Dynamic optical coherence tomography (DOCT) is a label-free technique that visualizes tissue dynamics by analyzing a long-time sequence of OCT images. Although it was successful for in vitro and ex vivo imaging, it is still challenging for in vivo imaging because of the sample motion. We address this issue by developing hardware- and software-based motion suppression methods and demonstrating in vivo DOCT imaging of human skin. The hardware method is a sample fixation spacer. The software method is an image-registration based motion correction. We used logarithmic intensity variance (LIV) method to image the tissue dynamics. LIV was calculated from 32 sequential OCT frames taken within a 6.35s time window. The interframe time interval was 204.8ms and the entire DOCT volume was measured in 52.4s. Furthermore, we measured OCT angiography (OCTA) by standard raster scan with four frame repeats. To quantitively analyzing the improvement of proposed methods, three regions of interests (ROIs), each measuring 176μm (in depth) × 217μm (in lateral direction), was select from one B-scan to calculate the mean LIV. The improvement was assessed by paired t-test. The motion correction methods significantly reduced the high-LIV artifacts and revealed very fine capillary structures that had been buried by motion artefacts. The paired t-test results showed that the combination of fixation spacer and the software correction significantly reduced LIV artifacts (p=0.0052, 0.0137 and 0.0068 for three ROIs).
|