Comparison of Longitudinal GCIPL Rates of Change between Two OCT Devices

Abstract

Purpose : To compare longitudinal rates of change in ganglion cell/inner plexiform layer (GCIPL) thickness and proportion of worsening and improving rates derived from Spectralis and Cirrus OCTs in eyes with central or advanced damage at baseline.

Methods : Spectralis and Cirrus OCT macular volume scans from 91 glaucoma eyes (91 patients) with >2 years of follow-up and ≥4 OCT images were exported. GCIPL data from both OCTs were centered on the fovea, registered, and segmented. Data from a central 4.8×4.0mm ellipse were extracted and divided into 6 sectors (Fig. 1). We estimated global and sectoral rates of change with linear regression of GCIPL thickness against time. Outcomes of interest were global and sectoral rates of change from the 2 OCT devices and the proportion of worsening and improving rates (negative vs. positive rates of change with p<0.05). Permutation analyses were used to control specificity with 2.5%ile cutoff point used to define true worsening.

Results : Median (IQR) follow-up time and the number of OCT images were 4.1 (3.2-5) years and 8 (6-11), respectively. Baseline global GCIPL thickness was 57.1μm and 59.5μm for Spectralis and Cirrus OCT. Spectralis rates of change were faster globally (p=0.04) and in the superotemporal (p<0.01) and superior (p=0.04) sectors (Fig. 1). Based on simple linear regression, a higher proportion of significant global/sectoral negative rates was observed with Spectralis in comparison to Cirrus except inferonasally; in contrast, the proportion of improving sectors was similar for both devices except in the inferotemporal sector (Spectralis: 14% vs. Cirrus: 2%). Permutation analyses also demonstrated a significantly higher proportion of deteriorating rates with Spectralis OCT globally and in superotemporal (p<0.001) and inferotemporal (p<0.001) sectors (Fig. 2).

Conclusions : Macular OCT measurements from Spectralis device are more likely to identify faster negative rates of change and to detect significant disease deterioration both globally and in sectors. OCT devices cannot be considered interchangeable with regard to the detection of glaucoma progression based on macular measurements.

This is a 2020 ARVO Annual Meeting abstract.

 

The distribution of global and sectoral GCIPL rates of change for Spectralis and Cirrus OCTs

 

Bar graph shows proportion of worsening GCIPL rates of change (negative rate with p<0.05) globally and in 6 sectors based on permutation analyses