Abstract
Purpose:
Research studies utilizing anterior segment OCT imaging have traditionally focused on acquiring either two or four images due to limitations in OCT technology. Recent technological advances permit the acquisition and analysis of up to 64 images per eye per scanning session. While the number of OCT images that can be acquired is rapidly increasing, the impact of analyzing an increased number of OCT images on angle measurements is unclear. In theory, increasing the number of images makes the measurements more sensitive to anatomical variations. This benefit, however, is decreased by the reliance of current analysis methods on manually identifying the scleral spur in each OCT image, which incrementally increases both the time needed to analyze the data and the measurement error associated with the analysis. This study examines the relationship between the number of anterior segment OCT images analyzed and angle measurement values obtained in order to determine the impact of analyzing all 64 OCT images.
Methods:
We analyzed 2, 4, 8, 16, 32, and 64 OCT images from one eye each of 10 patients randomly selected from the Chinese American Eye Study (CHES). We acquired measurements of 10 different angle parameters, including angle opening distance (AOD) and angle recess area (ARA). We then calculated the mean and standard deviation for each parameter across all 10 patients while varying the number of OCT images analyzed.
Results:
An increase in the number of OCT images analyzed caused angle measurement means and standard deviations to rapidly approach the values obtained by analyzing 64 OCT images. The deviation from the 64 image angle measurement mean for all parameters was 12% with 2 images analyzed, 4% with 4 images, and 1% or less with 8 or more images. The standard deviation was greater by 42% with 2 images analyzed, 20% with 4 images, and 6% or less with 8 or more images.
Conclusions:
The acquisition of additional OCT images should in theory increase the accuracy of anterior segment measurements, but this benefit is decreased by the measured error associated with manually identifying the scleral spur in each image. The results of this study suggest that analyzing 8 images provides angle measurements that are nearly identical to those obtained by analyzing 64 images, and that additional time spent provides significantly diminished returns per image analyzed.