Abstract
Purpose :
Color fundus photography represents the historical gold standard for diagnosis of geographic atrophy (GA)1. With the popularization of optical coherence tomography angiography (OCT) imaging at primary eye care, this study aimed to evaluate the identification of GA using en face OCT images against color fundus photo (CFP).
Methods :
This study comprised a total of 548 eyes that were seen in optometry practices. CFP and 6mmx6mm macular scan were both captured by Maestro2 (Topcon Healthcare, Tokyo, Japan) which is an OCT-fundus camera system. All 548 pairs of CFPs and en face OCT images were graded by one masked grader to determine the presence of GA, with the CFPs and OCT images being graded respectively. Images were graded by a four-point scale: GA positive, GA negative, uncertain, and poor quality. Response rate for effectiveness of OCT compared to CFP was calculated as proportion of eyes with en face OCT graded as GA positive, together with those where both en face OCT images and color fundus photo were graded GA negative, or OCT enface as GA negative, uncertain, or poor quality when the fundus images were also uncertain or of poor quality. Agreement of clinically useful grading was calculated using those images with definitive gradings for both modalities.
Results :
The response rate for GA identification by OCT en face compared to CFP was 86.0%. Agreement rate of clinically useful grading in the 159 eyes with definitive grades in both modalities was 86.0% (kappa of 0.702, representing substantial agreement). GA was identified in 83 CFP and 65 en face OCT images. Among them, 46 were seen in both, 37 were seen in CFP only, and 19 were seen in en face OCT only. Of the 197 uncertain cases by CFP, 111 (56.3%) had certain diagnosis (either positive or negative) by en face OCT, while 47 out of 63 uncertain cases with en face OCT (74.6%) were confirmed by CFP. The poor-quality rates of CFP and en face OCT were similar (23.9% vs. 23.7%).
Conclusions :
En face OCT image could be used to identify GA and had fair agreement with CFP. Even while en face OCT was not as efficient as CFP, it was a useful supplement that improved the identification of GA. Combining CFP and OCT that captured by a single device could be a convenient way to enhance clinical decision-making.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.