MOTION-FREE OCT IMAGE

[caption id="attachment_5764" align="alignnone" width="553"] Yiwei Chen PhD[/caption] The evolution of optical coherence tomography (OCT) has brought technology with faster scanning speeds that allow more data to be collected faster with a reduction in artefact. However, the scanning process of 3D OCT imaging still takes several seconds, and eye motion artefacts still occur. At the 2016 annual meeting of the Association for Research in Vision and Ophthalmology (ARVO) in Seattle, USA, researchers from the Computational Optics Group, University of Tsukuba, Japan, reported on a technique they have developed for motion-free 3D OCT. Their method uses a Lissajous scanning protocol and a custom motion correction algorithm. “Eye motion artefacts cause two potential problems. First, they can result in missing data that may contain important structures and features. In addition, the motion artefacts can corrupt shape information along the slow scanning direction. Our method provides the true shape in both the horizontal and vertical directions,” said Yiwei Chen PhD, who presented the technique. To assess its performance, imaging of the optic nerve head and macula was performed in five normal dominant eyes of five subjects. Both areas were scanned twice. Data were excluded if there was severe vignetting or blinking. Comparisons of en-face projection images, acquired before and after lateral motion correction, showed the technique had a 90 per cent success rate as only two of the 20 cases exhibited severe artefacts. In the remaining images, continuity of vessels and structure clarity were obvious after applying lateral motion correction. GROUND TRUTH Further demonstrating the performance of the technique, Dr Chen noted perfect co-registration between lateral motion-corrected images and those obtained using scanning laser ophthalmoscopy (SLO). “Because the SLO measurement is so quick, we can ignore the potential for any motion artefact and treat it as a ground truth,” he said. Repeatability of the lateral motion correction was shown through comparison of two motion-corrected en-face projection images from volume OCT obtained in a single eye. “Because each volume measurement takes 5.2 seconds, we would expect eye motion to be different during the two acquisitions. Considering the repeatability of the lateral motion correction, we believe this method is useful for total motion-free OCT angiography,” Dr Chen said. Performance correcting for axial motion was demonstrated by comparing images obtained in the same eye with and without motion correction, serial images obtained with motion correction, and using a horizontal B-scan image taken with a fast raster scan protocol (24 milliseconds) as a “ground truth” reference. Yiwei Chen: cyw198788@gmail.com