Test-retest Variability of Optical Coherence Tomography (OCT) and OCT Angiography in a Mobile OCT Module and level of Agreement with Conventional Table-top OCT

Abstract

Purpose : The FLEX module(Heidelberg Engineering) is equipped with the SPECTRALIS OCT2 technology mounted on an extendable arm, enabling OCT assessment in reclined and supine patients. We assessed the test-retest variability of OCT and OCT angiography(OCTA) with the FLEX in different positions and agreement with the table-top OCT.

Methods : 15 healthy subjects underwent macular imaging using the FLEX whilst sat upright and lying down, each repeated 3 times. The same protocol was repeated on the table-top OCT. OCTA macula was additionally performed in 13 subjects. Central macular thickness(CMT) and thickness of the four inner quadrants (ETDRS) around the center-point were compared for all scans. Fractal dimension(FD) and vessel density(VD) of the superficial vascular plexus were used to compare variability and agreement of OCTAs. Images were skeletonized and fractal box-counting analysis was conducted using MATLAB. VD was the ratio of black to white pixels in binary images. Coefficient of variation(COV) was used to evaluate the effect of FLEX positioning on within-subject variability(ANOVA), and Bland-Altman plots were used to evaluate agreement between the FLEX and table-top OCT.

Results : Within-subject variability of macular thickness for sitting and lying position(FLEX), and on the table-top was low(<2% COV). There was a statistically significant difference in variability across macular thickness scans in the inferior and nasal quadrants across the 3 scan positions(p<0.001), with the largest variability seen in the lying position(COV 2.6±1.5%, approximately 9µm±5µm). The mean difference and limits of agreement(LoA) between FLEX and table-top macular thickness measurements were 4µm(1.5-6.5), 2.4µm(-2.2–7.0), -0.2µm(-2.4–2.0),1.9µm(-1.9-5.7),-0.4µm(-4.1–3.3) for central, superior, inferior, nasal and temporal quadrants respectively. Mean difference(LoA) for FD was -0.016(-0.039–0.007) and for VD was 0.003(-0.012–0.018).

Conclusions : Within-subject variability on the FLEX module was low regardless of positioning and similar to variability on the table-top device. There is good agreement for macular thickness and retinal vessel measurements between the FLEX and table-top OCT. These findings suggest the FLEX module is a reliable device for acquiring macular OCT and OCTA scans, and can be used interchangeably with the table-top OCT.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.