Intervisit specificity of a combined retinal ganglion cell index in glaucoma

Abstract

Purpose : Glaucoma results in retinal ganglion cells (RGC) loss and is commonly assessed from structural changes detected by optical coherence tomography (OCT) and functional changes detected by perimetry. Progression analyses are available for both types of data to indicate when change exceeds test-retest variability and have been shown to have acceptable specificity. A structure-function RGC index (RGCI) provides a combined model to monitor progression instead of isolated structure or functional metrics¹. In this short-term, multi-visit clinical study, we investigated the specificity of several progression methods based on RGCI.

Methods : Visual field (VF) and OCT data were acquired from 74 eyes of 74 glaucoma subjects at 5 repeat visits within 4 months, using HFA™ II-i (ZEISS, Dublin, CA) and CIRRUS™ HD-OCT (ZEISS, Dublin, CA). At each visit, SITA Standard 24-2 VFs and Optic Disc 200x200 and Macula 200x200 cube scans were acquired.

RGCI progression was flagged and specificities calculated for two sets of methods. One set used results of linear regression (trend), assuming a visit interval of 0.5 years: a) reg_nonzero – significant non-zero slope; b) reg_less_cross, reg_less_long – slopes significant and less than previously reported cross-sectional and longitudinal rates, respectively^1,2; and c) reg_less_zero_confirm – slope significant and negative at final two visits (confirmation). The second set was patterned on change from baseline (event) based OCT GPA analyses using RGCI variability previously reported³: a) cfb_possible – “Possible” or “Likely” progression at final visit and b) cfb_likely – “Likely” progression at final visit.

Results : Mean age was 63.6 (SD: 10.1; range: 35.7 to 79.6) years. Mean MD was -3.9 (SD: 4.1; range: -18.2 to 1.2) dB. Specificities were 93.2% for reg_nonzero, 98.6% for reg_less_cross, reg_less_long, reg_less_zero_confirm, cfb_possible, and 100% for cfb_likely (see Table 1).

Conclusions : The combined RGCI shows excellent specificity in a short-term, multi-visit study design in a glaucoma population where no progression, only random fluctuation, is expected. As such, the RGCI may be a reasonable parameter for monitoring glaucomatous progression for both event-based and trend-based progression analyses.

References
¹Medeiros et al. Arch Ophth 2012; 130(9). ²Medeiros et al. AJO 2012; 154(5). ³Yu et al. IOVS 2018; 59(9): Abstract 2124.

This is a 2021 ARVO Annual Meeting abstract.

 

Table 1. Summary of specificities and 95% confidence intervals