Retinal cavitations in macular telangiectasia type 2 (MacTel): longitudinal structure-function correlations

Abstract

Background/Aims:

To quantify retinal cavitation size over time in macular telangiectasia type 2 (MacTel) and to correlate changes with visual acuity and area of ellipsoid zone loss.

Methods:

OCT macula volume scans from sham eyes included in a prospective, phase 2 clinical trial of human ciliaryneutrophic factor for MacTelat baseline, 1, and 2 years of follow-up were analyzed. Cavitationswere segmented by two independent readers. Total cavitation volume was compared to area of ellipsoid zone loss and best-corrected visual acuity (BCVA).

Results:

Fifty-one eyes from 51 unique patients (mean age 62years, range 45 – 79 years) were included. Intra-class correlation between readers for cavitation volume was excellent (>0.99). Average cavitation volume was 0.0109mm³, 0.0113mm³, and 0.0124mm³at baseline, 1, and 2 years, respectively. The average rate of cavitation volume change was +0.0039mm³/year. 10 eyes (20%) had a significant change in cavitation volume during the study (3 decreased, 7 increased). Eyes with increased cavitation volume had worse BCVA compared to eyes with no change/decreased cavitation volume(71.5 vs. 76.1 ETDRS letters respectively). Cavitation volume was negatively correlated to BCVA (r=−0.37) but not to area of ellipsoid zone loss. Cavitation volume was negatively predictive of BCVA in both univariate and multivariate mixed-effects modeling withellipsoid zone loss.

Conclusions:

Retinal cavitations and their rate of change in MacTelcan be reliably quantified using OCT. Cavitations are negatively correlated with visual acuity and may be a useful OCT-based biomarker for disease progression and visual function in MacTel.

Precis:

Cavitations are signs of neurodegeneration in macular telangiectasia type 2. In this study, we show that increased cavitation is predictive of worsening visual acuity independent of ellipsoid zone loss.

Introduction:

Macular telangiectasia type 2 (MacTel) is an idiopathic,bilateral, degenerative condition characterized by perifovealtelangiectatic vessels and neurosensory retinal atrophy.[1]Clinical findings on biomicroscopy include graying of the perifoveolar retina, dilated right-angled retinal venules, retinal pigment plaques, superficial crystalline retinal deposits, and progressive focal retinal atrophy sometimes accompanied by neovascular complexes in late stages of the disease.[1,2]Once thought to be a primary retinal vascular disorder, there is increasing evidence to suggest that the vascular abnormalities are secondary to a degenerative process in the neurosensory retina, perhaps due to the loss of the Muller cells.[3–5]

Spectral-domain optical coherence tomography (SD-OCT) captures many of the characteristic features of MacTel. Retinal atrophy can be seen as the loss of the ellipsoid zone (EZ, also known as photoreceptor inner and outer segment junction). Additionally, SD-OCT provides detailed visualization of peculiar hyporeflectiveintraretinal spaces known as cavitations,[6]which are thought to form as the result of tissue loss in the neurosensory retina underlying areas of intact internal limiting membrane.[7,8]Compared to cysts seen in exudative conditions, cavitations in MacTel are thought to be secondary to degeneration since they possess less circularity and lower internal reflectivity on SD-OCT.[6,9]

Retinal cavitations in MacTel are known to fluctuate with time, perhaps due to progressive loss of retinal structural integrity with degenerating Muller cells and reactive gliosis.[10]Currently, the clinical significance of these cavitations and their relevance to visual function are unclear. Given the critical role of Muller cells in the pathophysiology of MacTel, fluctuations in cavitation size could serve as an imaging-based biomarker to predict disease activity. This study aims to characterize the natural history of fluctuations in retinal cavitations in MacTel and to correlate them to measures of visualfunction and disease progression.

Materials and Methods:

Subjects and Image Identification:

This retrospective post-hoc analysis was approved by the Institutional Review Board of the Duke Eye Center and adhered to the tenets of the Declaration of Helsinki. Anonymized data from sham eyes included in a previously-published phase 2, multicenter clinical trial (Clinicaltrials.gov identifier: NCT01949324)of human ciliaryneurotrophic factor (CNTF) for MacTel were identified for review.[11]In brief, participants were eligible if they had at least one eye with a diagnosis of MacTel, were age 21 years or older, and had disruption in the photoreceptor ellipsoid zone on spectral domain optical coherence tomography (SD-OCT) between 0.16mm² and 4.00mm².[11]The best-corrected visual acuity (BCVA) had to be 20/50 or better (Early Treatment Diabetic Retinopathy Study visual acuity test score of 64 or better). BCVAs were recorded at baseline, 1 year, and 2 years of follow up. Eyes with subretinalneovascular proliferation were excluded. The overall disease severity of included participants corresponded to stages 1–4, as described by Gass and Blodi (stage 5 was excluded).[1]Sixty-seven participants were enrolled in the study.[11]In 35 participants, a single study eye (16 implant, 19 sham) was randomized. In 32 participants with two eligible study eyes, the right eye was randomized and the left eye received the alternative treatment. Eyes receiving the implant were excluded from the study. Comprehensive eye examinationwas performed at baseline through 2 years. Data from baseline, 1, and 2 years of follow-up were included for analysis.

Image Acquisition and Analysis:

SD-OCT macula volume scans centered at the fovea covering an area of 20° x 20° with 97 B-scans were obtained with Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany). Scans from the baseline, 1 year, and 2 years of follow-up were analyzed.Cavitations, defined as irregularly-shaped,hyporeflective intra-retinal spaces, were manually segmented with the Duke OCT Retinal Analysis Program (DOCTRAP, version 65.2) by two independentreaders (example shown in Figure 1).[12]The cavitation volume was defined as the areaof the cavitation in each B-scan multiplied by the thickness of the B-scan (distance across all B-scans divided by the number of B-scans). The final cavitation volume was the average between the two readers.The area of ellipsoid zone loss was obtained from orthogonal topographic maps (“en face” images) of the ellipsoid zone as previously described.[13]

Figure 1:

Example of spectral-domain optical coherence tomography B-scan showing segmentation of cavitations in macular telangiectasia (A). Output from the Duke OCT Retinal Analysis Program (DOCTRAP) of the cavitation segmentation (outlined in pink) from reader 1 (B) and reader 2 (C) showing minimal differences in overall cavitation measurement.

Statistical Analysis:

Summary statistics were used to describe the average cavitation volume in the cohort. The change in cavitation volume and ellipsoid zone loss was calculated as the average change per year over the two years of follow-up. Intraclass correlation (ICC) based on a two-way mixed effect model was used to assess reliability of cavitation volume measurements. Anatomic variables of ellipsoid zone loss and cavitation volume were correlated to BCVA at each corresponding visit using Spearman’s rank-order correlation with a p-value <0.05 considered statistically significant. Multi-level mixed effects modelswith a random intercept for patient to account for repeated measures over time were generated totest the significance of variables. Wilcoxon signed-rank test and Kruskal-Wallis Htest, with Dunn’s test for multiple pairwise comparisons, were used to compare groups. All statistical analyses were performed using STATA (Stata Statistical Software, Version 14.1; StataCorp LP, College Station, TX).

Results:

Fifty-one eyes from 51 unique patients were included in the study.The mean age at baseline was 62 years (range 45 to 76 years). The average BCVAs at baseline, 1, and 2 years of follow up were 76.1 ± 6.7 letters, 75.9 ± 8.2 letters, and 75.5 ± 8.1 letters,respectively with a corresponding average ellipsoid zone loss of 0.77 ± 0.55 mm², 0.85 ± 0.53 mm², and 0.99 ± 0.76 mm².

Change in Cavitation Volume:

All eyes had cavitations seen on OCT in at least 1 of the 3 visits and only 7 eyes (13%) had visits without cavitations. Of the total 152 SD-OCT scans, 141 (93%) had cavitations.

Average cavitation volume was 0.0109 mm³, 0.0113mm³, and 0.0124mm³at baseline, 1, and 2 years, respectively. The average rate of cavitation volume change was +0.0039mm³/year. Using a change of >0.0026mm³as statistically significant (mean + standard deviation of the absolute difference between graders), 10 eyes (20%) had a significant change in cavitation volume during the study; 30% (n = 3) of these eyes had a decrease in cavitation volume and 70% (n = 7) had an increase in cavitation volume. Examples of eyes with expanding and collapsing cavitations are shown (Figure 2 and 3).

Figure 2:

B-scan through the fovea from an eye at baseline (A), 1 year (B), and 2 years (C) of follow-up along with the corresponding segmentation in DOCTRAP (D-F) demonstrating growth of cavitations and extent of ellipsoid zone loss (G-I). The margins of the ellipsoid zone are marked with arrows (white). The corresponding best-corrected visual acuity (BCVA) at those visits declined from 65, 57, to 52 letters. The areas of ellipsoid zone loss were 0.15, 0.23, and 0.24 mm² respectively.

Figure 3:

B-scan from an eye with cavitationsthat collapsed from baseline (A), 1 year (B), to 2 years (C) of follow-up, as shown with corresponding DOCTRAP segmentation (D-F) along with corresponding extent of ellipsoid zone loss (G-I). The margins of the ellipsoid zone are marked with arrows (white).The best-corrected visual acuity (BCVA) during follow-up fluctuated from 66, 71, to 69 letters. The areas of ellipsoid zone loss were 0.67, 0.71, and 0.71 mm² respectively.

Inter-Reader Reliability:

Intra-class correlation between the readers for cavitation volume was excellent (>0.99). The mean absolute difference of cavitation volume measurements between readers was 0.0008± 0.0018mm³, which is less than 7.2% of the total mean cavitation volume.

Functional Correlations:

The 7 eyes with significantly increased cavitation volume had worse BCVA compared to eyes with no change or decreased cavitation volume (71.5 vs. 76.1 ETDRS letters respectively, p=0.03), but no differences in final area or rate of ellipsoid zone loss.The 3 eyes with significantly decreased cavitation volume had similar BCVA (70.6vs. 77.4 ETDRS letters respective, p=0.24), final area (1.94 vs. 0.93 mm², p=0.38)and rate of ellipsoid zone loss compared to eyes with increased or no changes in cavitation volume.

Cavitation volume was negatively correlated to BCVA (r=−0.37, p<0.0001) but not area of ellipsoid zone loss (r=−0.14, p=0.09). In univariate mixed-effects modeling, cavitation volume was negatively predictive of BCVA (p=0.02). In multivariate mixed-effects modeling using both cavitation volume and ellipsoid zone loss, cavitation volume remained a negative predictor of BCVA (p=0.01) while ellipsoid zone loss was no longer a predictor (p=0.15). The rate of cavitation volume change was not correlated to final BCVA, final area of ellipsoid zone loss, or rate of ellipsoid zone loss.

Initial cavitation volume was positively correlated to rate of ellipsoid zone loss (r=0.30, p=0.03). However, in univariate mixed-effects modeling, initial cavitation volume was not a significant predictor of rate of ellipsoid zone loss.

Ellipsoid zone loss was negatively correlated to BCVA (r=−0.20, p=0.0001). When ellipsoid zone loss was divided into quartiles (mean ± standard deviation, 0.32 ± 0.08mm², 0.56 ± 0.07mm², 0.82 ± 0.11mm², 1.62 ± 0.63mm² in each quartile), the mean BCVA in the first through fourth groups was 76.6, 77.4, 77.9, and 72.8 letters respectively. Eyes in the fourth quartile with the largest ellipsoid zone loss had worse BCVA compared to the other groups on Kruskal-Wallis H test followed by pairwise comparisons (p<0.0001). On univariate mixed-effects modeling, ellipsoid zone loss was not a significant predictor of BCVA.

Discussion:

Prior reports in MacTel have suggested that cavitations can change shape over time, disappear on follow-up examination, sometimes re-appear, and collapse in advanced atrophy.[7,10,14,15]To our knowledge, our study is the first to track the natural history progression of cavitations in a large cohort of MacTel patients. We show that the rate of change in cavitations can be reliably quantified using OCT and that typically, there is a slow enlargement of the cavitations over time, likely indicative of ongoing retinal degenerationin this population.Functionally, cavitations appear to indicate worse disease and faster disease progression.Largecavitationspredictworse visual acuity, and those with expanding cavitations have worse BCVA than those with cavitations that collapsed or stayed relatively stable.

In this study, we demonstrate that cavitation volume is predictive of visual acuity loss. This suggests that small cavitations might appear in early disease and can affect the visual acuity as they enlarge in advanced disease.Of note,there were 3 eyes in this cohort with collapsing cavitations that appeared to have worse BCVA and larger area of ellipsoid zone loss. These differences were not statistically significant, however, it is likely that there was inadequate power to detect differences with such a small number of eyes in this category. Larger cohorts are needed to further study this subset of MacTel patients with collapsing cavitations to better understand the functional significance.

Comparatively, ellipsoid zone area is perhaps the most well-established marker of disease progression in MacTel; however, although it correlates highly with scotoma size, it does not necessarily correlate to visual acuity loss, especially in early disease.[16,17]The outer retinal loss in MacTel typically starts in the temporal macula and enlarges to encroach upon the fovea affecting the visual acuity only in late disease.[16,17]Data from our cohort was consistent with this process—the area of ellipsoid zone loss was negatively correlated to BCVA, and patients with the largest area of ellipsoid zone loss had worse vision compared to those with a smaller area of loss.

Interestingly, we did not find a relationship between the area of EZ loss and cavitation volume even though one might expect the two to be correlated (i.e. small cavitations correlate to small area of EZ loss, indicative of early disease). The lack of correlation between EZ loss and cavitation volume could be because of our sample sizewas too small to detect any statistical relationship or that the two are related on a non-linear basis. Alternatively, however, the data could suggest two independent structural pathophysiologic processes at play. For example, a primary Muller cell related process may be responsible for the formation of cavitationswhereas a primary photoreceptor process may be causative for the disappearance of the EZ.

Supportive of this theory, a similar result is seen when analyzing the functional aspects of cavitation volume and ellipsoid zone loss. Specifically, our results suggest that cavitation volume predicts visual acuity loss independent of ellipsoid zone loss.The explanation for this photoreceptor-independent visual acuity loss is unknown but is likely related to the foveal location of cavitations. For example, on qualitative assessment of the 7 eyes with increasing cavitation volume, all of these cavitationsinvolved the center of the fovea but most had an intact EZ at this location (e.g., Figure 2). These data strongly suggest that the visual acuity loss seen in this group is related to the cavitations rather than loss of photoreceptors, which appear structurally intact. One pathophysiologic mechanism explaining the effect of foveal cavitations on visual acuity is the role of Muller cells as waveguides for light propagation in addition to their role as support cells to maintain the retina’s structural integrity.[18–20]In other words, foveal cavitations may disrupt the transmission of light even if the foveal photoreceptors remain structurally intact. This is a provocative finding that argues for the exploration of cavitation volume as an imaging-based biomarker, independent of ellipsoid zone area, for future studies in MacTel.

Importantly, since the presence of cavitations typically precedes the loss of EZ, cavitation volume may be a more useful biomarker earlier in the disease course. Although the data here demonstrate a correlation of cavitation volume with visual acuity, further studies using longitudinal measures of retinal function such as microperimetry may be helpful to clarify the structure-function relationship between cavitations and ellipsoid zone loss even outside of the fovea and shed further light on the pathophysiology of this disease.Until such studies are done, ellipsoid zone loss, which is well correlated with functional measures including loss of retinal sensitivity, should remain the primary end-point for clinical trials.[13,17,21,22]

This study has several limitations including its retrospective design and limited number of patients. Although the present report constitutes one of the larger studies of MacTel,our study may not have been powered to detect the small changes in cavitation volume. Only MacTel patients with a defined area of ellipsoid zone loss were included in the analysis thus the results may not be generalizable to those with mild or very advanced disease. Cavitation size can fluctuate over time, and only images from each year over a two-year span were included; we did not analyze the fluctuations from each monthly visit. Additionally, for the purposes of statistical analysis, we did not distinguish between the intraretinal location of cavitations, whether in the outer or inner retina, or include sub-analyses based on cavitation number, which can be difficult to determine without 3-dimensional segmentation that is technically much more challenging. Incorporation of machine learning automatic algorithms for quantification of the morphological properties of retinal layers and cysts are expected to facilitate analyzing OCT data in large-scale studies.[23,24]Further study in larger cohorts with longer periods of follow-up may be helpful to further characterize the influence of these factors.

Overall, this study illustrates that cavitations can be reliably quantified in patients with MacTel. The relationship between cavitation volume and visual acuity, independent of ellipsoid zone loss, suggests that cavitations may be a useful imaging-based biomarker to follow disease progression and treatment response in clinical trials of MacTel.