun JK, Radwan SH, Soliman AZ, et al. Neural Retinal Disorganization as a Robust Marker of Visual Acuity in Current and Resolved Diabetic Macular Edema. Diabetes. 2015;64(7):2560–2570. doi: 10.2337/db14-0782. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mimouni M, Segev O, Dori D, et al. Disorganization of the Retinal Inner Layers as a Predictor of Visual Acuity in Eyes with Macular Edema Secondary to Vein Occlusion. Am J Ophthalmol. 2017;182:160–167. doi: 10.1016/j.ajo.2017.08.005. 2017. [DOI] [PubMed] [Google Scholar]
- 5.Grewal DS, O’Sullivan ML, Kron M, Jaffe GJ. Association of Disorganization of Retinal Inner Layers with Visual Acuity in Eyes with Uveitic Cystoid Macular Edema. Am J Ophthalmol. 2017;177:116–125. doi: 10.1016/j.ajo.2017.02.017. [DOI] [PubMed] [Google Scholar]
- 6.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. 2015;43(8):735–741. doi: 10.1111/ceo.12557. [DOI] [PubMed] [Google Scholar]
- 7.Moein HR, 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. 2017;0(0):1–8. doi: 10.1097/IAE.0000000000001902. [DOI] [PubMed] [Google Scholar]
- 8.Park JJ, Soetikno BT, Fawzi AA. Characterization of the Middle Capillary Plexus Using Optical Coherence Tomography Angiography in Healthy and Diabetic Eyes. Retina. 2016;36(11):2039–2050. doi: 10.1097/IAE.0000000000001077. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhang M, Hwang TS, Campbell JP, et al. Projection-resolved optical coherence tomography angiography. Biomed Opt Express. 2016;7(3):816–828. doi: 10.1364/BOE.7.000816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Campbell JP, Zhang M, Hwang TS, et al. Detailed Vascular Anatomy of the Human Retina by Projection-Resolved Optical Coherence Tomography Angiography. Sci Rep. 2017;10(7):42201. doi: 10.1038/srep42201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitude-decorrelation angiography with optical coherence tomography. Opt Express. 2012;20(4):4710–4725. doi: 10.1364/OE.20.004710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hwang TS, Zhang M, Bhavsar K, et al. Visualization of 3 distinct retinal plexuses by projection-resolved optical coherence tomography angiography in diabetic retinopathy. JAMA Ophthalmol. 2016;134(12):1411–1419. doi: 10.1001/jamaophthalmol.2016.4272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Nesper PL, Roberts PK, Onishi AC, et al. Quantifying Microvascular Abnormalities With Increasing Severity of Diabetic Retinopathy Using Optical Coherence Tomography Angiography. Invest Ophthalmol Vis Sci. 2017;58(6):BIO307–BIO315. doi: 10.1167/iovs.17-21787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Onishi AC, Nesper PL, Roberts PK, et al. Importance of Considering the Middle Capillary Plexus on OCT Angiography in Diabetic Retinopathy. Invest Ophthalmol Vis Sci. doi: 10.1167/iovs.17-23304. In Review. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Scarinci F, Nesper PL, Fawzi AA. Deep Retinal Capillary Nonperfusion is Associated with Photoreceptor Disruption in Diabetic Macular Ischemia. Am J Ophthalmol. 2016;168:129–138. doi: 10.1016/j.ajo.2016.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nesper PL, Scarinci F, Fawzi AA. Adaptive Optics Reveals Photoreceptor Abnormalities in Diabetic Macular Ischemia. PLoS One. 2017;12(1):e0169926. doi: 10.1371/journal.pone.0169926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Weinhaus RS, Burke JM, Delori FC, Snodderly DM. Comparison of fluorescein angiography with microvascular anatomy of macaque retinas. Exp Eye Res. 1995;61(1):1–16. doi: 10.1016/s0014-4835(95)80053-0. [DOI] [PubMed] [Google Scholar]