Multilayer Retinal Correspondence of the Structural and Vascular Anomalies in Eyes With Early Macular Telangiectasia Type 2

Abstract

Purpose: To assess the correspondence between interdigitation zone (IZ) reflectivity, ellipsoid zone (EZ) loss, inner retinal layer reflectivity, patterns of capillary dilation, and telangiectasia in eyes with early macular telangiectasia type 2 (MacTel).

Patients and Methods: Twenty-eight eyes of 22 patients with grade 0–2 MacTel (according to the MacTel project classification) and 28 healthy control eyes were included in this study. Multimodal imaging, including optical coherence tomography (OCT) angiography, adaptive optics flood illumination ophthalmoscopy (AO-FIO) and blue light reflectance (BLR), was performed. The EZ, IZ, and outer plexiform layer (OPL) were analyzed on the structural OCT C-scans. The vascular density (VD) was measured on the binarized and skeletonized angiograms of the superficial vascular plexus and deep capillary complex. The vascular diameter index (VDI) was calculated by dividing the binarized VD by the skeletonized VD.

Results: On AO-FIO, cone density in the MacTel zone was significantly lower in MacTel eyes than in controls, even in areas located outside the EZ loss (P < 0.001). A distinctive pattern of IZ reflectivity attenuation extended beyond the area of EZ attenuation. The shape and size of a strong OPL hyper-reflectivity corresponded to the MacTel white area (MacTel zone) seen on BLR. Capillary dilation and rarefaction were colocalized with this area, extending beyond visible telangiectasia. The VDI was higher in MacTel eyes than in controls (P < 0.001).

Conclusions: These findings suggest that in early MacTel eyes, photoreceptor signal alteration, OPL hyper-reflectivity, and capillary dilation, potentially associated with Müller cell dysfunction, precede the EZ loss.

Idiopathic macular telangiectasia type 2 (MacTel) is a bilateral neurodegenerative disease affecting 0.06% to 0.1% of the population.¹ The current pathogenetic concept describes an altered serine and lipid metabolism resulting in the loss of Müller cells (MC).^2,3 

Ellipsoid zone (EZ) loss is a well-known marker of MacTel progression.^4,5 Clinical trials assessing treatment with ciliary neurotrophic factor implants in MacTel have used this outcome to assess treatment efficacy.^6,7 Furthermore, it has been shown that the EZ loss leads to capillary proliferation at the edge of the outer retina.^8–11 

Attenuation of the interdigitation zone (IZ) has also been proposed as a potential novel marker¹² because the IZ may be attenuated and fragmented beyond the visible EZ alteration.^13–16 However, the association between anomalies of the photoreceptors, inner retinal layers, and retinal capillary layers has not been systematically investigated at an early stage. 

The aim of this study was to analyze the correspondence between the IZ reflectivity, the EZ loss, the inner retinal layer reflectivity, the capillary dilation pattern and telangiectasia in early MacTel eyes, before the appearance of any outer retinal hyper-reflectivity (ORHR) and outer retinal neovascularization. Studying this association could help to define early markers of disease progression and treatment response. 

Patients and Methods

Patients with MacTel participating in the Natural History and Observation Registry Study¹⁷ were included in this monocenter study conducted at the Ophthalmology Department of Lariboisière Hospital, Paris, France. The study adhered to the tenets of the Declaration of Helsinki, and all participants gave their informed consent to participate. 

Patients with grade 0, 1 and 2 MacTel were selected based on the recent MacTel project classification.¹⁸ Patients could have small central or noncentral EZ breaks but could not show any ORHR, pigment or neovascularization. ORHR was defined as hyper-reflective lesions appearing as bright as the retinal pigment epithelium layer on optical coherence tomography (OCT), which may be linear or appear as mounds emerging from the retinal pigment epithelium and extending beyond the external limiting membrane of the retina.¹⁸ 

Cases with hemorrhage, pigment, or epiretinal membrane in the inner retina, which could have attenuated the EZ and IZ, were excluded. Normal healthy eyes of nondiabetic subjects were also included as controls. 

Patients and controls underwent a full clinical examination, including best-corrected visual acuity testing. The visual acuity (VA) was measured using a Snellen chart and converted into logarithm of the minimum angle of resolution. 

Multimodal Imaging

Different devices were used depending on the objectives or comparisons. Blue light reflectance (BLR) imaging, fluorescein angiography, and spectral-domain OCT were acquired with the HRA/Spectralis device (Heidelberg Engineering, Heidelberg, Germany). 

OCT angiography was acquired with the Plex Elite 9000 Swept-Source OCT-A device (Carl Zeiss Meditec, Jena, Germany). PlexElite is a swept-source device with an A-scan depth of 3.0 mm in tissues, an axial optical resolution of about 6.3 µm, and a transverse resolution, calculated at the beam size of the pupil, estimated at about 20 µm. 

The vascular plexuses were analyzed on 3 × 3-mm macular angiograms. The predefined boundaries provided by OCT-A software were used to characterize the retinal layers. The superficial vascular plexus (SVP) boundaries were set at the inner limiting membrane and inner plexiform layer/inner nuclear layer (IPL/INL) interface and the deep capillary complex (DCC) boundaries at the IPL/INL and outer plexiform layer/outer nuclear layer (OPL/ONL) interfaces. 

Adaptive optics flood illumination ophthalmoscopy (AO-FIO) was acquired with the rTX1 device (rtx1; Imagine Eyes, Orsay, France). AO-FIO uses wavefront sensors to detect aberrations in the ocular optics and deformable mirrors to compensate for the aberrations to improve retinal imaging performance. AO-FIO is an en-face reflectance imaging system that provides a lateral resolution of up to 2 µm in the human retina.¹⁹ For each subject, a series of 25 4° × 4° images was acquired with a 50% overlap between adjacent images to cover a 12° × 12° field in the central macula. At each fixation point, 40 raw AO-FIO images were automatically recorded and averaged to improve the signal-to-noise ratio using the software supplied by the manufacturer (ck_v0_1b; Imagine Eyes). The montage of the images was performed with i2k Retina using affine transformation parameters (DualAlign, LLC, Clifton Park, NY, USA). 

Imaging Processing

All images were processed using ImageJ software (National Institutes of Health, Bethesda, MD, USA). 

Cone Identification and Cone Density (CD) Measurements on AO-FIO

All montages were cropped to fit the 3 × 3 mm angiogram of the C-scan. In the final images of AO-FIO, the cones correspond to definite bright round structures. In the absence of cones or when the cone outer segments (OS) are altered, the image is blurry and dark. Rods were not visible on AO-FIO because their size was below the resolution limit. 

The edges of the pixels were sharpened using the unsharp mask (radius = 0.99, mask = 1), and the morphological filter plugin was then used to extract round particles of 4-pixel radius corresponding to the cones. According to Curcio et al.,²⁰ the cone radii are 3.3 µm at 100-200 µm from the foveal center (4-5 pixels on AO-FIO images). The cone diameter increases from the fovea to the periphery. The visualization of these particles was enhanced by applying a threshold. The binarized images obtained were used to analyze the en face pattern of the cone mosaic. The find maxima function was used to count the particles.²¹ This function automatically detects the central coordinates of small circular spots with a stronger brightness than that of the surrounding background. A single point and a segmented particle diagram were obtained. The segmented polygons were considered to correspond to the cell surface. The CD map was generated by calculating the density for each cell (one per cell surface) and the regions of interest color coder plugin (Fig. 1). 

Figure 1.

 

AO-FIO montage, enhanced cone visualization and CD mapping of a healthy right eye. (A) AO-FIO montage, consisting of 25 frames of 4° × 4° including the MacTel zone. The montage was cropped to fit the 3 × 3 mm angiogram acquired by OCT. (B) The visualization of round particles, corresponding to the cones, was enhanced using a morphological filter and a threshold applied to the montage. (C) CD mapping in this healthy control eye. Because of the limitations of AO-FIO in resolving cones from the center of the fovea, this area was excluded from the analysis (white circle on B and C).

Figure 2.

 

Binarization and skeletonization of the angiogram. (A) C-scan angiogram of the deep capillary complex of the right eye of a 68-year-old patient with grade 0 MacTel. (B) Result of the binarization process, which preserved the capillary dilation. (C) Result of the skeletonization process applied to the binarized image. The central ring of the ETDRS grid was superimposed on the image. Measurements were performed in the temporal, superior, nasal, and inferior perifoveolar quadrants.

The CD was measured in the four quadrants of the central ring of the ETRDS grid using the single point diagrams. The central 1000-µm ring was excluded from the analysis because AO-FIO images obtained at retinal eccentricities within 2.5° from the foveal center have been shown not to be reliable for obtaining cone metrics on AO-FIO.²² 

EZ and IZ Assessment

The EZ and IZ were assessed on the structural SD-OCT B-²³ and C-scans. For the segmentation of the EZ, the boundaries of the slab were set 40 and 75 µm above the Bruch's membrane. For the segmentation of the IZ, the boundaries were set 20 and 40 µm above the Bruch's membrane. 

The reflectivity pattern of the EZ and IZ was analyzed on the C-scans. The correlation between a reduction in reflectivity on the en face images and the attenuation or disappearance of the outer retinal lines on the B-scans was assessed. 

The reflectivity of a 2.5 × 3-mm oval zone centered on the foveal center was also assessed by measuring the mean gray value on the C-scans of MacTel and control eyes. This zone corresponds to the MacTel zone.²⁴ 

Areas of EZ loss were drawn on the C-scans and validated by the loss of the EZ zone on the B-scans as previously described with the OCT Spectralis.⁴ EZ loss referred to the complete disappearance of the EZ line on the B-scan. An attenuation of the EZ signal alone was not considered a break. 

Inner Retinal Layer Reflectivity

The reflectivity of the ganglion cell layer (segmented between the internal limiting membrane and the IPL) and the INL (between the IPL and the OPL) were assessed on the structural C-scans. An additional segmentation at the OPL alone was also performed, using two boundaries parallel to the Bruch's membrane spaced by 40 µm. The reflectivity of the 2.5 × 3 mm oval zone centered on the foveal center at these different layers was assessed by measuring the mean gray value on the C-scans of MacTel and control eyes. 

Vascular Diameter Index (VDI) Measurement (Fig. 2)

The vascular density (VD) was measured on binarized images. A binarization method preserving the vascular diameter that effectively removes isolated signals was used. The background was removed using a top hat filter. Then a gray-scale attribute filtering of the morphological plugging filter was used to isolate particles with an area of 100 pixels and a connectivity of 4. Binarized images were generated by applying a mean auto-threshold. The VD was measured by calculating the percentage of white pixels on these binarized images. 

Binarized images were processed using a Hessian filter (smallest smoothing = 4) and skeletonized. The percentage of white pixels on these skeletonized images, corresponding to the vessel length, was also calculated. 

The VDI was measured by dividing the binarized VD by the vessel length.²⁵ Measurements were performed in the four quadrants of the parafoveal ring of the ETDRS grid. 

Statistical Analyses

Statistical analyses were performed using Excel stat software. Shapiro-Wilk tests were used to determine if data were distributed normally. A Mann-Whitney test was used to compare MacTel eyes versus controls. All P values < 0.05 were considered statistically significant. A χ² test was used to determine whether there were gender-related differences within each group. 

Results

Twenty-eight eyes of 22 patients with MacTel, with a mean age of 60 years (43–70 years) were included. There were 11 men and 11 women. The mean VA was 0.2 ± 0.1 logMAR (20/50-20/20 Snellen equivalent). Seven patients (32%) had type 2 diabetes without diabetic retinopathy. According to the 2023 MacTel classification, 10, four, and 14 eyes had grade 0 (i.e. no EZ break), grade 1 (i.e. non-central EZ break, no ORHR), and grade 2 (i.e. central EZ break) MacTel, respectively. 

Twenty-eight healthy eyes of 21 nondiabetic subjects, with a mean age of 54 years (33–78 years) were also included as controls. There were 9 men and 12 women. Patients and control subjects were matched for age and sex (P = 0.36). 

EZ and IZ Break and Attenuation on OCT and Cone Mosaic Disorganization on AO-FIO

On the OCT C-scans, a specific topography of the EZ and IZ reflectivity attenuation was observed in MacTel eyes. The IZ, which was uniformly hyper-reflective in controls, showed an inhomogeneous grey surrounded by a dark border in MacTel eyes. This relative hyporeflectivity was also visible, but less marked, on the en face image of the EZ (Fig. 3). The mean gray values of the IZ and EZ measured in an oval area of 2.5 mm × 3 mm were significantly lower in MacTel eyes than in controls (P < 0.0001 for both) (Supplementary Fig. S1). 

Figure 3.

 

Pattern of EZ and IZ alterations on the OCT C-scan and AO-FIO images of the eye of a 43-year-old woman with grade 1 MacTel (upper row) and a control eye (lower row). (A, E) C-scan of the EZ. The areas of EZ loss are marked in red in the MacTel eye. (B, F) C-scan of the IZ. The central macula, an area with a high cone density, showed a relatively uniform reflectivity in the control eye (F), whereas it appeared as an inhomogeneous gray in the MacTel eye. This area was surrounded by a dark border (B). (C, G) AO-FIO montages, processed to enhance cone visualization. A similar pattern of cone loss, resembling the reflectivity attenuation of the IZ, and to a lesser extent, the EZ was observed in the MacTel eye. (D, H) The attenuation of the EZ and IZ layers is highlighted by yellow brackets corresponding to the yellow dashed lines in B. The IZ showed a greater attenuation than the EZ. The red bracket shows a break in the EZ that is also seen in B. (H) OCT B-scan of a normal control eye. Because of the limitations of AO-FIO in resolving cones from the center of the fovea, this area was excluded from the analysis (white circle on C and G).

The dark border was exactly concentric, although slightly smaller than the MacTel zone seen on BLR. The attenuation and loss of the IZ and EZ on the C-scans were confirmed by examining their reflectivity on the B-scans (Fig. 3). 

On the enhanced cone AO-FIO montages, a pattern similar to that seen on the IZ C-scans was observed. The CD measured on AO-FIO was also significantly lower in MacTel eyes than in controls in the four perifoveal quadrants (all P < 0.001) (Supplementary Table S1). The lowest CDs were observed in the temporal quadrants, but the entire MacTel zone showed a decreased CD that did not correlate with the decrease in VA (Fig. 4). It thus appeared that the reflectivity attenuation pattern observed on the IZ and EZ OCT C-scans was related to the alteration of the cone OS on AO-FIO. 

Figure 4.

 

Decrease in CD in MacTel eyes compared to control eyes. (A) AO-FIO montage processed to enhance cone visualization in a left eye of a 60-year-old woman with grade 2 MacTel. Visual acuity was 20/32. The darker areas corresponded to an alteration in the cone mosaic. An area of EZ break is shown in red. (B) On the optical coherence tomography B-scan, the yellow brackets highlight areas of reflectivity attenuation in the IZ corresponding to the yellow line in A. The area of EZ loss is shown in red. (C) AO-FIO CD mapping, showing the decrease in density in areas of IZ and EZ reflectivity attenuation. (D) Box plot illustrating CD measurements in MacTel eyes compared to control eyes in the four quadrants of the central ETDRS grid. The CD was significantly lower in MacTel eyes.

Inner Retinal Layer Hyper-Reflectivity on the OCT C-Scans: Comparison With Healthy Eyes and the MacTel Zone Seen on BLR

The OCT C-scans of the inner retinal layers showed a hyper-reflective image, the shape and size of which corresponded exactly to the MacTel zone (i.e. the white area visible on BLR). The strongest reflectivity was observed on the C-scans segmented at the OPL. This hyper-reflective pattern was not observed in control eyes (Fig. 5; Supplementary Fig. S2). The mean gray value in the 2.5 × 3-mm oval area centered on the fovea was significantly higher in MacTel eyes than in controls (P < 0.0001) (Supplementary Fig. S2). 

Figure 5.

 

Reflectivity of the inner retinal layers in the macula. (A) MacTel zone seen on the BLR image of the right eye of 70-year-old man with grade 1 MacTel. (B) OCT C-scan, segmented at the OPL, showing a hyper-reflective oval area, the size and shape of which corresponded to the MacTel zone. (C) Surface plot showing the grayscale values of this hyper-reflective area. This hyper-reflective pattern was not observed in control eyes in which the OPL was uniformly reflective throughout the image surface (D–F). (G) Measurements of the mean grayscale values within the MacTel zone (3 mm × 2.5 mm, indicated by the yellow oval area in E) were significantly higher in MacTel eyes than in control eyes.

VDI

The VDs in the SVP and DCC measured on the binarized images were similar between MacTel and control eyes in the temporal and nasal quadrants. However, the VDs in the SVP and DCC measured on the skeletonized images were significantly lower in MacTel eyes than in control eyes (all P < 0.001). The VDI was also higher in MacTel eyes in all four quadrants (all P < 0.01) than in control eyes (Fig. 6) (Supplementary Table S2). This difference indicated a capillary rarefaction masked by a capillary dilation of the remaining capillaries. 

Figure 6.

 

Capillary rarefaction and dilation in the MacTel zone in the right eye of a 51-year-old woman with grade 2 MacTel. (A, D) Structural OCT C-scan segmented at the outer plexiform layer, showing a hyper-reflective area corresponding to the MacTel zone (A) and a homogeneous reflectivity in a control eye (D). (B, F) Vascular diameter index map, highlighting telangiectasia, and diffuse capillary dilation in the MacTel zone (B) compared to a control eye (F). (C, F) Vascular density map, measured on the skeletonized image, showing capillary rarefaction in the MacTel zone (C) compared to a control eye (F).

Multilayer Correspondence

The highest VDIs were found in the temporal quadrants where telangiectasia were subjectively identified. The VDI was also greater than in control eyes in the other quadrants, even in the absence of telangiectasia (Supplementary Table S2). Moreover, the area of increased VDI was co-localized with the area of hyper-reflectivity observed at the OPL. 

The decrease in CD was also more pronounced in the temporal quadrant, in the area of telangiectasia, although the decrease involved the entire MacTel zone. The CD decrease correlated with the VD decrease on the skeletonized images, illustrating the capillary rarefaction associated with an increase in VDI (Fig. 7). 

Figure 7.

 

Multilayer retinal changes in the right eye of a 59-year-old woman with grade 2 MacTel. (A) MacTel zone seen on BLR. (B) Hyper-reflectivity of the outer plexiform layer, corresponding to the MacTel zone on the OCT C-scan. (C) Dye leakage from telangiectasia on fluorescein angiography. The leakage area was smaller than the MacTel zone. (D) OCT-angiography of the deep capillary plexus showing focal telangiectasia and diffuse capillary dilation in the MacTel zone. (E) Structural OCT C-scan showing three foci of the ellipsoid zone (EZ) breaks (marked in red). (F) The cone density map on adaptive optics showed widespread cone disturbance, with the most severe alterations corresponding to the telangiectasia area. Because of the limitations of the rTX1 device in resolving cones from the center of the fovea, this area was excluded from the analysis (white circle). (G) OCT B-scan showing an EZ break (red brackets), and the attenuation of the EZ and interdigitation zone (IZ) reflectivity (yellow brackets). (H) On the structural OCT C-scan, the pattern of IZ reflectivity attenuation was more severe than that of the EZ (E) and closely resembled the alteration in the cone mosaic (F).

In summary, in eyes with early MacTel, the area of cone loss was co-localized with capillary rarefaction and dilation in the MacTel zone. This remodeling of the capillaries was also associated with OPL hyper-reflectivity. The most pronounced dilation, visible as telangiectasia, was co-localized with the area of most severe cone loss. 

Discussion

In this study, we observed that structural retinal anomalies in MacTel eyes were present in several retinal layers across the entire MacTel zone. As already suggested, a loss or dysfunction of MC could be the cause of these anomalies.^2,26,27 

In early MacTel eyes (grade 0–2), we observed an alteration of the perifoveal cone mosaic on AO-FIO and an attenuation of the IZ reflectivity on OCT, extending well beyond the disruptions of the EZ, and affecting the entire MacTel zone as seen on BLR. Interestingly, these findings were identified as early as grade 0. 

In previous studies, we and others have shown that the CD was lower than normal in MacTel eyes, even in areas located outside the EZ loss.^13,28–30 We have observed that this decrease in CD corresponded to an impairment of the IZ and a narrowing of the space between the EZ and the retinal pigment epithelium on OCT, with minimal changes in EZ reflectivity.^13,28 

Normally, the EZ aligns with the ellipsoid portion of the inner segments, whereas the IZ band corresponds to an ensheathment of the cone OSs through the apical processes of the retinal pigment epithelium, forming a structure known as the contact cylinder.³¹ The reduction in cone mosaic density on AO-FIO appears to be related to a reduction in photoreceptor OS alteration or misalignment, despite the presence of cone photoreceptor cells and their inner segments. These changes could occur before any visible interruption of the EZ on a B-scan. 

Histologically, Powner et al.² report a strong correlation between the EZ loss and rod depletion. Furthermore, the histological connectomic reconstruction of a MacTel macula has shown that many cones in the MacTel zone showed highly disordered disc membranes with extensive vesicularization and fragmentation.³² 

This region of cone mosaic alteration showed a distinct demarcation marked by a dark border visible on the OCT C-scans and AO-FIO. Ooto et al.³⁰ also observed peculiar dark, ringlike areas surrounded by small patches in the cone mosaic using AO scanning laser ophthalmoscopy. This ring was smaller than the white lesion seen on BLR. This is also what we observed in our study. The anatomical rationale for this demarcation line is not known, but a distinct border delineating the extent of altered MC in the MacTel zone has also been observed histologically using a connectomic approach.³² 

Previous studies have shown that residual cone structure can persist within small EZ lesions, which can show changes in appearance over time, as observed by AO scanning laser ophthalmoscopy. However, because our patients were examined only once, we did not have the opportunity to observe this phenomenon.^33,34 

Hyper-Reflectivity of the OPL

We identified a hyper-reflective region, the size and shape of which corresponded to the MacTel zone at the inner retinal layers. This hyper-reflective region was more visible when we adjusted the slab at the OPL. In control eyes, the OPL hyper-reflectivity was uniform over the entire 3 × 3-mm surface of the C-scan. The normal hyper-reflectivity of the OPL can be attributed to the multilaminar distribution of mitochondria in the OPL.³⁵ Mitochondrial alterations have been observed in all retinal layers within the MacTel zone.³² It has also been noted that, on electronic microscopy, cone terminals appear dense, compressed, and misshapen with disrupted and poorly defined postsynaptic invaginations and triadic complexes. The white area observed on BLR could therefore be a consequence of increased reflectivity within the OPL, which could be related to mitochondrial derangement at this level. 

This effect could be exacerbated by the loss of macular pigment.²⁴ Although BLR and OCT do not use the same wavelength and optical properties, a change in mitochondrial reflectivity could have resulted in hyper-reflectivity on both modalities. 

Capillary rarefaction and dilation in MacTel have been imaged using OCT-A since 2015. Our previous work has demonstrated both rarefaction and dilation, as well as irregularities in capillary arrangement in the SCP and DCC.³⁶ Subsequent studies have found a significant decrease in foveal vessel density;³⁷ a progressive rarefaction of retinal microvasculature, including a decrease in mean VD, skeleton density, and fractal dimension in the DCP;³⁸ and a decrease in flow density and capillary length in the SVP and DCC.⁹ 

The particularity of our series was to demonstrate, in the early stages of the disease, before outer retinal vascular proliferation sets in, the presence of both a reduction in the length of perfused capillaries and an enlargement of the remaining capillaries. Furthermore, the mapping showed that capillary dilation extended beyond visible telangiectasia, corresponding well to areas of hyper-reflectivity in the OPL and MacTel zone. The greatest dilation, visible as telangiectasia, corresponded closely to the areas of most severe cone loss (Fig. 7). 

The association between capillary rarefaction/dilation and the loss of MC is supported by several publications and could result from an imbalance between pro- and anti-angiogenic factors typically expressed by MC. Selective MC ablation has been shown to induce photoreceptor degeneration and vascular telangiectasia.²⁷ Early Müller glia dysregulation could trigger the retinal vascular pathology observed in MacTel.³⁹ 

In a relatively early stage of MacTel, a severe degeneration of MC has been observed within the MacTel zone.³² MC depletion in MacTel has been shown to induce photoreceptor degeneration, capillary rarefaction and dilation, and synaptic disorganization at the OPL. We were able to identify the relationship between these consequences, the most relevant finding being the extent of photoreceptor OS disorganization throughout the MacTel zone before the EZ loss. 

Our study has some limitations, mainly because of the small sample size, which is due to the selection of early MacTel cases with good fixation and clear media. Nevertheless, this small cohort allowed obtaining highly significant findings, meaning that the changes observed were substantial and real. 

Furthermore, the methods used for OCT-A image binarization thresholding and histography could have had an impact on the quantitative measurements, as previously described.^40,41 To mitigate this potential bias, we systematically applied the same quantification method to both MacTel and control eyes. The accuracy of the results depends on the quality of the segmentation, regardless of the level of precision applied. 

Because there are no established normative data and standardized measurement methods for CD, we compared MacTel eyes to an age-matched control group, using the same methodology. We could not perform CD measurement in the foveal area because of the limitations of the RTX1 device. 

Conclusions

In the early stages of MacTel, OCT-A allowed detecting tissue and vascular abnormalities at different levels that corresponded well with each other and with other imaging modalities (AO-FIO, BLR). The attenuation of the IZ extended well beyond the EZ break and corresponded to the decrease in CD seen on AO-FIO. The increase in OPL reflectivity corresponded to the white zone of the MacTel zone. Capillary dilation extended beyond the telangiectasia clusters throughout the MacTel zone. Capillary dilation and IZ attenuation in the MacTel zone preceded the EZ loss. 

Acknowledgments

Presented at the MacTel Meeting, October 2023 at Lowy Medical Research Institute, New York, USA. 

Disclosure: V. Krivosic, None; Z. Dobbels, None; C. Duliere, None; A. Zureik, None; R. Tadayoni, None; A. Gaudric, None