Visible-light optical coherence tomography oximetry based on circumpapillary scan and graph-search segmentation

Degeneration of Melanin-Containing Structures Observed Longitudinally in the Eyes of SOD1-/- Mice Using Intensity, Polarization, and Spectroscopic OCT

Purpose

Melanin plays an important function in maintaining eye health, however there are few metrics that can be used to study retinal melanin content in vivo.

Methods

The slope of the spectral coefficient of variation (SSCoV) is a novel biomarker that measures chromophore concentration by analyzing the local divergence of spectral intensities using optical coherence tomography (OCT). This metric was validated in a phantom and applied in a longitudinal study of superoxide dismutase 1 knockout (SOD1^−/−) mice, a model for wet and dry age-related macular degeneration. We also examined a new feature of interest in standard OCT image data, the ratio of maximum intensity in the retinal pigment epithelium to that of the choroid (RC ratio). These new biomarkers were supported by polarization-sensitive OCT and histological analysis.

Results

SSCoV correlated well with depolarization metrics both in phantom and in vivo with both metrics decreasing more rapidly in SOD1^−/− mice with age (P < 0.05). This finding is correlated with reduced melanin pigmentation in the choroid over time. The RC ratio clearly differentiated the SOD1^−/− and control groups (P < 0.0005) irrespective of time and may indicate lower retinal pigment epithelium melanin in the SOD1^−/− mice. Histological analysis showed decreased melanin content and potential differences in melanin granule shape in SOD1^−/− mice.

Conclusions

SSCoV and RC ratio biomarkers provided insights into the changes of retinal melanin in the SOD1^−/− model longitudinally and noninvasively.

Translational Relevance

These biomarkers were designed with the potential for rapid adoption by existing clinical OCT systems without requiring new hardware.

Eye-specific 3D modeling of factors influencing oxygen concentration in the lamina cribrosa

Our goal was to identify the factors with the strongest influence on the minimum lamina cribrosa (LC) oxygen concentration as potentially indicative of conditions increasing hypoxia risk. Because direct measurement of LC hemodynamics and oxygenation is not yet possible, we developed 3D eye-specific LC vasculature models. The vasculature of a normal monkey eye was perfusion-labeled post-mortem. Serial cryosections through the optic nerve head were imaged using fluorescence and polarized light microscopy to visualize the vasculature and collagen, respectively. The vasculature within a 450 μm-thick region containing the LC - identified from the collagen, was segmented, skeletonized, and meshed for simulations. Using Monte Carlo sampling, 200 vascular network models were generated with varying vessel diameter, neural tissue oxygen consumption rate, inflow hematocrit, and blood pressures (arteriole, venule, anterior boundary, and posterior boundary). Factors were varied over ranges of baseline ±20% with uniform probability. For each model we first obtained the blood flow, and from this the neural tissue oxygen concentration. ANOVA was used to identify the factors with the strongest influence on the minimum (10th percentile) oxygen concentration in the LC. The three most influential factors were, in ranked order, vessel diameter, neural tissue oxygen consumption rate, and arteriole pressure. There was a strong interaction between vessel diameter and arteriole pressure whereby the impact of one factor was larger when the other factor was small. Our results show that, for the eye analyzed, conditions that reduce vessel diameter, such as vessel compression due to elevated intraocular pressure or gaze-induced tissue deformation, may particularly contribute to decreased LC oxygen concentration. More eyes must be analyzed before generalizing.

Computational refocusing of Jones matrix polarization-sensitive optical coherence tomography and investigation of defocus-induced polarization artifacts

Here we demonstrate a long-depth-of-focus imaging method using polarization sensitive optical coherence tomography (PS-OCT). This method involves a combination of Fresnel-diffraction-model-based phase sensitive computational refocusing and Jones-matrix based PS-OCT (JM-OCT). JM-OCT measures four complex OCT images corresponding to four polarization channels. These OCT images are computationally refocused as preserving the mutual phase consistency. This method is validated using a static phantom, postmortem zebrafish, and ex vivo porcine muscle samples. All the samples demonstrated successful computationally-refocused birefringence and degree-of-polarization-uniformity (DOPU) images. We found that defocusing induces polarization artifacts, i.e., incorrectly high birefringence values and low DOPU values, which are substantially mitigated by computational refocusing.

Visible-Light Optical Coherence Tomography Fibergraphy for Quantitative Imaging of Retinal Ganglion Cell Axon Bundles

Purpose

To develop a practical technique for visualizing and quantifying retinal ganglion cell (RGC) axon bundles in vivo.

Methods

We applied visible-light optical coherence tomography (vis-OCT) to image the RGC axon bundles, referred to as vis-OCT fibergraphy, of healthy wild-type C57BL/6 mice. After vis-OCT imaging, retinas were flat-mounted, immunostained with anti-beta-III tubulin (Tuj1) antibody for RGC axons, and imaged with confocal microscopy. We quantitatively compared the RGC axon bundle networks imaged by in vivo vis-OCT and ex vivo confocal microscopy using semi-log Sholl analysis.

Results

Side-by-side comparison of ex vivo confocal microscopy and in vivo vis-OCT confirmed that vis-OCT fibergraphy captures true RGC axon bundle networks. The semi-log Sholl regression coefficients extracted from vis-OCT fibergrams (3.7 ± 0.8 mm^–1) and confocal microscopy (3.6 ± 0.3 mm^–1) images also showed good agreement with each other (n = 6).

Conclusions

We demonstrated the feasibility of using vis-OCT fibergraphy to visualize RGC axon bundles. Further applying Sholl analysis has the potential to identify biomarkers for non-invasively assessing RGC health.

Translational Relevance

Our novel technique for visualizing and quantifying RGC axon bundles in vivo provides a potential measurement tool for diagnosing and tracking the progression of optic neuropathies.

Retinal capillary oximetry with visible light optical coherence tomography

Assessing oxygen saturation (sO₂) remains challenging but is nonetheless necessary for understanding retinal metabolism. We and others previously achieved oximetry on major retinal vessels and measured the total retinal oxygen metabolic rate in rats using visible-light optical coherence tomography. Here we extend oximetry measurements to capillaries and investigate all three retinal vascular plexuses by amplifying and extracting the spectroscopic signal from each capillary segment under the guidance of optical coherence tomography (OCT) angiography. Using this approach, we measured capillary sO₂ in the retinal circulation in rats, demonstrated reproducibility of the results, validated the measurements in superficial capillaries with known perfusion pathways, and determined sO₂ responses to hypoxia and hyperoxia in the different retinal capillary beds. OCT capillary oximetry has the potential to provide new insights into the retinal circulation in the normal eye as well as in retinal vascular diseases.

Deep spectral learning for label-free optical imaging oximetry with uncertainty quantification

Measurement of blood oxygen saturation (sO2) by optical imaging oximetry provides invaluable insight into local tissue functions and metabolism. Despite different embodiments and modalities, all label-free optical-imaging oximetry techniques utilize the same principle of sO2-dependent spectral contrast from haemoglobin. Traditional approaches for quantifying sO2 often rely on analytical models that are fitted by the spectral measurements. These approaches in practice suffer from uncertainties due to biological variability, tissue geometry, light scattering, systemic spectral bias, and variations in the experimental conditions. Here, we propose a new data-driven approach, termed deep spectral learning (DSL), to achieve oximetry that is highly robust to experimental variations and, more importantly, able to provide uncertainty quantification for each sO2 prediction. To demonstrate the robustness and generalizability of DSL, we analyse data from two visible light optical coherence tomography (vis-OCT) setups across two separate in vivo experiments on rat retinas. Predictions made by DSL are highly adaptive to experimental variabilities as well as the depth-dependent backscattering spectra. Two neural-network-based models are tested and compared with the traditional least-squares fitting (LSF) method. The DSL-predicted sO2 shows significantly lower mean-square errors than those of the LSF. For the first time, we have demonstrated en face maps of retinal oximetry along with a pixel-wise confidence assessment. Our DSL overcomes several limitations of traditional approaches and provides a more flexible, robust, and reliable deep learning approach for in vivo non-invasive label-free optical oximetry.

Monitoring retinal responses to acute intraocular pressure elevation in rats with visible light optical coherence tomography

Elevated intraocular pressure (IOP) is an important risk factor for glaucoma. However, the role of IOP in glaucoma progression, as well as retinal physiology in general, remains incompletely understood. We demonstrate the use of visible light optical coherence tomography to measure retinal responses to acute IOP elevation in Brown Norway rats. We monitored retinal responses in reflectivity, angiography, blood flow, oxygen saturation ( sO 2 ), and oxygen metabolism over a range of IOP from 10 to 100 mmHg. As IOP was elevated, nerve fiber layer reflectivity was found to decrease. Vascular perfusion in the three retinal capillary plexuses remained steady until IOP exceeded 70 mmHg and arterial flow was noted to reverse periodically at high IOPs. However, a significant drop in total retinal blood flow was observed first at 40 mmHg. As IOP increased, the venous sO 2 demonstrated a gradual decrease despite steady arterial sO 2 , which is consistent with increased arterial-venous oxygen extraction across the retinal capillary beds. Calculated total retinal oxygen metabolism was steady, reflecting balanced responses of blood flow and oxygen extraction, until IOP exceeded 40 mmHg, and fell to 0 at 70 and 80 mmHg. Above this, measurements were unattainable. All measurements reverted to baseline when the IOP was returned to 10 mmHg, indicating good recovery following acute pressure challenge. These results demonstrate the ability of this system to monitor retinal oxygen metabolism noninvasively and how it can help us understand retinal responses to elevated IOP.

Longitudinal deep-brain imaging in mouse using visible-light optical coherence tomography through chronic microprism cranial window

We longitudinally imaged both the superficial and deep cortical microvascular networks in brains of healthy mice and in a mouse model of stroke in vivo using visible-light optical coherence tomography (vis-OCT). We surgically implanted a microprism in mouse brains sealed by a chronic cranial window. The microprism enabled vis-OCT to image the entire depth of the mouse cortex. Following microprism implantation, we imaged the mice for 28 days and found that that it took around 15 days for both the superficial and deep cortical microvessels to recover from the implantation surgery. After the brains recovered, we introduced ischemic strokes by transient middle cerebral artery occlusion (tMCAO). We monitored the strokes for up to 60 days and observed different microvascular responses to tMCAO at different cortical depths in both the acute and chronic phases of the stroke. This work demonstrates that the combined microprism and cranial window is well-suited for longitudinal investigation of cortical microvascular disorders using vis-OCT.

Speckle reduction in visible-light optical coherence tomography using scan modulation

We present a technique to reduce speckle in visible-light optical coherence tomography (vis-OCT) that preserves fine structural details and is robust against sample motion. Specifically, we locally modulate B-scans orthogonally to their axis of acquisition. Such modulation enables acquisition of uncorrelated speckle patterns from similar anatomical locations, which can be averaged to reduce speckle. To verify the effectiveness of speckle reduction, we performed in-vivo retinal imaging using modulated raster and circular scans in both mice and humans. We compared speckle-reduced vis-OCT images with the images acquired with unmodulated B-scans from the same anatomical locations. We compared contrast-to-noise ratio (CNR) and equivalent number of looks (ENL) to quantify the image quality enhancement. Speckle-reduced images showed up to a 2.35-dB improvement in CNR and up to a 3.1-fold improvement in ENL with more discernable anatomical features using eight modulated A-line averages at a 25-kHz A-line rate.

Fast esophageal layer segmentation in OCT images of guinea pigs based on sparse Bayesian classification and graph search

Endoscopic optical coherence tomography (OCT) devices are capable of generating high-resolution images of esophageal structures at high speed. To make the obtained data easy to interpret and reveal the clinical significance, an automatic segmentation algorithm is needed. This work proposes a fast algorithm combining sparse Bayesian learning and graph search (termed as SBGS) to automatically identify six layer boundaries on esophageal OCT images. The SBGS first extracts features, including multi-scale gradients, averages and Gabor wavelet coefficients, to train the sparse Bayesian classifier, which is used to generate probability maps indicating boundary positions. Given these probability maps, the graph search method is employed to create the final continuous smooth boundaries. The segmentation performance of the proposed SBGS algorithm was verified by esophageal OCT images from healthy guinea pigs and the eosinophilic esophagitis (EoE) models. Experiments confirmed that the SBGS method is able to implement robust esophageal segmentation for all the tested cases. In addition, benefiting from the sparse model of SBGS, the segmentation efficiency is significantly improved compared to other widely used techniques.