Longitudinal Structure-Function Relationship between Macular Vessel Density and Thickness and Central Visual Field in Early Glaucoma

Longitudinal OCTA vessel density loss in macula and optic nerve head in healthy, glaucoma suspect and established glaucoma eyes

BACKGROUND/AIMS

To examine longitudinal optical coherence tomography angiography (OCTA) changes in macula and optic nerve head (ONH) in healthy, glaucoma suspect (GS) and primary open-angle glaucoma (POAG) eyes.

METHODS

Healthy, GS and POAG eyes from Diagnostic Innovations in Glaucoma Study with ≥2 years follow-up and four visits of macular/ONH OCTA imaging were included. Rates of macular wiVD (whole-image vessel density) and ONH wiCD (whole-image capillary density) changes were calculated for each diagnosis group using join mixed-effect modelling. Differences in wiVD/wiCD change rates across diagnoses were examined through pairwise comparison. Relationships of baseline 24-2 visual field (VF) mean deviation (MD) with wiVD/wiCD change rates were evaluated for POAG eyes.

RESULTS

There were 36, 57 and 79 eyes (25, 38 and 50 subjects) in healthy, GS and POAG groups, respectively. Rates of wiVD (range:-0.72 to -0.92 %/year) and wiCD (range:-0.28 to -0.66 %/year) loss were different from zero in all groups (p<0.05). The rates of wiCD loss differed across all diagnosis groups (p<0.001), while wiVD change rates did not increase when comparing healthy to GS eyes (p=0.167). Baseline VF MD showed a significant but modest correlation with the rates of both wiVD and wiCD loss (p<0.05), and the correlation with wiCD change rate was slightly stronger (R2=0.27 vs 0.16).

CONCLUSIONS

In glaucoma, there is earlier microvasculature loss in the ONH than in the macula. Moreover, ONH VD loss shows a slightly stronger association with baseline VF than macular VD. Observing ONH VD loss with OCTA may help to monitor early glaucoma, which should be confirmed by future larger studies.

Longitudinal changes in optical coherence tomography angiography characteristics in normal-tension glaucoma with or without high myopia

PURPOSE

To evaluate the structural, microvascular, and functional progression of normal tension glaucoma (NTG) with or without high myopia by examining longitudinal changes in optical coherence tomography angiography (OCTA) and visual field (VF) parameters.

METHODS

We evaluated 61 NTG eyes and classified 25 of the eyes with axial lengths (ALs) of ≥26 mm as highly myopic. We assessed the rate of change in OCTA parameters, namely radial peripapillary capillary (RPC) vessel density (VD), parafovea VD, deep parafovea VD, retinal nerve fibre layer (RNFL) thickness, and ganglion cell complex thickness. We evaluated the correlation of the rate of change in OCTA parameters with VF loss and AL.

RESULTS

Among the 61 NTG eyes, rates of loss of RPC VD, parafovea VD, deep parafovea VD, and RNFL thickness were significantly different from zero despite the nonsignificant rate of change in VF mean deviation (MD). Changes in these OCTA parameters did not differ significantly in highly myopic NTG eyes. The rate of change in VF MD was significantly correlated with the rate of change in parafovea VD in highly myopic and non-highly myopic NTG eyes. In highly myopic NTG eyes, AL was negatively correlated with the rates of loss of RNFL thickness, VF MD, and VF PSD.

CONCLUSION

NTG eyes with a relatively stable VF exhibited loss of VD and RNFL thickness. VF progression in NTG was correlated with decreasing parafovea VD, indicating a structure-function correlation. Greater AL may indicate faster VF loss and RNFL thinning in highly myopic NTG eyes.

Comparison of the Circumpapillary Structure-Function and Vasculature-Function Relationships at Different Glaucoma Stages Using Longitudinal Data

Purpose

This study investigated the global and regional correlations between longitudinal structure-function (S-F) and vasculature-function (V-F) data using circumpapillary retinal nerve fiber layer thickness (cpRNFLT) measurements from optical coherence tomography (OCT), circumpapillary vessel density (cpVD) from OCT angiography (OCTA), and the corresponding visual field mean sensitivities at different glaucoma stages.

Methods

A total of 107 eyes from 107 glaucoma patients with progressive visual field (VF) changes followed up for an average of 3.33 ± 1.39 years were enrolled, including early-to-moderate (51 eyes) and advanced (56 eyes) stages. The rates of longitudinal change in the VF mean deviation (MD), cpRNFLT, and cpVD were evaluated using linear mixed-effects models and compared between different glaucoma stages. Longitudinal global and regional S-F and V-F relationships were assessed by repeated measures correlation analysis by glaucoma stage.

Results

No significant differences were found in the rates of VF MD and cpVD changes (P > 0.05) between the two glaucoma stage groups. CpRNFLT decreased more rapidly in the early-to-moderate stage group (P < 0.001) in which significant longitudinal global and regional correlations were found in both S-F and V-F relationships (all P < 0.05), except for the nasal sector. Significant global and regional correlations were only found in V-F relationship in advanced stage cases (all P < 0.05).

Conclusions

Significant longitudinal V-F relationships exist globally and regionally regardless of glaucoma stage but no longitudinal S-F relationship is present in advanced glaucoma. Longitudinal follow-up of cpVD parameters may be useful for monitoring glaucomatous VF progression at all disease stages.

Deep Learning Estimation of 10-2 Visual Field Map Based on Macular Optical Coherence Tomography Angiography Measurements

PURPOSE

To develop deep learning (DL) models estimating the central visual field (VF) from optical coherence tomography angiography (OCTA) vessel density (VD) measurements.

DESIGN

Development and validation of a deep learning model.

METHODS

A total of 1051 10-2 VF OCTA pairs from healthy, glaucoma suspects, and glaucoma eyes were included. DL models were trained on en face macula VD images from OCTA to estimate 10-2 mean deviation (MD), pattern standard deviation (PSD), 68 total deviation (TD) and pattern deviation (PD) values and compared with a linear regression (LR) model with the same input. Accuracy of the models was evaluated by calculating the average mean absolute error (MAE) and the R2 (squared Pearson correlation coefficients) of the estimated and actual VF values.

RESULTS

DL models predicting 10-2 MD achieved R2 of 0.85 (95% confidence interval [CI], 74-0.92) for 10-2 MD and MAEs of 1.76 dB (95% CI, 1.39-2.17 dB) for MD. This was significantly better than mean linear estimates for 10-2 MD. The DL model outperformed the LR model for the estimation of pointwise TD values with an average MAE of 2.48 dB (95% CI, 1.99-3.02) and R2 of 0.69 (95% CI, 0.57-0.76) over all test points. The DL model outperformed the LR model for the estimation of all sectors.

CONCLUSIONS

DL models enable the estimation of VF loss from OCTA images with high accuracy. Applying DL to the OCTA images may enhance clinical decision making. It also may improve individualized patient care and risk stratification of patients who are at risk for central VF damage.