Purpose:
Spectral domain optical coherence tomography (SDOCT) has demonstrated clinical potential for in vivo high-resolution and high-speed imaging of human retinal structure. Advances in Doppler SDOCT have demonstrated several image acquisition schemes that enabled real-time, high-resolution, volumetric display of flow maps for the diagnosis of various retinal pathologies. Current Doppler SDOCT techniques are based on calculating the Doppler frequency shift of flow at a single spatial position from the corresponding temporal phase shifts, thus requiring oversampling A-Scans at each lateral position. While providing flow velocities and 3D flow maps, these techniques are limited in scanning speed due to inherent oversampling and require manual segmentation of vessels. Here we demonstrate single-pass volumetric flow imaging for high-speed, segmentation-free SDOCT of human retinal structure and vessels.
Methods:
Single-pass volumetric flow imaging was implemented on a high-speed SDOCT retinal system. An axial carrier frequency was imposed across each lateral B-Scan by offsetting the scanning sample beam on the lateral scanning mirror. By taking a Hilbert transform of the lateral component of the acquired signal, any flow in the image with unidirectional Doppler frequency greater than the carrier frequency was imaged to one half of the imaging plane while the non-moving structure was imaged onto the opposite half.
Results:
3D flow and structure volumes (2048x1000x200pix) were acquired at the full A-Scan rate of 17.5kHz. Volumes were acquired for positive and negative carrier frequencies yielding in vivo 3D volumetric maps of bidirectional flow at the human optic nerve-head (Fig. 1). Flow maps showed vessel structure in areas of high flow velocities identical to summed voxel projections (SVP) and fragmented reconstructions in smaller vessels.
Conclusions:
Single-pass retinal flow imaging was demonstrated using an off-pivot sample beam. While several approaches for creating Doppler flow maps have been demonstrated, single-pass flow imaging enables 3D volumetric vessel reconstruction without manual segmentation and oversampling.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • retina • optic flow