We propose a new toolset to characterize the role of corneal nerves in ocular conditions such as dry eye disease (DED). By automating analysis (nerve segmentation, density, detection of whorl patterns and nerve beading) we quantify the nerve structure over whole ex vivo corneas in a mouse model of evaporative DED. We also created a mouse line of DED with genetically encoded calcium markers and successfully imaged neural activity over time in live corneas. This combined structural and functional quantitative approach will further our understanding of corneal nerves in disease and accelerate the development of novel therapies.
Advances in three-dimensional (3D) microscopy are providing never-before-seen images of coronary microvasculature organization. However, it remains inaccessible to researchers due to difficult sample preparation and image analysis. We present a deep learning network that can segment the coronary microvasculature in 3D microscopy without vessel staining. The network is based on 3D U-net and accepts DAPI (nuclei) and autofluorescence (tissue structure) volumes as inputs. The network detects vessels with high accuracy when compared to the ground truth obtained from isolectin staining. Contrast-free segmentation of vessels simplifies sample preparation, frees fluorescent channels during imaging and opens the door toward user-friendly 3D microscopy.
Advances in tissue clearing and three-dimensional microscopy require new tools to analyze the resulting large volumes with single-cell resolution. Many existing nuclei detection approaches fail when applied to the developing heart, with its high cell density, and elongated myocytes. We propose a new regression-based convolutional neural network that detect nuclei centroids in whole DAPI-stained embryonic quail hearts. High nuclei detection accuracy was obtained in two different hearts where our algorithm outperformed other deep learning approaches. Once nuclei were identified we were also able to extract properties such as orientation and size, which enables future studies of heart development and disease.
Optical coherence tomography (OCT) has been applied to investigate heart development because of its
capability to image both structure and function of tiny beating embryonic hearts. Labeling heart
structures is necessary for quantifying mechanical functions such as cardiac motion, wall strain, blood
flow and shear stress, of looping hearts. Since manual segmentation is time-consuming and labor-
intensive, this study aimed to use deep learning to automatically extract dynamic shapes including the
myocardium, the endocardial cushions, and the lumen of beating embryonic hearts from 4-D OCT
images. This will benefit research on heart development, especially studies requiring large cohorts of
embryos.
Optical coherence tomography (OCT) has been applied for understanding heart development because of its capability of imaging both the structure and function of tiny beating embryonic hearts. Labeling endocardial cushions is necessary for quantifying morphological characteristics of the looping hearts. Since manual segmentation is time-consuming and prone to subjectivity, this study aims to use V-net to automatically segment endocardial cushions from OCT images. This will benefit research on heart development, especially studies requiring large cohorts of embryos, for example those investigating the teratogenic effects of ethanol or drugs and the prevention of these effects on the developing hearts.
New techniques are needed to study the coronary microvasculature in the embryo and to quantify the differences between healthy and diseased coronary development. We combine our custom optical clearing method (LIMPID), fluorescent staining with DiI and DAPI, and 3D confocal microscopy to visualize the coronary vasculature at two developmental stages (day 9, 13) in quails. We discovered a highly organized coronary vessel network that is aligned with surrounding myocardium cells even at early developmental stages. By characterizing the normal heart vasculature, this experiment provides a baseline for future studies on how diseases affect coronary and myocardial orientation in the embryo.
Optical coherence tomography (OCT) is a standard for retinal imaging and has been integrated to surgical microscopes to evaluate tissue-instrument interactions during macular surgery. One common procedure during such surgery, membrane peeling, is done under a white light microscope. Indocyanine green (ICG) can be used to specifically dye the inner limiting membrane (ILM) and facilitates this surgery. However, there is no equivalent contrast mechanism to specifically target the ILM on OCT images. We propose to use photothermal OCT (PT-OCT) to detect ICG in the OCT field-of-view, which would increase contrast between the ILM and other structures of the retina. As preliminary data for this project, we have collected PT-OCT images of different ICG phantoms over a wide range of laser powers and ICG concentrations, including concentrations lower than the clinical standard. We have also detected a PT-OCT signal from ICG on a mouse tail with low photothermal laser powers (0.56 mW) to evaluate the feasibility of this technique for in vivo ocular imaging. Finally, we have collected PT-OCT images of a fixed monkey retina after the ILM was dyed with ICG, and obtained a PT-OCT signal from the ICG and the melanin present in the retinal pigment epithelium and the choroid. Those preliminary results indicate that ICG can be detected with PT-OCT at low concentrations and low laser powers. PT-OCT has never been demonstrated in the human eye and has only been recently demonstrated in the mouse eye. This experiment establishes feasibility for PT-OCT in clinical applications.
KEYWORDS: Nanorods, 3D acquisition, 3D image processing, Gold, Optical coherence tomography, Retina, In vivo imaging, Eye, 3D imaging standards, Retinal scanning
Optical coherence tomography (OCT) has become standard in retinal imaging at the pre-clinical and clinical level by allowing non-invasive, three-dimensional imaging of the tissue structure. However, OCT lacks specificity to contrast agents that could be used for in vivo molecular imaging. We have performed in vivo photothermal optical coherence tomography (PT-OCT) of targeted gold nanorods in the mouse retina after the mice were injected systemically with the contrast agent. To our knowledge, we are the first to perform PT-OCT in the eye and image targeted gold nanorods with this technology. As a model of age-related wet macular degeneration, lesions were induced by laser photocoagulation in each mouse retina (n=12 eyes). Untargeted and targeted (anti-mouse CD102 antibody, labeling neovasculature) gold nanorods (peak absorption λ=750nm) were injected intravenously by tail-vein injection five days after lesion induction, and imaged the same day with PT-OCT. Our instrument is a spectral domain OCT system (λ=860nm) with a Titanium:Sapphire laser (λ=750nm) added to the beam path using a 50:50 coupler to heat the gold nanorods. We acquired PT-OCT volumes of one lesion per mouse eye. There was a significant increase in photothermal intensity per unit area of the lesion in the targeted gold nanorods group versus the saline control group and the untargeted gold nanorods group. This experiment demonstrates the feasibility of PT-OCT to image the distribution of molecular contrast agents in the mouse retina, including in highly scattering lesions. In the future we will use this method to identify new biomarkers linked with retinal disease.
OCT has become a standard in retina imaging at the pre-clinical and clinical level by allowing non-invasive, three-dimensional imaging of the tissue structure. However, OCT lacks specificity to contrast agents that could be used for in vivo molecular imaging. We have performed in vivo photothermal optical coherence tomography (PTOCT) of gold nanorods in the mouse retina after the mice were injected intravenously with the contrast agent. To our knowledge, we are the first team to perform PTOCT in the eye. Four lesions were induced by laser photocoagulation in each mouse retina (n=6 mice) and gold nanorods (untargeted and targeted with anti-mouse CD102 antibody, which labels neovasculature, peak absorption λ=750nm) were injected intravenously by tail-vein injection five days later in four mice (two mice are controls). The mice were imaged with PTOCT the same day. Our instrument is a spectral domain OCT system (λ=860nm) with a Titanium:Sapphire laser (λ=750nm) added to the beam path using a 50:50 splitter to target the gold nanorods. We acquired PTOCT B-scans over one lesion per mouse eye. There was a significant increase in photothermal intensity at the center of the lesion in the gold nanorod group versus the control group. This experiment demonstrates the feasibility of PTOCT to image the distribution of contrast agents in the mouse retina. In the future we will use this method to optimize drug delivery to the retina in pre-clinical models.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.