Abstract
Purpose:
We aim to analyze the accurateness of the scan position of a cross-sectional image (B-scan) in multiple follow up examinations on an OCT.
Methods:
The cross-sectional images of the retina were acquired by a Heidelberg Spectralis OCT. For each scan an infrared image (768 x 768 pixel or 1536 x 1536 pixel, 256 gray scales) was internally used to compensate eye-movements which covers a 30° area at the fundus. It was saved along with the B-scans and their positions on the infrared image. This study included 11 subjects age 58 to 80 years at baseline examination. Each subject had 1 baseline and between 4 and 10 follow up measurements at the Spectralis OCT over a time period of 11 to 44 month. These 11 datasets hold 4 subjects with high resolution scans and 7 subjects with scans in high speed mode. All baseline examinations were inspected for a B-scan with at least 2 vessels with diagonal orientation relative to the B-scan and an approximately 90° angle relative to each other. Only the B-scans in one dataset with the same ID number where further analyzed. The angles of all dominating vessels in the chosen B-scan where measured out by use of the corresponding infrared image. The vessel borders where automatically detected in all B-scans by segmenting the ILM and RPE layer followed by a gradient vector approach to determine the vessel positions. A combined intensity profile based on the already segmented layers was calculated per vessel segment and the border positions where detected by a full width at half height approach. Based on the vessel border positions the centerline and the vessel diameter was calculated. The distance between the vessel centerlines was calculated and corrected by subtracting two times the standard deviation of the involved vessel diameter.
Results:
The mean distance difference over all datasets (n=11) was 0.193 ± 0.099 mm. The smallest distance difference of one dataset (n=5) was 0.028 ± 0.011 mm (max 0.047 mm) whereas the biggest distance difference (n=6) was 0.202 ± 0.085 mm (max 0.339 mm).
Conclusions:
The results demonstrate that the follow up data of a B-scan have to be viewed critically regarding the eyetracking accuracy. A variation of the scan position may lead to incorrect morphometric measurements and a misinterpretation of the therapeutic outcome.
Keywords: 550 imaging/image analysis: clinical •
552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)