Abstract
Purpose:
Traditionally choroidal vasculature is imaged using indocyanine green (ICG) angiography, which is an invasive method and provides poor resolution images of choroidal vessels. With improved visualization of choroidal vessels on enhanced depth imaging using SD-OCT, quantitative measurement of choroidal vasculature is possible. Manual measurements of choroidal vasculature have been recently reported, which is very time consuming and intra observer variability is high, hence we have built an automated algorithm for detection and measurement of area or diameter. We hereby present the validation of results of our newly constructed automated algorithm for individual choroidal vessel measurements.
Methods:
A retrospective analysis of 30 eyes of 17 subjects who underwent enhanced depth imaging using SD-OCT was performed. Choroidal vessel measurements including area and diameter were obtained manually and by using automated algorithm from OCT images. Observers were masked to each other's and their own previous readings. Intraobserver repeatability and correlation between automated and manual technique was calculated. The agreement between the intraobserver measurements or interobserver measurements was assessed using the concordance correlation coefficient (CCC). Bland-Altman plots were used to assess the clinically relevant magnitude of the differences between the measurements and techniques.
Results:
Intraobserver CCC for diameter and area for manual technique was 0.5 and 0.6 respectively. CCC between manual and automated segmentation for diameter and area was 0.49 (95% CI, 0.44-0.54) (p=0.3) and 0.44 (95% CI, 0.41-0.47) (p=0.1) respectively. Mean difference between two techniques for diameter and area was 0.15±0.13 mm and 0.002±0.08 mm2, which is clinically insignificant. Coefficient of reproducibility (CR) for diameter and area was 0.26 mm (95% CI, 0.21 - 0.30 mm) and 0.163 mm2 (95% CI, 0.13 - 0.19 mm2).
Conclusions:
Present study reports a clinically insignificant difference between manual and automated technique. It also validates that our automated choroidal vessel measurement algorithm. It is proposed that it can be used as an independent tool for vessel measurements for detection of pathologies.
Keywords: 549 image processing •
550 imaging/image analysis: clinical •
688 retina