New methods to measure and quantify tissue molecular composition and metabolism are a major driver of discovery in basic and clinical research. Optical methods are well suited for this task based on the non-invasive nature of many imaging and spectroscopy techniques, the variety of exogenous fluorescent probes available, and the ability to utilize label-free endogenous absorption signatures of tissue chromophores including oxy- and deoxy-hemoglobin, water, lipid, collagen, and glucose.
Despite significant advances in biomedical imaging, there remain challenges in probing tissue information in a fast, wide-field, and non-invasive manner. Moreover, quantitative in vivo mapping of endogenous biomarkers such as water and lipids remain relatively less explored by the biomedical optics community due to their characteristic extinction spectra, which have distinct spectral features in the shortwave infrared, a wavelength band that has been traditionally more challenging to measure.
The work presented in this dissertation was focused on developing instrumentation and algorithms for non-invasive quantification of tissue optical properties, fluorophore concentrations, and chromophore concentrations in a wide-field imaging format. All of the imaging methods and algorithms developed in this thesis extend the capability of the emerging technique called Spatial Frequency Domain Imaging (SFDI). First, a new imaging technique based on SFDI is presented that can quantify the quantum yield of exogenous fluorophores in tissue. This technique can potentially provide a new non-invasive means for in vivo mapping of local tissue environment such as temperature and pH. Next, an angle correction algorithm was developed for SFDI for more accurate estimation of tissue optical properties as well as chromophore concentrations in highly curved tissue, including small animal tumor models. Next, a wide-field label-free optical imaging system was developed to simultaneously measure water and lipids using the shortwave infrared (SWIR) wavelength region. Last, to break the bottleneck of processing speed in optical property inversion, new deep learning based models were developed to provide over 300× processing speed improvement.
Together, these projects substantially extend the available contrasts and throughput of SFDI, providing opportunities for new preclinical and clinical applications. / 2020-10-22T00:00:00Z
| Identifer | oai:union.ndltd.org:bu.edu/oai:open.bu.edu:2144/32674 |
| Date | 23 October 2018 |
| Creators | Zhao, Yanyu |
| Contributors | Roblyer, Darren |
| Source Sets | Boston University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
| Rights | Attribution-NonCommercial-NoDerivatives 4.0 International, http://creativecommons.org/licenses/by-nc-nd/4.0/ |
Page generated in 0.0024 seconds