Longitudinal Ellipsoid Zone Dynamics Following Macular Hole Repair in the DISCOVER Study: Structure-Function Assessment

Abstract

Purpose

To evaluate association of baseline macular hole (MH) geometric features and longitudinal ellipsoid zone (EZ) integrity with visual acuity (VA) outcome following surgical repair.

Methods

This was a post-hoc analysis of eyes in the DISCOVER study undergoing vitrectomy repair for MH. Anatomical and functional data were collected through one year post-operatively. An automated retinal layer segmentation platform was utilized for the assessment of outer retinal metrics and volumetric reconstruction of MH. Association of longitudinal EZ features and baseline MH height, width, and volume with VA outcomes were investigated.

Results

84 eyes with MH were included. Mean baseline VA was 20/114 and increased to 20/45 (p<0.001) at postoperative month 12 (N=45). Successful MH closure was achieved in 98.8% of cases. EZ integrity metrics significantly improved from baseline (p = 0.002) and post-operative month 1 (p<0.001) to post-operative month 12. EZ metrics independently correlated with VA at all follow-up visits (p <0.05). Increased baseline MH width and volume negatively correlated with VA at post-operative month 12 (p < 0.001). Pre-operative VA and EZ integrity on OCT were predictors for post-operative VA.

Conclusions

Baseline MH volumetric parameters and EZ parameters were associated with VA outcomes following repair.

Summary Statement

Baseline macular hole (MH) geometric features and longitudinal ellipsoid zone integrity are associated with visual acuity outcomes following intraoperative OCT-assisted MH repair.

INTRODUCTION

Full-thickness macular hole (MH) most commonly results from vitreomacular traction with subsequent foveal defect with absence of all retinal layers from internal limiting membrane (ILM) to retinal pigment epithelium (RPE) at fovea. Symptoms of MH include distorted vision and central vision loss. The prevalence of MH is estimated to be 0.02 % to 0.33%.¹ Vitrectomy with ILM peeling is the mainstay of treatment with a high rate of success. Optical coherence tomography (OCT) is the current gold standard for the diagnosis, staging and monitoring of MHs.²

OCT has enabled new insights into MH features and associated retinal findings. This includes insight to the pathogenesis of MH and allowed for enhanced feature extraction of MH geometric features.^3–5 Linear measurements (e.g., width, height) have been associated with surgical success and visual outcomes.^{4, 6–8} Novel software analysis systems now enable unique assessment of specific retinal and MH features. This includes MH volumetric segmentation and ellipsoid mapping/integrity assessment.^9–11 These higher-order features (e.g., MH volume, ellipsoid zone (EZ) alterations, and ELM features) have been associated with successful closure, foveal normalization, and visual acuity.^{3, 12–14}

The utilization of intraoperative OCT (iOCT) has provided new information related to the anatomic alterations that occur during surgical interventions.^15–17 In MH surgery, iOCT has identified subclinical alterations that occur both to MH geometry and to the EZ/RPE relationship.^{3, 9, 10, 18} Although these features have been evaluated in relationship to the early anatomic and functional outcomes, limited data exists on the predictive value of these features at baseline for long-term outcomes and the long-term dynamics that occur in these features following surgical repair.

This study provides an evaluation of higher order OCT features in subjects in the DISCOVER study that underwent MH repair and serial OCT imaging to evaluate the anatomical outcomes and their associations with functional outcomes.

METHODS

The DISCOVER study is propspective IRB-aprpoved clinical study evaluating the feasibility and utility of microscope-integrated intraoperative OCT systems. All subjects gave written informed consent. This report included all patients enrolled in DISCOVER who underwent surgery for MH repair.¹⁹ For this analysis, eyes were included that had availability of both preoperative spectral domain OCT (SD-OCT) and post-operative SD-OCT at any of 3, 6, or 12 months. Exclusion criteria were SD-OCT scans of limited quality not appropriate for analysis. The surgical outcomes of this cohort have been previously reported.¹⁹

OCT Image Analysis

Macular cube scans with 512×128 A-scans were performed using a Cirrus HD-OCT preoperatively and at least at one of post-operative 3, 6, and/or 12 months. OCT data of all available timepoints were exported for image analysis. The OCT data was imported into a multi-layer and MH feature extraction platform as previously described.^{11, 20} (Figures 1 and 2). Automated segmentation was performed for ILM, EZ, RPE and MH segmentation. All time-points were reviewed. Automated segmentation errors were corrected as needed by a trained image analyst. Multiple MH and retinal parameters including MH volume, height, width and EZ-RPE thickness were exported and evaluated for analysis.

Figure 1. Segmentation Metrics.

(A) En face OCT image of a representative case with macular hole, demonstrating fovea (plus), central subfield area (small white circle) and central macular area (big white circle). (B) Volumetric rendering demonstrating macular hole in red, intraretinal fluid in blue, ellipsoid zone (EZ) layer in green and retinal pigment epithelium layer (RPE) in pink. Boarders of central subfield (0.5 mm distant from the fovea) and central macular (1.0 mm distant from the fovea) areas are marked with small and big gray circles respectively. Pink areas in central subfield circle correspond to EZ attenuation with underlying visibility of the RPE (pink). Volume calculated between green and pink layers is EZ- RPE volume in the designated location (e.g., central subfield, central macular, macular). (C) Foveal B-scan and corresponding automated segmentation (D) demonstrate volumetric macular hole segmentation in red, internal limiting membrane in blue, EZ layer with yellow line and RPE layer with pink line. Arrowheads and arrows represent the boarders of central subfield and central macular areas respectively. Long vertical green line between ILM (blue line) and RPE (pink line) represents retinal thickness. EZ-RPE thickness is demonstrated by short vertical green line. Central subfield and central macular thickness metrics are the mean thickness in areas marked by small and big circles, respectively.

Figure 2. Volumetric Segmentation.

B-scan (A), automated segmentation (B) and volumetric rendering (C) example of a case with large full thickness macular hole. Red area represents macular hole, blue line represents internal limiting membrane, yellow line represents ellipsoid zone (EZ) and pink line represents retinal pigment epithelium (RPE). EZ-RPE thickness metric is the mean distance between yellow and pink lines, demonstrated by green vertical line. EZ is attenuated between arrowheads. The EZ line (yellow) approaches the top of the RPE line (pink) in areas of EZ attenuation.

Statistical Analysis

Anatomical outcomes assessed by OCT scans at 1, 3, 6 and 12 months were compared with baseline using paired t-tests. Pearson correlations between different OCT variables were derived and it was tested whether they differed significantly from zero. A first multivariable linear regression analyses was performed to assess the effect of baseline OCT variables on best corrected visual acuity (BCVA) change from baseline to post-operative month 12 and a second multivariable linear regression analyses to assess the effect of OCT variables at post-operative month 1 on BCVA at post-operative month 12. R version 3.6.1 (R Project for Statistical Computing) were used for statistical analysis at a significance level of 5%.

RESULTS

Baseline Clinical Features

Overall, 84 eyes from 81 subjects were included in this analysis with a mean age of 70 ± 7.6 years with a mean follow-up time was 10.6 ± 3.5 months.¹⁹ The preoperative mean BCVA was 20/172 (range: 20/40 to CF). The mean preoperative MH width was 364 ± 193 μm, height was 404 ± 83 μm, and volume was 0.099 ± 0.071 mm³.

Visual Acuity and Ellipsoid Zone Integrity Outcomes

As reported, the mean BCVA was 20/68 (P< 0.001, N=74) at 1 month, 20/48 (P<0.001, N=73) at 3, 20/45 (P<0.001, N=51) at 6, and 20/45 (P<0.001, N=45) at 12 months.¹⁹ Hole closure was achieved for 82 of 84 eyes (97.6%) with single surgery. One case of hole closure failure was re-operated with successful MH closure. OCT parameter values at baseline and at post-operative month 12 are shown in Table 1. All outer retinal parameters improved significantly from baseline to month 12. Mean central macular EZ-RPE distance increased from 27.7 ± 5.8 μm to 31.3 ± 7.7 μm at postoperative month 1 (P < 0.001). At postoperative month 12, mean central macular EZ-RPE distance was 35.7 ± 6.1 μm (P < 0.001). Percentage of EZ attenuation decreased from 2.7 ± 2.4 % to 0.7 ± 1.6 % (P < 0.001) (Figure 3).

Table 1.

Comparison of preoperative and postoperative month 12 values for OCT Retinal Parameters

	Preoperative	Postoperative Month 12	P
Central subfield retinal thickness, μm	356 ± 83	262 ± 63	<0.001
Central subfield retinal volume, mm3	0.27 ± 0.06	0.21 ± 0.05	<0.001
Central macular retinal thickness, μm	378 ± 56	300 ± 33	<0.001
Central macular retinal volume, mm3	1.19 ± 0.17	0.94 ± 0.10	<0.001
EZ-RPE thickness ≤ 20 μm, area coverage (%)	2.7 ± 2.4	0.7 ± 1.6	<0.001
EZ-RPE thickness = 0 μm, area coverage (%)	2.1 ± 2.0	0.4 ± 1.1	<0.001
Central subfield EZ-RPE thickness, μm	14.7 ± 12.6	36.3 ± 9.0	<0.001
Central macular EZ-RPE thickness, μm	27.7 ± 5.8	35.8 ± 6.2	<0.001

EZ= ellipsoid zone; RPE= retinal pigment epithelium. Central subfield is the circular region that is defined by a 0.5 mm radius centered at the fovea. Central macula is the circular region defined by a 1.0 mm radius centered at the fovea. Statistical analysis is conducted with paired t-test.

Figure 3. Longitudinal Ellipsoid Zone (EZ) Recovery.

A representative case with full thickness macular hole that underwent vitrectomy with internal limiting membrane peel and visualization of the changes in EZ – retinal pigment epithelium (RPE) thickness in color coded en face EZ integrity mapping and horizontal foveal B-scan images. Area between arrowheads demonstrate EZ attenuation which is represented by pink (0 μ EZ-RPE thickness) on the enface EZ-RPE integrity map. The RPE appears intact.in the foveal B-scan at postoperative month 1 and 3. Decrease in attenuation from postoperative month 1 to postoperative month 3 demonstrates continuing EZ reconstitution following macular hole surgery.

Functional and Anatomic Correlation

OCT parameters were significantly correlated with BCVA throughout the study. At baseline, MH width (r = 0.467, P < 0.001), volume (r = 0.443, P < 0.001) and foveal MH area (r = 0.481, P < 0.001) were negatively correlated with visual acuity. At baseline, multiple outer retinal parameters including mean central subfield EZ-RPE thickness and mean central macular EZ-RPE thickness demonstrated significant correlation with BCVA.

At post-operative month 12, BCVA was significantly correlated with preoperative MH width (r = 0.552, P < 0.001) and MH volume (r = 0.531, P < 0.001) (Table 2). Pre-operative central subfield EZ-RPE thickness (r = −0.459, P = 0.002) and central macular EZ-RPE thickness (r = −0.551, P < 0.001) were also significantly correlated with BCVA at post-operative month 12.

Table 2.

Correlation Coefficients between Best-Corrected Visual Acuity at Postoperative month 12 and OCT Retinal Parameters at Preoperative and Postoperative Month 1.

	Preoperative (N=45)	Postoperative Month 1 (N=39)
Central subfield mean retinal thickness, μm	−0.524	−0.537
Retinal volume, mm3	−0.186*	−0.376
Macular hole central foveal area, μm	0.410	NA
Macular hole height, μm	−0.194*	NA
Macular hole width, μm	0.552	NA
Macular hole volume, mm3	0.531	NA
Macular hole map area, mm2	0.595	0.666
EZ-RPE thickness ≤ 20 μm, area coverage (%)	0.555	0.436
EZ-RPE thickness = 0 μm, area coverage (%)	0.558	0.378
Central subfield EZ-RPE thickness, μm	−0.459	−0.578
Central macular EZ-RPE thickness, μm	−0.551	−0.607
Central subfield ILM-RPE thickness, μm	−0.524	−0.531
Central ILM EZ-RPE thickness, μm	−0.279*	−0.470

EZ= ellipsoid zone; RPE= retinal pigment epithelium. Asterisk indicates a correlation coefficient that is not significantly different from zero. Area Central subfield is equivalent to 0.5 mm distant from the fovea is defined as central subfield and area. Central macula is equivalent to 1.0 mm distant from the fovea is defined as central macula..

Multivariable linear regression analysis on the BCVA change from baseline to post-operative month 12 included baseline BCVA, MH volume, MH width, central macular EZ-RPE thickness and zero-micron EZ-RPE coverage before MH surgery (Adjusted R² = 0.621,). Block-wise selection was used to select MH and EZ integrity variables. Among those variables baseline BCVA (r= 0.723, P< 0.001) and zero-micron EZ-RPE coverage (r = 0.039, P = 0.006) had a significant effect on BCVA change. Multivariable linear regression analysis on BCVA at post-operative month 12 with post-operative month 1 BCVA, central subfield retinal thickness, zero-micron EZ-RPE coverage and mid-subfield EZ-RPE thickness at (Adjusted R² = 0.625,) demonstrated that post-operative month 1 BCVA (r= 0.351, P < 0.001) and macular EZ-RPE volume (r= −0.013, P = 0.006) are significantly correlated with post-operative month 12 BCVA.

DISCUSSION

This analysis demonstrated longitudinal improvement EZ integrity following MH repair. A semi-automated segmentation tool allowed for successful quantitative assessment of EZ-RPE parameters and demonstrated consistent correlation with BCVA preoperatively and postoperatively. Pre-operative MH features including width, length, volume and area correlated significantly with BCVA at baseline and post-operative month 12. Multivariable analysis revealed pre-operative BCVA and zero-micron EZ-RPE coverage percentage to be predictors of BCVA at postoperative month 12. Additionally, BCVA and mid-subfield EZ-RPE volume at postoperative month 1 significantly correlated with BCVA at post-operative month 12.

Multiple studies suggested that MH width is a predictor post-operative visual acuity.^{3, 4, 6, 7} Consistent with previous reports, MH width were negatively correlated with post-operative BCVA in our study. Additionally, volumetric measurement of MH was feasible, and it was also negatively correlated with postoperative BCVA in univariable analysis. In addition to MH parameters, pre-operative EZ-RPE metrics demonstrated significant correlation with BCVA at post-operative month 12.

Multivariable linear regression analysis on change in BCVA with pre-operative variables showed baseline BCVA has the most robust correlation and it was followed with zero-micron EZ-RPE coverage. MH volume did not remain significant in multivariable analysis. Pre-operative EZ integrity is found to be prognostic for vision recovery after MH surgery. Similarly, macular EZ-RPE thickness at post-operative month 1 significantly correlated with post-operative month 12 BCVA. These results highlight the importance of outer retina integrity as a predictive measure for visual outcome expectations from MH surgery.

The correlation between EZ-RPE metrics and BCVA demonstrated at all timepoints in this report was in alignment with previous studies investigation EZ-BCVA relation in other pathologies ^{11, 17} Previous studies with qualitative assessment of EZ continuity following MH repair reported better functional outcomes in eyes with EZ recovery.^12–14 To the best of our knowledge this is the first study where volumetric EZ integrity is quantitively assessed following MH surgery. We reported overall impact of EZ-RPE thickness and EZ attenuation on BCVA at a given time-point.

This study has several limitations. ELM restoration has been shown to be an important preceding parameter of photoreceptor layer healing and improvement in vision. We did not measure ELM metrics which may have provided additional insight. In this study, the EZ integrity and dynamics were the focus and alterations in the RPE were not specifically explored. Further research regarding RPE alterations may also be valuable for future studies. We also did not include post incision data acquired by iOCT in our analysis, because it was not consistently available. The differences in surgical techniques of participation surgeons were not considered for their effects on anatomical outcome.

This study represents the quantitative OCT assessment of the largest cohort that underwent of iOCT assisted MH surgery. Quantitative assessment of MH features and retinal layers pre-operatively and post-operatively provided more insight on expected visual outcomes of MH surgery. EZ mapping enabled to demonstrate the impact of reconstructive changes following successful surgery on visual acuity. Future prospective studies with more detailed outer retinal layer segmentation are needed to further investigate the factors leading to reconstructive changes and better visual outcomes.