Line-Scan Spectral-Domain Optical Coherence Tomography for Cellular Resolution Structural and Vascular Imaging of Biological Tissues

Abstract

Optical coherence tomography (OCT) is an optical interferometric technique for non-invasive contactless imaging of the cellular-level structures of biological tissues. However, the application of OCT for in-vivo volumetric cellular resolution imaging of the human anterior eye is challenging due to artifacts induced by involuntary eye motion and the contradictory requirements for high lateral resolution and extended depth of focus. This thesis addresses these challenges by developing: (i) a broadband line-scan (LS) spectral-domain (SD) OCT system that combines micrometer-scale spatial resolution and ultrafast image acquisition rate; (ii) an image reconstruction method for restoring the diffraction-limited lateral resolution of the LS SD-OCT system along a large depth range. In addition, a novel flow velocimetry method is developed for extending the LS SD-OCT system's functionality.

The novel LS SD-OCT system combines a broadband light source and an ultrafast area camera to achieve a nearly isotropic spatial resolution of ~2.3 𝜇m in free space and an image acquisition rate of up to 3000 frames/second. The central sensitivity of the system is 92 dB near the zero optical delay with a 6 dB rolloff depth range of 0.78 mm. The system's performance was evaluated by imaging in-vivo a healthy volunteer's cornea and limbus. The motion artifacts are not noticeable in most volumetric images. Cornea epithelial cells, sub-basal corneal nerves, keratocytes in the stroma, cornea endothelial cells, palisaded of Vogt (POV) in the limbus, limbal epithelial cells between the POVs, and hyperreflective line structures underneath POVs are resolved in the 3D images within a limited depth range.

Digital adaptive optics (DAO) is commonly used to correct the monochromatic wavefront aberrations in heterodyne imaging techniques. We show that interference-induced phase destruction, spatial-spectral crosstalk, and chromatic aberrations are the three primary artifacts obscuring diffraction-limited resolution restoration with standard DAO in images acquired with the broadband LS SD-OCT system. We demonstrate that phase destruction can be minimized with appropriate optics alignment. In addition, we show that spatial-spectral crosstalk and chromatic aberrations can be efficiently suppressed by registration of monochromatic aberration corrected sub-band tomograms. The image reconstruction method for recovering the diffraction-limited lateral resolution has been validated using different test objects such as standard resolution target, microbeads phantom, and different biological tissues imaged ex-vivo.

The novel decorrelation-based transverse flow velocimetry, developed specifically for LS SD-OCT, extends the current dynamic light scattering flow speed measurement techniques. We take advantage of the phase stability within each B-scan and digitally generate a low-resolution OCT signal. By introducing the lateral resolution contrast in the temporal autocorrelation function of the OCT signals, this method allows for precisely measuring the transverse intralipid flow velocity in the low time resolution and low SNR conditions. The proposed method was validated by comparing with the phasebased OCT velocity measurement methods in phantom-based experiments.

The combination of the broadband LS SD-OCT system and the proposed image reconstruction method allows aberration-free volumetric cellular resolution imaging of biological tissues. The high image acquisition rate suppresses the motion-induced image artifacts, making high-resolution in-vivo imaging of the human eye in an extended depth of focus possible. The novel flow velocimetry can be used to monitor the flow dynamic, which extends the LS SD-OCT's functionality.