Label-free vibrational stimulated Raman microscopy surpasses the diffraction limit by employing structured illumination, opening new avenues for super-resolution imaging of biological targets.
Label-free chemical imaging by stimulated Raman scattering (SRS) has seen rapid developments and widespread biomedical applications. Nevertheless, the cross-section of the Raman process sets up a physical limit that impedes further evolution of SRS. Computational imaging is an appealing approach to break the physical barrier through the synergistic integration of advanced instrumentation and data science. In this talk, I will introduce a series of computational SRS work to push the physical limits in the aspects of imaging speed, signal fidelity, chemical specificity, and spatial resolution.
Breast cancer is one of the most common cancers among women worldwide. The process of diagnosis is completed with conventional histology (Hematoxylin and eosin staining followed by immunohistochemistry staining), which is known to be labor-intensive, time-consuming, and subject to inter- as well as intra-laboratory variations. We present here a multi-color Stimulated Raman Histology (SRH) technique to overcome these limitations. The multi-color SRH can map lipid, cellular protein, and collagen in a label-free manner to differentiate cancerous breast tissue from healthy tissue. Our technique can be a promising candidate for labor-efficient, fast, and highly sensitive breast cancer histology.
The gut microbiome has a considerable impact on human health. Furthermore, it is very sensitive to the in vivo environment changes. Entacapone (ENT) is one important human-targeted drug for the treatment of Parkinson’s disease shown to largely impact the growth of individual organisms in the gut microbiome. To resolve ENT incorporation and impact on gut microbiome activity in situ, we employed stimulated Raman scattering - fluorescence in situ hybridization (SRS-FISH) along with ultrafast delay line tuning. The hyperspectral chemical images acquired with this method could better preserve and separate the chemical information, thus enabled high-throughput identity-specific single cell metabolism analysis and mapping of drug incorporation by individual microbiome cells.
Engineered microbes can produce valuable chemicals but often require extensive optimization to reach sufficient production yields. Current quantitation methods such as gas chromatography-mass spectrometry are based on ensemble measurements that ignore the genetic or phenotypic variation among cells. Here, we use hyperspectral stimulated Raman scattering (hsSRS) to directly image free fatty acid synthesis in metabolically engineered Escherichia coli and uncover substantial heterogeneity among and within colonies. Using optimized laser exposures, we perform longitudinal SRS imaging to link the dynamics of biofuel synthesis with cell growth. Further, we develop an analytical protocol to enable mass spectrometry-like chemical specificity estimates of fatty acids.
Hyperspectral stimulated Raman scattering (hSRS) is a label-free microspectroscopic modality that enables live-cell metabolic imaging with chemical specificity. Yet, hSRS in the CH region has low throughput and poor chemical specificity, which limits its application to a broader range of metabolic studies. We propose a high-content, high-throughput hSRS imaging method by a sparsity-driven spectral unmixing and active spectral sub-sampling. We unprecedently generate chemical maps of four major metabolic species (lipid, protein, nucleic acid and carbohydrate) in a Mia PaCa-2 cell using seven spectral frames in the CH region, improving the acquisition speed by over an order of magnitude.
Fingerprint stimulated Raman scattering (SRS) produces label-free chemical maps of molecules in living systems with higher specificity compared to CH vibration region. However, due to the weak signal levels in the fingerprint window, it remains challenging for fingerprint SRS to study highly dynamic or large-scale samples. Here, we push the design space of SRS using deep learning, which can recover the signal-to-noise ratio to the levels comparable to measurements with 100 times longer integration time. Combined with an ultrafast 50-kHz delay-line tuner, we can generate real-time images of cholesterol, fatty acid, and proteins of living cells and large-area tissues including the whole brain.
Type 2 diabetes is an increasingly prevalent disease, with more than 400 million people worldwide diagnosed in 2016. As a stable and accurate biomarker, glycated hemoglobin (HbA1c) is clinically used to diagnose type 2 diabetes with a threshold of 6.5% HbA1c among total hemoglobin (Hb). Current methods such as boronate affinity chromatography or enzymatic assay involve complex processing of large-volume blood samples, which inhibits real-time measurement in clinic. Moreover, these methods cannot measure the HbA1c fraction at single red blood cell level, thus unable to separate the contribution by diabetes from other factors such as diseases related to lifetime of red blood cells. Here, we demonstrate a transient absorption imaging approach that is able to differentiate HbA1c from Hb based on the excited state dynamics measurement. HbA1c fraction inside a single red blood cell is derived quantitatively through phasor analysis. HbA1c fraction distribution for diabetic blood is found apparently different from that for healthy blood. A mathematical model is developed to derive the long-term glucose concentration in the blood. Our technology provides a new way to study heme modification and to derive clinically important information avoid of glucose fluctuation in the bloodstream.
Spectroscopic stimulated Raman scattering (SRS) is a label-free chemical imaging modality enabling visualization of molecules in living systems with high specificity. Among various spectroscopic SRS imaging methods, a convenient way is through linearly chirping two femtosecond lasers and tuning their temporal delay, which in turn corresponds to different Raman shifts. Currently, the acquisition speed using a resonant mirror is 3 seconds (80 microseconds per spectrum), which is insufficient for imaging samples with high motility. In this work, we aim to push the imaging speed using a 50-kHz polygon scanner as a delay line tuner, achieving a speed of 20 microseconds per spectrum. At such high speeds, to overcome the signal level decrease due to reduced signal integration time, we apply a U-Net deep learning framework, which first takes pairs of spectroscopic SRS images at different speeds as training samples, with high-speed, low-signal images as input and low speed, high-signal ones as output. After training, the network is capable of rapidly transforming a low-signal spectroscopic image to a high-signal version. Consequently, our design can generate ultrafast spectroscopic SRS image while maintaining the signal level comparable to the output with longer signal integration time.
A hyperspectral image corresponds to a data cube with two spatial dimensions and one spectral dimension. Through linear un-mixing, hyperspectral images can be decomposed into spectral signatures of pure components as well as their concentration maps. Due to this distinct advantage on component identification, hyperspectral imaging becomes a rapidly emerging platform for engineering better medicine and expediting scientific discovery. Among various hyperspectral imaging techniques, hyperspectral stimulated Raman scattering (HSRS) microscopy acquires data in a pixel-by-pixel scanning manner. Nevertheless, current image acquisition speed for HSRS is insufficient to capture the dynamics of freely moving subjects. Instead of reducing the pixel dwell time to achieve speed-up, which would inevitably decrease signal-to-noise ratio (SNR), we propose to reduce the total number of sampled pixels. Location of sampled pixels are carefully engineered with triangular wave Lissajous trajectory. Followed by a model-based image in-painting algorithm, the complete data is recovered for linear unmixing. Simulation results show that by careful selection of trajectory, a fill rate as low as 10% is sufficient to generate accurate linear unmixing results. The proposed framework applies to any hyperspectral beam-scanning imaging platform which demands high acquisition speed.
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.