Purchase this article with an account.
Damon Wong, Jacqueline Chua, Bingyao Tan, Xinwen Yao, Rachel S. Chong, Chelvin Sng, Rahat Husain, Tin Aung, Leopold Schmetterer; Evaluating the Structure-Function Relationship using Structural and Vascular Measures in Glaucoma. Invest. Ophthalmol. Vis. Sci. 2021;62(8):3383.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the global structure-relationship with optical coherence tomography (OCT) structural measures and OCT angiography (OCTA) vascular measures in subjects with primary open angle glaucoma (POAG).
147 eyes from 101 Chinese subjects with POAG underwent OCT and OCTA scans centered on the optic nerve head, and 24-2 visual field testing using standard automated perimetry. A customized algorithm was applied to remove major vessels from the OCTA scans. For each eye, the mean circumpapillary retinal nerve fiber layer thickness from OCT (MRNFL) and mean binarized capillary density from OCTA (MCD) were calculated. Multi-variate regression analysis was performed with the visual field mean deviation (MD) as the dependent variable, and MRNFL and MCD as the predictor variables. Likelihood ratio testing was used to assess the fit of the model with MRNFL and MCD against univariate models trained using either measures separately.
Mean age of the subjects was 63.48±1.05 years and mean visual field severity was -3.74±0.25 dB. Coefficients of MRNFL and MCD from the multi-variate regression analysis were both significant (P<.001). Likelihood ratio testing showed that the multi-variate model was significantly better (P<.001) than the univariate models using MCD or MRNFL. Predicted mean deviation from the multi-variate model achieved a Pearson correlation coefficient of 0.48 with MD, compared to 0.40 with MCD and 0.37 with MRNFL.
Structure-function modelling using both structural and vascular measures was better than univariate models using either structural or vascular measures, suggesting that combining both can provide complementary information to improve the structure-function relationship.
This is a 2021 ARVO Annual Meeting abstract.
This PDF is available to Subscribers Only