In vivo ultra-high-sensitive OCT angiography with massive-frame-repeat and hardware and software motion suppression

Optical coherence tomography angiography (OCTA) can visualize flow structures of skin. Its sensitivity depends on the number of frames per position and the time duration of the sequential frame acquisition (i.e., the time window). OCTA typically uses a few frames and a few-to-ten-millisecond time window. Although more frames and a longer time window may enable the visualization of finer vasculature, sample motion hampers OCTA measurement. We demonstrate ultra-high-sensitive (UHS-) OCTA of in vivo human skin using 32 frames per position and a very long (a few seconds) time window. A custom scanner with a sample-fixation tip and an image-correlation-based motion correction were used to restrain sample motion. The outer forearm of 10 subjects was measured using a 1,310-nm swept-source OCT. A 6 × 6 mm² lateral field was divided into 8 blocks. Each block was scanned by a 32-repeating raster scan. The time window was 6.35 seconds. We adopted three modalities for UHS-OCTA: the logarithmic intensity variance (LIV), the late OCT auto-correlation decay speed (OCDSl) and multiple-frame correlation-mapping OCA (cmOCA). We also computed complex-correlation-based OCTA. These three types of UHS-OCTAs were individually observed and fused into a pseudo-color UHS-OCTA. For reference, the sample was also scanned using the standard 4-frame OCTA protocol, and standard OCTA was computed. En face average projections were generated at depth slabs of 0 to 200-μm, 200 to 400-μm, 400 to 600-μm, and 600 to 800-μm from the skin surface. Subjective observation clarified that pseudo-color UHS-OCTA provides very high contrast visualization of fine vessels and revealed significantly more vessels than standard OCTA. The vessel density of LIV-based UHS-OCTA and standard OCTA was computed. The UHS-OCTA showed significantly higher vessel density than standard OCTA (p < 0.001 for all slabs).