Swept-source optical coherence tomography angiography (SS-OCTA) is a novel diagnostic tool enabling the presentation of en face images of retinal vessels at different retinal and choroidal layers. In this technique, the OCT signal is transformed to visualize flow. As in every imaging technique, artifacts may appear. Several of these were already described (eg, false-negative flow, false-positive flow, quilting defect, etc.)1 Their exact identification might improve our ability to correctly interpret the images we receive. The aim of the current study is to describe image artifacts that may occur in SS-OCTA imaging of full-thickness macular holes (FTMHs).
The main cause of FTMH formation is tangential vitreomacular traction. However, vitreous detachment is a normal aging process and macular holes do not form in all cases of posterior vitreous detachment. FTMH formation has also been reported in eyes with detached vitreous,2 or after vitrectomy.3 Macular holes develop in the fellow eyes of macular hole patients with a much higher frequency than in the general population (ie, in 1.2% to 28% of cases).4 Thus, it may be extrapolated that in the fovea morphology a predisposing factor may exist, which increases the risk of FTMH formation.
Thus, the second aim of the current study was to evaluate macular microcirculation with SS-OCTA in eyes with FTMHs before and after surgery and in their fellow eyes.
Patients and Methods
This is a retrospective, observational study. The OCTA database of our outpatient clinic was reviewed in order to assemble patients in whom idiopathic FTMH had been diagnosed before and/or after surgery. Data of their unaffected fellow eyes, if available, were also evaluated. The institutional review board approved the study. The research adhered to the tenets of the Declaration of Helsinki. The inclusion criteria were: idiopathic macular holes before or after surgery. Patients with opaque media, diabetic retinopathy, and vein occlusion in any eye were excluded.
Besides complete ophthalmic examination implemented by a retina specialist, SS-OCT and SS-OCTA (Triton; Topcon, Tokyo, Japan) were performed. In five cases, an additional spectral-domain OCTA (SD-OCTA) was executed (RTVue; Optovue, Fremont, CA) the same day. SS-OCTA operates at 1,050 nm and 100,000 A-scans per second to acquire OCT angiography volumes consisting of four repeated B-scans. The scan area of 3 mm × 3 mm (about 10° angle of a view), centered on the fovea, was used in this study. The device is also capable to perform 6 × 6 and 9 × 9 scans. The total OCT scan time lasts approximately four seconds if a 3 × 3 image is performed.
Two experienced examiners (ZM, JN) measured the minimum and maximum macular hole diameter in SS-OCT scans. In all patients in whom only postoperative SS-OCTA data were available, preoperative SS-OCT measurements of the size of macular hole were included into the analysis.
We analyzed the four following OCT angiography images: superficial retinal vessels, deep retinal vessels, retinal pigment epithelium, and choriocapillaris. We measured the size (diameter and area) of the foveal avascular zone (FAZ) in the superficial and deep retinal vessels layer.
Additionally, a measurement of the diameter and area of the FAZ in the superficial and deep retina vessels layer was performed in all examined eyes.
Vessel density was qualitatively analyzed by both authors and compared with healthy eyes.
Statistical analysis was performed using SigmaStat 11 for Windows (Microsoft, Redmond, WA). Correlations were calculated with the Pearson Product Moment and Spearman Rank Order tests. The Mann- Whitney test was applied to compare different groups.
We examined 33 eyes with FTMHs and 45 eyes of different patients after successful repair of FTMHs (mean: 13 months after surgery; range: 1 month to 71 months). Additionally, 16 unaffected fellow eyes of the above-mentioned patients were examined. Seven eyes from this group had a small macular hole (< 250 μm), 18 patients had a medium macular hole (250 μm to 400 μm), and the rest had large macular holes.5 The mean age of our patients was 70 years. There were no statistically significant differences in the measurements performed by the two examiners.
In SS-OCTA, the superficial vessels layer of stage two macular holes, with an undetached operculum still attached to the edge of the hole, appeared similar to those of a normal eye and the avascular zone appeared normal. In the deep retinal vascular layer, a slight disarrangement at the margin of the FAZ was visible. In the choriocapillaris layer, we observed a hyporeflective circle and a slightly disarranged reflectivity when compared to healthy subjects (Figure 1). In larger macular holes, a series of artifacts was noted.
Swept-source optical coherence tomography angiography (SS-OCTA) comparison between a stage two macular hole (top) and a healthy eye (bottom). Orange frame: Retina at the level of superficial vessels. Green frame: Retina at the level of deep retinal vessels. Light blue frame: Retina at the level of retinal pigment epithelium. Dark blue frame: Level of choriocapillaris. (A) Full-thickness stage two macular hole. At the level of deep retina vessels, slight disarrangement at the margin of the foveal avascular zone is visible. This corresponds to cystoid spaces visible in SS-OCT. At the level of choriocapillaris, a hyporeflective circle is visible, corresponding to the maximum diameter of the macular hole. (B) Healthy macula.
Inappropriate layer segmentation: In all cases of stage three and four of FTMHs, a hyperreflective circle was noted prior to surgery at the level of choriocapillaris (Figures 2 and 3). Such artifacts may be the result of inappropriate layer segmentation. The segmentation lines, delineating deep retina vessels, and the avascular zone are automatically inserted by the software erroneously, below the retinal pigment epithelium (Figure 2). Such errors occur especially in macular holes larger than 400 μm (ie, in 84% of cases).
Swept-source optical coherence tomography angiography (SS-OCTA) of a full-thickness macular hole. Measurements performed in the current study and artifacts. White lines indicate the measured area of foveal avascular zone in the superficial retinal vessels layer (orange frame) and the area of deteriorated deep retina vessels (green frame). Orange frame: Retina at the level of superficial vessels. Green frame: Retina at the level of deep retinal vessels. Light blue frame: Retina at the level of retinal pigment epithelium. Dark blue frame: Level of choriocapillaris. Lower left: SS-OCT with presented segmentation in representative colors. Lower middle: Image of all vessels in representative colors. Lower right: Fundus image. The arrow indicates a hyperreflective oval shape, which is visible in all retinal layers, corresponding to an artifact occurring due to increased light transmission through the macular hole.
Swept-source optical coherence tomography angiography of full-thickness macular hole. Manual artifact removal. The upper images present a macular hole in superficial (left) and deep (right) retinal vessels. The oval hyperreflective shape in the middle is an artifact. The lower pictures were transformed with temporal and de-flickering noise reduction algorithms and plugins in Adobe After Effects in order to remove the central artifact. We noted areas of decreased perfusion on the upper margin of the macular hole. These probably correspond to retinal edema around the macular hole.
“Jellyfish-like” appearance: In three eyes, we noted a “jellyfish-like” appearance (Figures 2 and 4). In these cases, at the deep retinal layer we noticed a hyperreflective circle around the fovea with adhering hyperreflective spikes. In some cases, inside the circle, another hyperreflective, round shape was visible. The latter corresponds to the light transmission artifact described above (Figure 2). The jellyfish-like appearance was observed in cases with especially large cystoid spaces visible in SS-OCT at the margins of the macular hole coexisting with detachment of the margins of the macular hole.
Swept-source optical coherence tomography at the level of the deep vascular plexus resembling a jellyfish in shape. Probable cystoid spaces are represented as “jellyfish digestive canals.” The hyperreflectiveness of those canals might suggest that some blood flow is present in the area of retinal edema around the macular hole. Vessels might be simply transposed because of the presence of the macular hole.
These artifacts were comparable in both spectral-domain OCTA and SS-OCTA. The only difference was that SS-OCTA gave a better visualization of the reflectivity irregularity in reflectivity in the choriocapillaries layer (Figure 5). It was then possible to remove them with the software included with the Triton (Figure 6).
Comparison of two different optical coherence tomography (OCT) techniques of the same eye on the same day. (A) Spectral-domain OCT angiography (OCTA) (RTVue). (B) Swept-source OCTA (Triton).
Swept-source optical coherence tomography angiography (SS-OCTA) of a full-thickness macular hole before (A, B) and after (C, D) “jellyfish” artifact removal by adjustment of segmentation with included software (Triton). (A) From left: Superficial retinal vessels plexus, deep retina vessels plexus with visible “jellyfish artifact,” avascular zone, choriocapillaris. (B) From left: SS-OCT with visible layer segmentation, color OCTA of multiple layers, color photo. (C) From left: Superficial retinal vessels plexus, deep retina vessels plexus without the “jellyfish artifact,” avascular zone, choriocapillaris. (D) From left: SS-OCT with visible layer segmentation after correction, color OCTA of multiple layers, color photo.
Full-Thickness Macular Holes Before Surgery
Baseline data are listed in Table 1. The area of the avascular zone in the deep retina layers in eyes with FTMHs correlated with the minimum and maximum diameter of the macular hole (P = .0003, r = 0.683; P = .001, r = 0.612, respectively; Pearson Product Moment). The correlation between the largest linear diameter of the avascular zone in deep retina layers and minimum and maximum diameter of the macular hole was also statistically significant (P = .0008, r = 0.714; P = .0009, r = 0.649, respectively).
Patients With FTMHs and SS-OCT/SS-OCTA Data Collected Preoperatively
Full-Thickness Macular Holes After Surgery
Baseline data are listed in Table 2. No significant correlation was observed between elapsed time after surgery and size of the FAZ.
Patients With FTMHs and SS-OCT/SS-OCTA Data Collected Postoperatively
The postoperative central retinal thickness correlated negatively with preoperative minimum macular hole diameter (P = .004, r = −0.45), and with maximum macular hole diameter (P = .01, r = −0.38).
Postoperative central retinal thickness correlated negatively with the area of the FAZ in deep retina vessels measured after surgery (P = .04, r = −0.324; Spearman rank product) (Figures 7 and 9).
Increased size of the foveal avascular zone (FAZ) at the level of the deep retina vessel plexus corresponds with central retinal thickness 1 month after surgery. The green circle represents the area of the FAZ. Upper from left: Swept-source optical coherence tomography angiography (SS-OCTA) at the level of the superficial retinal vessels, SS-OCTA at the level of deep retinal vessels, and SS-OCT. Lower from left: SS-OCTA) at the level of the superficial retinal vessels, SS-OCTA at the level of deep retinal vessels, and SS-OCT.
Localization of vessels in particular retina and choroidal layers. Orange circles represent superficial retina vessel plexus. Green circles represent deep retina vessels plexus. Red circles represent choriocapillaries.
Visual acuity improved significantly with time after surgery (P = .05, r = 0.36; Pearson correlation).
We examined an additional 16 unaffected fellow eyes of eyes in which vitrectomy for FTMHs had been successfully performed with the temporal inverted internal limiting membrane flap technique. In the deep retina layers the largest linear diameter of the FAZ in the fellow eyes was 770 μm and the area was 521 μm2. The liner diameter was statistically significantly larger than in the eye with FTMH, but not in the area of the avascular zone (P = 0.015 and P = 0.1 respectively, Mann-Whitney Test). In all cases, an increased irregularity of the reflectivity of choriocapillaris was observed when compared to healthy eyes (Figure 8).
Swept-source optical coherence tomography angiography comparison between the fellow eye of full-thickness macular hole (FTMH) (top) and a healthy eye (bottom). Orange frame: Retina at the level of superficial vessels. Green frame: Retina at the level of deep retinal vessels. Light blue frame: Retina at the level of retinal pigment epithelium. Dark blue frame: Level of choriocapillaris. Arrows: Area of deteriorated reflectivity at the choriocapillaris layer in the fellow eye of FTMH.
SS-OCTA is a noninvasive tool to examine retinal and choroidal microcirculation. Its usefulness was already described for age-related macular degeneration, macular telangiectasia, and retina vascular diseases.1,6,7,8 As with every imaging technique, it is not always free from artifacts.
In SS-OCTA, all moving objects (erythrocytes in blood vessels) are represented as hyperreflective. The round or oval hyperreflective zone visible in the middle of the fovea in multiple layers of the SS-OCTA image (Figure 2, white arrow; Figures 3 and 5) is, in fact, an artifact. The device software automatically finds and tracks levels of the various tissue types, and it may encounter difficulties in maintaining its level when it encounters a large macular hole. This often leads to the software falsely situating the scanning lines of the deep retinal layer and the avascular zone below the retinal pigment epithelium. This might be confirmed by the fact that in small macular holes, as presented in Figure 1, the artifact is missing. The segmentation algorithm fails in these cases in various SS-OCTA devices (Figure 5). Such artifact may be deleted with additional software. Glittenberg et al. described this phenomenon during the Association for Research in Vision and Ophthalmology congress in 2016. They suggested that the raw data might be exported and enhanced using temporal as well as de-flickering noise reduction algorithms and specialized plugins in Adobe After Effects (San Jose, CA) (Figure 3). The newest software version (Triton) enables automatic correction of segmentation (Figure 6).
The jellyfish-like appearance (Figures 4, 5, and 6) is visible in some patients with cystoid spaces at the margins of the macular hole. The striae are of similar reflectivity to retina tissue, and no flow is visible in the representation of “digestive canals.” Such an appearance represents a segmentation failure, which was also earlier described by our group.9 It must be considered that vessels might have several junctions not only in the X-axis but also in the Y-axis, thus SS-OCTA results must be evaluated very carefully. Rizzo et al. analyzing 13 eyes with SD-OCTA also suggested that “vascular sliding” at the borders of cystoid spaces is consistent with persistent microvasculature.10
Retinal and Choroidal Blood Flow in FTMHs
In this paper, we report that blood flow in the deep retinal layers and choriocapillaris might be altered in FTMHs and their fellow eyes, both before and after successful surgery (Figure 8). Teng et al. similarly observed that flow in the choriocapillaris does not completely normalize 1 month after surgery and is especially decreased in eyes with larger macular holes.11 They hypothesized that the restoration of blood flow may take a longer period of time. The exact count of vessel density was not the topic of this publication. However, in our group, some choriocapillaris defects were visible even many years after macular hole surgery. The exact count of these defects would require further studies.
We found a statistically significant correlation between the diameter and area of the FAZ in deep retina layers and postoperative central retinal thickness (Figure 7 and Figure 9). Such a correlation was also observed for the superficial retinal FAZ by other authors.12 It was previously reported that the size of the superficial FAZ inversely corresponds to central retinal thickness also in healthy eyes.13 The diameter of the FAZ is always slightly larger than the minimum diameter of the macular hole and slightly smaller than the maximum diameter of the macular hole. This might be explained by the fact that the size of the macular hole gradually increases from the inner retinal layers up to the retinal pigment epithelium, and the layer of deep retinal vessel lies somewhere in between. The lack of or decreased density of vessels around the macular hole in the deep retinal vessels may explain why we are unable to achieve complete functional recovery in some cases. This may be caused by some perfusion abnormalities, especially in long-standing macular holes.
The jellyfish-like shape is sometimes visible at the level of the deep retinal vessels (Figures 4, 5, and 6). It is visible in macular holes with large parafoveal cystoid spaces and suggests that even in such macular holes, some perfusion exists around the macular hole, even if the vessels are pushed aside by the cystoid spaces. They may reperfuse after surgery. This is based on the fact that in some cases of FTMH we observe cystoid spaces on the margin of the macular hole. In SS-OCTA at the level of deep retina vessels, we see a representation of these cystoid spaces in the strange jellyfish shape. It might also correspond to reduced blood flow around the macular hole.
We observed deteriorated appearance of the choriocapillaries layer in macular holes before and after surgery, and in fellow eyes. One possible explanation may be that perhaps a prolonged traction creating the macular hole may furthermore cause some defects in the choriocapillaris. This might explain incomplete visual recovery in some cases. On the other hand, taking into consideration that it occurs also in fellow eyes, and that fellow eyes develop FTMHs significantly more frequently than in the general population, it might be extrapolated that such an appearance of the choriocapillaris layer might be a predisposing factor to FTMH formation.
A bias of the current study is that it was retrospective in nature, and we did not observe the same patients before and after vitrectomy. Further observational studies on the same eyes before and after surgery may enhance our knowledge of FTMHs.
To conclude, in this study we report two possible artifacts appearing while visualizing FTMHs with SS-OCTA: the central hyperreflective circle caused by increased light transmission and segmentation failure, and the jellyfish-like appearance caused by segmentation failure. Even after manual removal of those artifacts, certain deteriorations in deep retinal vasculature and in the choriocapillaris are visible. The correlation between the size of the postoperative FAZ in the deep retina layers plexus and central retinal thickness may add some explanation to incomplete recovery of visual acuity despite successful surgery.
- Spaide RF, Fujimoto JG, Waheed NK. Image artifacts in optical coherence tomography angiography. Retina. 2015;35(11):2163–2180. doi:10.1097/IAE.0000000000000765 [CrossRef]
- Smiddy WE. Macular hole formation without vitreofoveal traction. Arch Ophthalmol. 2008;126(5):737–738. doi:10.1001/archopht.126.5.737 [CrossRef]
- Medina CA, Ortiz AG, Relhan N, Smiddy WE, Townsend JH, Flynn HW Jr, . Macular hole after pars plana vitrectomy for rhegmatogenous retinal detachment. Retina. 2017;37(6):1065–1072. doi:10.1097/IAE.0000000000001351 [CrossRef]
- Niwa H, Terasaki H, Ito Y, Miyake Y. Macular hole development in fellow eyes of patients with unilateral macular hole. Am J Ophthalmol. 2005;140(3):370–375. doi:10.1016/j.ajo.2005.03.070 [CrossRef]
- Duker JS, Kaiser PK, Binder S, et al. The International Vitreomacular Traction Study Group classification of vitreomacular adhesion, traction, and macular hole. Ophthalmology. 2013;120(12):2611–2619. doi:10.1016/j.ophtha.2013.07.042 [CrossRef]
- Rispoli M, Savastano MC, Lumbroso B. Capillary network anomalies in branch retinal vein occlusion on optical coherence tomography angiography. Retina. 2015;35(11):2332–2338. doi:10.1097/IAE.0000000000000845 [CrossRef]
- Muakkassa NW, Chin AT, de Carlo T, et al. Characterizing the effect of antivascular endothelial growth factor therapy on treatment- naïve choroidal neovascularization using optical coherence tomography angiography. Retina. 2015;35(11):2252–2259. doi:10.1097/IAE.0000000000000836 [CrossRef]
- De Carlo TE, Chin AT, Bonini Filho MA, et al. Detection of microvascular changes in eyes of patients with diabetes buy not clinical diabetic retinopathy using optical coherence tomography angiography. Retina. 2015;35(11):2364–2370. doi:10.1097/IAE.0000000000000882 [CrossRef]
- Michalewska Z, Nawrocki J. Atlas of Swept Source Optical Coherence Tomography. Switzerland: Springer; 2017;11.3:108–112.
- Rizzo S, Savastano A, Bacherini D, Savastano MC. Vascular features of full-thickness macular hole by OCT angiography. Ophthalmic Surg Lasers Imaging Retina. 2017;48(1):62–68. doi:10.3928/23258160-20161219-09 [CrossRef]
- Teng Y, Yu M, Wang Y, Liu X, You Q, Liu W. OCT angiography quantifying choriocapillary circulation in idiopathic macular hole before and after surgery. Graefes Arch Clin Exp Ophthalmol. 2017;255(5):893–902. doi:10.1007/s00417-017-3586-0 [CrossRef]
- Baba T, Kakisu M, Nizawa T, Oshitari T, Yamamoto S. Superficial foveal avascular zone determined by optical coherence tomography angiography before and after macular hole surgery. Retina. 2017;37(3):444–450. doi:10.1097/IAE.0000000000001205 [CrossRef]
- Samara WA, Say EA, Khoo CT, et al. Correlation of foveal avascular zone size with foveal morphology in normal eyes using optical coherence tomography angiography. Retina. 2015;35(11):2188–2195. doi:10.1097/IAE.0000000000000847 [CrossRef]
Patients With FTMHs and SS-OCT/SS-OCTA Data Collected Preoperatively
|Mean minimum diameter of macular hole||497.8 μm|
|Mean maximum diameter of macular hole||922 μm|
|Mean linear diameter of the FAZ in deep retinal layers||504.4 μm|
|Mean area of the FAZ in deep retinal layers||541.2 μm2|
|Mean linear diameter of the lower/disturbed perfusion zone in deep retinal layers||1,277 μm|
|Mean area of the lower/disturbed perfusion zone in deep retinal layers||1,189 μm2|
Patients With FTMHs and SS-OCT/SS-OCTA Data Collected Postoperatively
|Mean minimum diameter of macular hole||419.7 μm|
|Mean maximum diameter of macular hole||878.2 μm|
|Mean linear diameter of the fovea avascular zone in deep retinal layers||638.3 μm|
|Mean area of the fovea avascular zone in deep retinal layers||404.9 μm2|
|Central retinal thickness||167 μm|