Foveal Topographic and Volumetric Measures: New Biomarkers for Detection of Early Glaucoma

Abstract

Purpose : To test the hypothesis that newly developed topographic and volumetric foveal outcome measures can discriminate patients with early perimetric glaucoma from normal subjects.

Methods : We used new measures defining inner foveal shape and central macular volume that were recently developed based on parametric modeling of central macula with cubic Bézier (Figure). Global, hemiretinal, and quadrant parameters were calculated for a group of 50 glaucoma eyes (50 patients) with MD of -6 dB or better and 181 eyes of 96 normal subjects. 8×8 arrays of ganglion cell layer (GCL) and inner plexiform layer (IPL) measurements within 3°×3° superpixels were exported and GCIPL thickness parameters were estimated for 5 macular sectors in the superior and inferior hemiretinas. Multivariate logistic regression and area under ROC curves were used to compare the performance of foveal topographic and volumetric measures to that of thickness measurements.

Results : The average (±SD) visual field mean deviation was –3.6 (±1.5) in the glaucoma group. The main global parameters discriminating glaucoma from normal eyes were the macular rim volume and the area of the maximum foveal slope (AUC:0.786). Among the regional topographic/volumetric parameters, a combination of four temporal hemiretinal parameters (rim height, inner rim volume, foveal pit depth, and pit volume) performed best for discrimination of glaucoma from normal eyes (AUC:0.903; 95% CI=0.858-0.948). A combination of 13 regional parameters selected with elastic net regression resulted in AUC of 0.931 (95% CI:0.899-0.966). In comparison, the AUC for the best GCIPL thickness parameter (inferior sector 2) was 0.895 (95% CI:0.844-0.946).

Conclusions : The newly defined topographic and volumetric foveal biomarkers are able to detect early perimetric glaucoma with clinically relevant performance. Such biomarkers do not depend on segmentation of intraretinal layers and may be helpful where adequate segmentation of macular layer cannot be achieved; they may be used to define macular phenotypes in glaucoma and could assist for detection of disease progression in the later stages of glaucoma.

This is a 2021 ARVO Annual Meeting abstract.

 

Description of macular topographic and volumetric measures used in the study. For details see: Yadav SK et al. CuBe: parametric modeling of 3D foveal shape using cubic Bézier. Biomed Opt Express. 2017;8(9):4181-4199.