Abstract
Purpose :
To test and identify diagnostic accuracy of a single OCTA in Central Asians for differentiating among normal subjects, POAG suspects and POAG compared to clinical examination.
Methods :
Data from Central Asians (Kazakh, Korean, N=59) with confirmed POAG, POAG suspects and healthy subjects age 27 -76 yr were collected. Patients with different stages of glaucoma were included. Visual acuity, refraction, IOP (non contact tonometer, Shin-Nippon NCT -200), slit lamp and fundus examination, ultrasound biometry of lens and vitreous (Alcon Ultrascan B-Scan + A-scan),visual field (HFI 750i, Carl Zeiss), anterior segment (AngioVue OCTA ,Optovue Inc.) were obtained. Anterior chamber parameters were measured. The optic nerve was measured and analyzed in all patients with OCTA (AngioVue, Optovue Inc.). Parameters included RNFL analysis, ONH analysis, GCC analysis, TSNIT NDB Reference.
OCTA images interpretation with a single-page report were exported and analyzed by independent, experienced glaucoma imaging expert, according to a specified grading protocol. The OCTA reader was blinded to any clinical patient information, such as the patient’s medical history, visual acuity, IOP, fundus examination, etc.. The OCTA reader assessment was compared to the standard of care clinical examination which had access to exam and all ancillary testing.
Results :
An OCTA analysis gave an overall classification rate of 85%.
With two outputs (normal subject, POAG suspect) of OCTA analysis the specificity was 87% and sensitivity 91.4%.
The specificity of OCTA for the detection of POAG was 86.6% with sensitivity of 92.5%
Conclusions :
A single OCTA assessment demonstrated good diagnostic accuracy when compared to comprehensive exam in its ability to differentiate glaucoma patients from glaucoma suspect and normal subjects in Central Asians.
This findings demonstrates that OCTA is a valuable adjunctive tool to assist in glaucoma diagnosis and may provide a method for diagnosis remotely which has value during the COVID pandemic.
This is a 2021 ARVO Annual Meeting abstract.