A deep learning system can detect referable GON with high sensitivity and specificity. Coexistence of high or pathologic myopia is the most common cause resulting in false-negative results. Physiologic cupping and pathologic myopia were the most common reasons for false-positive results.

Segmentation and measurement of GCIPL thickness were successful in 50 of 51 subjects. All ICCs ranged between 0.94 and 0.98, but ICCs for average and superior GCIPL parameters (0.97-0.98) were slightly higher than for inferior GCIPL parameters (0.94-0.97). All COVs were <5%, with 1.8% for average GCIPL and COVs for superior GCIPL parameters (2.2%-3.0%) slightly lower than those for inferior GCIPL parameters (2.5%-3.6%). The TRTSD was lowest for average GCIPL (1.16 μm) and varied from 1.43 to 2.15 μm for sectoral GCIPL CONCLUSIONS: The Cirrus HD-OCT GCA algorithm can successfully segment macular GCIPL and measure GCIPL thickness with excellent intervisit reproducibility. Longitudinal monitoring of GCIPL thickness may be possible with Cirrus HD-OCT for assessing glaucoma progression.

No abstract

No abstract