This PDF file contains the front matter associated with SPIE Proceedings Volume 11655, including the Title Page, Copyright information, and Table of Contents.

Introduction to SPIE Photonics West BiOS conference 11655: Label-free Biomedical Imaging and Sensing (LBIS) 2021

Fluorescence lifetime imaging microscopy (FLIM) of reduced nicotinamide adenine dinucleotide and reduced nicotinamide adenine dinucleotide phosphate, together termed NAD(P)H, is often used to analyze and characterize the metabolic state of biological samples. The use of photon counting in FLIM of NAD(P)H often leads to long accumulation times which average out sub-second dynamics. This challenge is overcome with the presented two-photon fast FLIM system that uses an analog detection system and fast digitizer, enabling imaging up to 20 frames per second. Furthermore, GPU processing is used for real-time phasor analysis and visualization of collected data.

Fluorescence imaging has been used for decades to identify important elements in biological samples. However, the additional overhead of fixing and staining the sample limits the pervasiveness of the technique. Here we show a robust virtual fluorescence technique which uses deep learning to jointly optimize microscope design (i.e. illumination pattern) and fluorescence image reconstruction. Our results show how by combining the image capture and processing stages we can accurately and adaptively predict fluorescent images from unlabelled biological samples.

We propose a Single-Pixel Macroscopic Autofluorescence Imaging platform with supercontinuum excitation (440-690nm) and 16 parallel wavelength detection (475-1000nm) through a spectrophotometer coupled Single Photon Counting PMT. Recorded decays of FAD, POPOP and PPIX commercial auto-fluorophores serve to simulate training samples for UNMIX-ME, a deep learning algorithm that disentangles spectral overlaps. In silico mixed samples are reconstructed as a proof of concept and mixed in vitro samples prepared, measured and reconstructed to unmixed intensity and lifetime images. The results highlight the utility of the platform to macroscopically quantify autofluorescence lifetime in vitro and its future potential for in vivo autofluorescence imaging.

Type 1 diabetes occurs when insulin secreting beta cells in pancreatic islets are destroyed leading to elevated glucose and ill health. Islet transplantation is an effective therapy, but islets are often damaged by the isolation process resulting in numerous, repeated transplants to achieve insulin independence. We have applied hyperspectral microscopy to damaged islets and have shown through the assessment of native cell autofluorescence we can detect heterogenous forms of damage (elevated ROS, inflammatory signalling and warm ischemia) in mouse islets. This approach has great potential to be translated clinically to minimise the burden of suboptimal islet transplantations.

The reduced pyridine-nucleotides nicotinamide adenine dinucleotide (NADH) and nicotinamide adenine dinucleotide phosphate (NADPH) are ubiquitous metabolic cofactors playing significant, distinct roles in cellular metabolism. NADH and NADPH are primarily involved in cellular respiration and in maintaining antioxidant defenses, respectively, however their nearly identical fluorescence properties (the abbreviation NAD(P)H denotes this uncertainty) pose a challenge when interpreting and distinguishing autofluorescence signals. For sensing in turbid media such as tissue, additional challenges include the presence of multiple scattering, intrinsic absorption, and background fluorescence. Here, we assess an approach for distinguishing cellular-respiration and oxidative-stress responses when sensed in turbid media. Spectral phasor analysis, an analytical approach originally developed for the rapid segmentation of hyperspectral images, has been used on UV-excited autofluorescence for the real time monitoring of cellular NAD(P)H conformation. We showed previously that the spectral response to chemicals affecting NADH and NADPH pathways, e.g., in response to cyanide and hydrogen peroxide, does not follow two-component behavior and so could be distinguished in cell-only preparations. Here, we demonstrate pathway-level sensing in turbid media by monitoring the metabolic response of yeast cells embedded in a source for background emission. The distinguishability of UV-excited autofluorescence spectra to chemical perturbations affecting cellular respiration and oxidative stress are compared with previously reported cell-only observations.

Antitumor activity of T cells often determines treatment efficacy for immunotherapy and radiotherapy. Due to the immunosuppressive tumor microenvironment, tumor-infiltrating T cells often exhibit a hypofunctional exhausted phenotype. Current methods to evaluate T cell exhaustion use flow cytometry or cytokine production measurements that are either destructive to the sample or cannot capture single-cell heterogeneity, respectively. Here, we used fluorescence lifetime imaging of the metabolic co-enzyme NAD(P)H to evaluate T cell metabolism during exhaustion. Exhausted T cells have significantly different NAD(P)H lifetimes compared to functional T cells. This study demonstrates a label-free method to monitor T cell exhaustion in tumors.

Magnetic levitation is an effective tool for separating target cells within a heterogeneous solution by utilizing density differences among cell lines. However, magnetic levitation cannot be used to identify target cells which have similar density profile as the other cells in the solution. Therefore, cell identification accuracy can dramatically reduce. In this study, we introduce, for the first time, the use of deep learning-based object detection approach for label-free identification of rare cancer cells within levitated cells. As a result, our novel and hybrid detection strategy could be used to identify circulating tumor cells for diagnosis and prognosis of cancer.

We virtually generate multiple histological stains through a single deep-neural-network, using at its input autofluorescence images of the unlabeled tissue alongside a user-defined digital-staining-matrix. By feeding this digital-staining-matrix to the network, the user indicates which stain to apply on each pixel or region-of-interest, enabling virtual blending of multiple stains according to a desired micro-structure map. We demonstrated this technique by applying combinations of different stains (H&E, Masson’s Trichrome and Jones silver stain) on blindly-tested, unlabeled tissue sections. This technology avoids the histochemical staining process and enables newly-generated stains and stain-combinations to be used for inspection of label-free tissue microstructure.

We present a field-portable and high-throughput imaging flow-cytometer, which performs phenotypic analysis of microalgae using image processing and deep learning. This computational cytometer weighs ~1.6kg, and captures holographic images of water samples containing microalgae, flowing in a microfluidic channel at a rate of 100mL/h. Automated analysis is performed by extracting the spatial and spectral features of the reconstructed images to automatically identify/count the target algae within the sample, using image processing and convolutional neural networks. Changes within the measured features and the composition of the microalgae can be rapidly analyzed to reveal even minute deviations from the normal state of the population.

Deep ultraviolet microscopy (UV) enables high-resolution, label-free imaging of biological samples and yields diagnostically relevant quantitative molecular and structural information. We recently demonstrated that deep UV microscopy can serve as a simple, fast, and low-cost alternative to modern hematology analyzers that assess variations in the morphological, molecular, and cytogenetic properties of blood cells to monitor and diagnose blood disorders. We also introduced a pseudocolorization scheme that uses multi-spectral UV images (acquired at three different wavelengths) to generate images whose colors accurately recapitulate those produced by conventional Giemsa staining, and can thus be used for visual hematological analysis. Here, we present a deeplearning framework to virtually stain single-channel UV images acquired at 260 nm, providing a factor of three improvement in imaging speed without sacrificing accuracy. We train a generative adversarial network (GAN) using image pairs consisting of single-channel UV images of blood smears and their corresponding pseudocolorized images to generate realistic, virtually stained images. The virtual stained images are post-processed to improve contrast and yield consistent background colors. We quantify the performance of our framework in terms of the structural similarity index (SSIM) for each color channel. Our virtual staining scheme is the first step towards a completely automated hematological analysis pipeline that includes segmentation and classification of different blood cell types to compute metrics of diagnostic value. Our method eliminates the need to acquire images at different wavelengths and could potentially lead to the development of a faster and more compact label-free, point-of-care hematology analyzer.

Ovarian cancer is challenging due to poor detection rates and high mortality. Multispectral fluorescence imaging (MFI) has recently become a favorable method for cancer characterization. By utilizing MFI along with a characterized ovarian cancer mouse model and human fallopian tube histology sections, we were able to study cancer in its earliest stages with a promising modality for early disease detection. Fluorescence images with various emission combinations from 280nm to 550nm excitation and reflectance images from 320nm to 550nm were taken of 8-week-old mouse ovarian tissues. Human fallopian tube histology slides were also imaged with the same fluorescence images. Disease characterization was studied using Quadratic Discriminant Analysis (QDA) on image grayscale intensities for both tissues as well as the greylevel co-occurrence matrix (GLCM) in human tissue slides in order to determine a classification group with the highest predictive merit. Both tissues were able to be classified with greater than 80% accuracy, suggesting promise for MFI as a potential diagnostic candidate.

Here we explore the use of Photoacoustic Remote Sensing (PARS™) microscopy, a recently developed non-contact photoacoustic imaging modality, for visualizing subcellular structures label-free in tissues. Operating in an all-optical reflection-mode architecture PARS captures optical absorption contrast within bulk tissue samples. Presented here, by visualizing endogenous optical absorption of DNA and cytochromes, cellular morphology is captured with contrast analogous to the industry standard hematoxylin and eosin (H&E) staining. Subcellular features are recovered from human and murine, brain and gastrointestinal tissues. This work represents a significant step towards the development of a real-time microscopy system for label-free histopathological assessment of tissues in-situ.

Vitiligo is a disorder characterized by the loss of skin pigment due to the autoimmune destruction of epidermal melanocytes. Treatments that aim to repigment vitiligo regions are long, incomplete, and suffer from a highly variable success rate. This study aims to use non-invasive, depth-resolved, label-free, two-photon detection of reduced nicotinamide adenine dinucleotide (NADH) to quantitatively evaluate mitochondrial dynamics in vitiligo micrografting patients during the repigmentation process. This study indicates differences in the depth-dependent mitochondrial profiles between healthy and vitiligo tissues. This study also indicates that vitiligo regions responsive to treatment exhibit depth-dependent shifts in mitochondrial dynamics prior to repigmentation.

Prostate cancer is the most common non-cutaneous cancer in men. Moreover, identifying the most effective treatment strategy for an individual with prostate cancer strongly depends on the Gleason score (Grade group) of the disease. However, this task continues to be a significant clinical challenge in a subset of patients. Currently, the gold standard for grading prostate cancer is the Pathologist’s visual assessment of hematoxylin-and-eosin-stained histological sections, and designation of a Gleason score (Grade group) based on the top two most common Gleason grades. However, this process is subjective and thus prone to error and high variability, especially amongst non-Urologic Pathologists or those who are not aware of recent modifications to the grading system. In addition, variations in protocols for staining, could make quantitative analysis of stained tissues challenging in some cases. Therefore, there is a significant clinical need to develop additional complementary quantitative methods that can provide robust, objective, reproducible, and accurate information of the aggressiveness and grade of prostate cancer. Here we address this issue by imaging unstained tissue sections using multi-spectral deep-UV microscopy. This method enables us to obtain valuable quantitative insight into the aggressiveness and grade of the disease due to its subcellular spatial resolution and high sensitivity to many endogenous biomolecules, including nucleic acid and proteins. The approach uses a simple, fast and cost effective wide-field imaging configuration that is well-suited for pathology applications. Spectral signatures form wavelengths ranging from 220 nm to 450 nm are analyzed using a geometrical representation of principal component analysis. Our results reveal distinct morphological and molecular alterations in tissue as they progress from benign to cancerous, and as they become more aggressive (higher grade). In this research project, we delineated the design of the multispectral, deep UV microscope and described our quantitative and qualitative image analysis. Our results show that multispectral deep UV microscopy provides a quantitative measure to differentiate between prostate cancer patients with varying grades of cancer.

Label-free intensity and lifetime two photon excited fluorescence in combination with second harmonic generation images yield a wealth of morphological and functional information. Quantitative assessments of such images enable characterization of cellular metabolism and morphology in the context of important structural features of the surrounding matrix and/or biomaterial scaffold. Dynamic changes and interactions that are difficult to characterize using traditional staining or bulk tissue assessment techniques are revealed in the context of disease presence, tissue responses, and cell-matrix interactions in human and engineered tissues.

For investigating the potential for personalized treatment of pancreatic cancer, two- and three-photon autofluorescence and second and third harmonic generation images were simultaneously generated from in vivo pancreatic tumor patient-derived xenograft (PDX) mice. Two different treatment regimens were administered to PDX mice for 5 weeks and compared. Control mice were treated only with saline. Mice were classified into three groups according to their response to the treatment: responsive, resistant, and control. Optical redox ratios were calculated from tumor regions which showed differences between the responsive group compared to the resistant and control groups.

We have developed a novel multiphoton nonlinear microscopy with a highly integrated optical imaging system that offers numerous label-free techniques including two-photon excited fluorescence, second-harmonic generation, third-harmonic generation, fluorescence lifetime imaging, and spectral focusing coherent anti-Stokes Raman scattering in one platform. We have applied our system to investigate plant-microbe-mineral interactions in the rhizosphere. The system provides time efficient monitoring of the rhizosphere, offering an array of simultaneous biomolecular information without staining, three-dimensional sub-micron resolution with deeper penetration , and less photodamage. We believe that multiphoton nonlinear optical microscopy will become a valuable imaging tool in the rhizosphere and soil mineral sciences.

A single photon-excited confocal whole-cell autofluorescence lifetime system is described that uses fast field-programmable gate array (FPGA)-based time tagging electronics to implement time-correlated single photon counting (TCSPC) with simultaneous near-IR brightfield imaging. This system resolves variations in the fluorescence decay of the metabolic coenzyme NAD(P)H that provides high accuracy and precision for classifying activated and quiescent primary human T cells (ROC AUC = 0.98). This performance is similar to that obtained using commercial two-photon fluorescence lifetime imaging microscopy (FLIM). The small footprint, low cost, and high acquisition speed make this system attractive for monitoring cell biomanufacturing.

We present a new non-invasive label-free imaging method, which visualizes tissue dynamics/sub-cellular motion by analyzing the temporal fluctuation of optical coherence tomography (OCT) signals. Our modality has been utilized for visualization and quantification of the time course necrotic process of human breast adenocarcinoma spheroid (MCF7). The response of MCF7 spheroid against anti-cancer drug has also been investigated. The presented method is quantitative. So, the necrotic process was not only shown by images but the dynamics signal value is also plotted as a function of time. It showed clear degradation of tissue activity by time.

In this paper, we demonstrate polarization-sensitive optical coherence tomography (PS-OCT) based tissue dynamics imaging methods to investigate the activity of mouse livers. The methods include logarithmic intensity variance (LIV), temporal variance of birefringence, temporal degree-of-polarization-uniformity. Portal veins appeared as hypo-scattering structure in static OCT intensity. High dynamics appearance around the portal veins in all the dynamics contrasts was found just after the dissection of mouse liver. Although, high activity around the vessels reduced by time, a bright and high dynamics spot appeared in the vessel in LIV cross-sectional and en face image after 28-hours of the dissection.

Most soft tissues, such as cornea and sclera, exhibit mechanical anisotropy. Obtaining the anisotropy properties of the sclera can be helpful in understanding the mechanical properties of the eyes in the development and progression of glaucoma and myopia. In this work, we have demonstrated a method of quantifying the anisotropic Young’s modulus values for sclera, using a confocal Lamb wave based acoustic radiation force optical coherence elastography (ARF-OCE). We have measured mechanical anisotropy in porcine sclera by measuring the Young’s modulus along both the equator direction and anterior to posterior direction. The results demonstrate the Young’s modus in the equator direction is 15-30% higher than that in the anterior to posterior direction.

Photothermal optical coherence tomography (PT-OCT) is an extension of OCT that forms images based on both scattering and absorption of light. Conventional OCT measures the path length of elastically back-scattered light. Variation of the tissue’s local refractive index due to photothermal modulation results in a modulated OCT phase signal that relates to the absorption of light. Detailed understanding of the PT-OCT signal has the potential to provide insight into the chemical composition of tissue, and may pave a way to detecting and characterizing lipid-rich atherosclerotic plaques with intracoronary imaging in patients. Here, we investigated the dependence of the PT-OCT signal on concentration of a specific tissue component. We modeled the generated thermal wave field and the resulting PT-OCT signal as a function of concentration of this component using the bio-heat equation. We found that the significant parameters in determining the amplitude of the PT-OCT signals are the density, the absorption coefficient and the specific heat of the sample. All these parameters vary as a function of the sample composition, leading to a non-linear relation between PT-OCT signal and the concentration of the component of interest. Only in special cases, e.g. the oxygenation level of blood, when the absorption coefficient is the only varying parameter does a linear dependence arise. PT-OCT experiments on tissue-like samples, prepared by mixing mayonnaise (<80% lipid) and agar gel (<90% water) to mimic lipid-rich atherosclerotic plaques, confirmed the non-linear relation predicted by our model.

Quantitative phase digital holographic microscopy (QP-DHM) we have developed presents the great advantage, thanks in particular to its numerical capacity to apply both wave front correction and autofocusing, to accurately measure optical path difference (OPD) in screening mode involving multiwell cell culture plates. Experiments combining electrophysiology and QP-DHM have permitted to establish relationships between ions movements (transmembrane current) and the OPD signal. The capacity of QP-DHM to perform label-free image-based screening was validated by screening for the GABAA receptor activity expressed in HEK cells using a selection of active compounds in agonist, antagonist, and modulator modes.

The first-line treatment for brain cancer is surgery, which focuses on maximizing the percentage of the tumor removed during surgery (i.e., extent of resection) while minimizing damage to healthy brain tissue. Data show that extent of resection is one of the most critical factors associated with prolonged survival. However, differentiating between tumor and healthy tissue intraoperatively remains a significant clinical challenge, resulting in an exceedingly low 5-year survival rate of only ~35%. In this work, we show that quantitative oblique back illumination microscopy (qOBM), a novel label-free optical imaging technique that achieves tomographic quantitative phase imaging (QPI) in thick scattering samples, clearly differentiates between tumor and healthy tissue. Using a 9L gliosarcoma rat tumor model, we show that quantitative image features from qOBM provide a robust set of biomarkers for disease. In addition, tumor regions, including diffuse tumor, and healthy brain structures, show excellent structural agreement with H&E stained and sliced brightfield images, the gold standard for cancer detection. The unique attribute of qOBM—low-cost, easy-to-use, label-free, and real-time—make this technology ideally suited to help guide neurosurgery and address this important unmet need. Here we describe our free-space qOBM system and present quantitative results from the 9L gliosarcoma rat tumor model.

Many studies suggest that external forces applied to cells generate signals that are as potent as those of biochemical stimuli. To understand how these forces are transmitted to the molecular structures of cells, and how they might be transduced into biochemical reactions, require measuring both cell mechanical properties (CMPs) and biological pathways in a physiological environment. For this purpose, we will present measurements, through rheological approaches, of CMPs in a label-free manner performed thanks to an automated imaging system devoted to live cells combining digital holographic microscopy, environmental control, and microfluidic assays.

We propose a multimodal quantitative imaging technique by integrating off-axis interferometric phase microscopy (IPM) and fluorescence microscopy, for measuring morphological changes of cells induced by hyper-osmotic pressure. Based on the collected data, we calculated the temporal dependence of various physiological parameters during osmotic pressure induced by the cell surrounding-medium change. This technique is expected to be a useful for observing the dynamics of live cells, while obtaining both quantitative imaging capabilities and molecular specificity of specific cellular organelles.

We propose a label-free technique that performs analysis of live individual cells during flow and then sort the cells based on the analysis results. The cells rapidly flow in a droplet microfluidics device, which prevents two cells in the field of view in once. By acquiring holograms of each cell during flow and achieving the cell optical path delay (OPD) profile, we extract different features that differentiate between the cells, based on which rapid sorting is performed. The device presents the possibility for both high-throughput label-free analysis of individual cells and their real-time sorting.

Tomographic phase microscopy (TPM) is a well-known technique for imaging cells without staining and visualizing their inherent 3-D refractive index profile. We propose a new low-cost TPM approach using sample rotation of single cells. The cells are inserted into a capillary and trapped in a low-density gel. Cell rotation is performed by mechanically rotating the capillary. Projection acquisition is performed using an imaging interferometer. This technique is expected to be useful for label-free cell sorting and analysis.

We report a module employing shearing interferometry with constant off-axis angle (SICA) that works both for coherent light sources and low-coherence light sources, in which the shearing distance in uncoupled to the off-axis angle. In the module, two laterally sheared sample beams are generated by a grating, and the part of the sample area containing no sample details serves as the reference beam, as in any shearing interferometer. However, in our module, the shearing distance can be flexibly adjusted by controlling the axial location of the grating, while the off-axis interference angle remains constant.

Six-pack holography may be used to reject out-of-focus objects in dynamic samples. By illuminating the sample from six different angles in parallel, out-of-focus objects are laterally shifted in six different directions when projected onto the focal plane. Pixel-wise averaging of the six reconstructed images then reduces the impact of out-of-focus objects. The averaged images reduce the contribution of out-of-focus objects by up to 83%, depending on the out-of-focus object axial distance from the focused plane. Simulation results are shown, as well as experimental results for swimming microalgae and flowing microbeads, including numerical refocusing by Fresnel propagation.

Recent advances in ultrafast laser science have opened new prospects in probing living systems. Suddenly-excited molecular vibrations emit picosecond-duration coherent infrared light characteristic of the sample’s composition. Given the capacity of this electric-field molecular fingerprint (EMF) to cross-molecularly probe organic samples, we evaluate the feasibility of EMF for biomedical diagnostics. We report high stability of EMF over time and show that several cancer entities leave their signatures in the EMF of serum. Leveraging on the advances of stability, reproducibility and sensitivity of EMF, we provide foundation for applications of infrared molecular profiles in the preventive care model of health monitoring.

We report a label-free Surface Enhanced Raman Spectroscopy (SERS) for pleural fluid analysis to distinguish Lung cancer from controls patients. Herein, we have used a novel silver coated silicon Nanopillar (SCSNP) as SERS substrate to acquire multiple SERS spectra for each pleural fluid sample and advanced chemometrics methods. We report a classification accuracy of 85% along with sensitivity and specificity of 87% and 83% respectively for the detection of Lung cancer over control pleural fluid samples with a receiver operating characteristics (ROC) area under curve value of 0.92 using PLS-DA binary classifier to distinguish between lung cancer over control subjects.

We report a novel approach to coherent anti-Stokes spectroscopy based on multiple overtone resonances in both theory and experiment. A pump laser with the tunable wavelength at the near infrared range is used to excite the inner-band vibrational transition of the C-H stretching vibrations, commonly corresponding to biological molecules. Strong coherent signals with higher chemical specificity are observed in experiments. Our work points to a new route for applying resonantly enhanced nonlinear optical spectroscopy in the deeper tissue optical imaging with near infrared lasers, which is important for various biomedical applications.

Spontaneous Raman spectroscopy enables non-ionising, non-destructive, and label-free acquisition of a biochemical fingerprint for a given sample. However, the long integration times required largely prohibit high-throughput applications. Here, we present a comprehensive deep learning framework for extreme speed-up of spontaneous Raman imaging. Our deep learning framework enhances Raman imaging two-fold, effectively reconstructing both spectral and spatial information from low spatial resolution, low signal-to-noise ratio images to achieve extreme Raman imaging time speed-ups of 40-90x while mainting high reconstruction fidelity. As such, our framework could enable a host of higher-throughput spontaneous Raman spectroscopy applications across a diverse range of fields.

We present multicolor coherent Raman imaging (CRI) with two orders of magnitude enhanced sensitivity accomplished by a frequency modulation (FM) at 20 MHz of a portable fiber optical parametric oscillator (FOPO). The FM is combined with a rapid and wide tunability for accessing Raman bands between 700 and 3500 1/cm within only 5 ms. This development constitutes an important step for advancing CRI microscopy in terms of portability and sensitivity for applications in medical diagnostics or environmental sensing.

Fiber-based laser sources for Coherent Anti-Stokes Raman scattering (CARS) microscopy and Stimulated Raman Spectroscopy (SRS) have been successfully used for label-free bio-imaging in the last few years. However, these systems were limited so far to a few experimental setups in university labs. We present the first extremely compact commercially available dual-wavelength source for CARS and SRS. With a foot-print of <30cm x 50cm this alignment-free turn-key system offers tuning over the highly relevant frequency range from <500cm^-1 up to 5000cm^-1 within a few seconds without any detrimental thermal effects. The spectral width of the signals is <10cm^-1 enabling a multitude of applications.

We introduce a Fourier-transform hyperspectral microscope based on an ultrastable birefringent interferometer. The microscope enables wide field acquisition with broad spectral coverage, tunable spectral resolution, high sensitivity and short acquisition time. We present the prototype of an add-on to a commercial microscope. We provide examples of applications in biology and solid state physics. The microscope is suited for fluorescence and Raman imaging.

Understanding cellular iron homeostasis is critical to understanding cancer cell survival and proliferation, as this process includes balancing iron uptake with storage and utilization. Iron-bound transferrin (holo-Tf) will bind to the transferrin receptor (TfR) at the cell surface and undergo endocytosis where iron is released into a mildly acidified endosome. Inside the endosome the iron is reduced for transport across the membrane for utilization or into the cytosol for storage in ferritin (Ft). It remains unclear whether iron uptake and storage regulation remain coordinated processes in breast cancer cells. Normally, it is expected for TfR and Ft protein expression to be inversely related based on their regulation via iron regulatory proteins (IRP1/2); however, increased expression of both TfR and Ft have been expressed in heterogenous breast cancer populations. To address the heterogeneous populations, single-cell analysis with Raman hyperspectral imaging could evaluate the relationship of iron uptake and storage through identification of iron-bound Tf and Ft in unlabeled cells. Raman hyperspectral imaging at 532 nm excitation has facilitated the imaging of iron-bound Tf in unlabeled cells. It indicated disrupted Tf iron-release in triple-negative breast cancer cells (MDAMB231), but not in the luminal A breast cancer line (T47D). Our data suggests that 532 nm excitation of Ft results in unique spectra. Currently we are collecting data on the unlabeled breast cancer cells to determine the relationship of iron-bound Tf and Ft by single-cell Raman hyperspectral imaging. This method will accelerate our understanding of iron homeostasis in breast cancer cells.

Optical detection of small nanoparticles and single molecules often relies on fluorescence, but limited photophysics and the need for labeling pose severe restrictions for the broad application of this approach. The ubiquitous process of Rayleigh scattering offers a powerful alternative. Although the common intuition might be that detection of individual nanoparticles and single molecules is not within reach via the measurement of their Rayleigh scattering, interferometric scattering (iSCAT) microscopy introduced in 2004 has demonstrated the contrary. In this presentation, I will present the most recent advances in iSCAT analysis, reaching an exquisite sensitivity for detecting single unlabeled proteins as light as 20 kDa. I also discuss results on the detection and microscopy of SARS-Cov-2 viruses in a high-security biosafety environment.

Mid-infrared photothermal (MIP) imaging is an emerging technique where a visible beam is deployed to sense the photothermal lensing effect induced by infrared absorption of molecules. This technology provides sub-micron spatial resolution defined by the visible probe beam. Yet, the photothermal lensing effect is a weak effect for most materials. For instance, the diffraction coefficient of poly (methyl methacrylate) changes around 0.1% per Kelvin. Thus, highly sensitive probes are desired to improve the detection sensitivity. Here, we report a fluorescence-enhanced mid-infrared photothermal (FE-MIP) microscope with high sensitivity. MIP microscopy uses a pump-probe strategy in which a mid-infrared light vibrationally excites a sample and a visible light probes the thermal effect. Instead of measuring the scattering modulated by mid-infrared absorption, we deploy thermal sensitive fluorescent dyes as the probe, and measure the modulated fluorescence intensity in both confocal mode and wide field mode. We demonstrate high imaging sensitivity and component specificity through fluorescence labeling.

The current paradigm for chemotherapy in the treatment of colorectal cancer (CRC) is based on the maximum-tolerated dose approach (MTD), which uses high doses of cytotoxic agents with the aim of killing tumor cells. However, the toxicity of MTD chemotherapy requires drug delivery in a cyclic manner with rest periods that allow the tumor to recover. In contrast, metronomic chemotherapy (MET) uses low but continuous doses that target the endothelial cells of tumors, thereby directly affecting the development of new blood vessels (angiogenesis). Because the potential impact of MET when used as neoadjuvant chemotherapy (NAC) in CRC has not been explored and contrasted with the effects in perfusion of the standard MTD regimen, this study uses an azoxymethane-induced primary model of CRC to assess changes in perfusion and angiogenic markers. Diffuse reflectance spectroscopy (DRS) was used to longitudinally quantify oxygen saturation (StO2) and compare these changes with the change in expression of VEGF-A between MET and MTD treatments. In the MTD group, an increase in StO2 from 65 to 81% was observed by week 4 of treatment and a reduction of VEGF-A expression from 81% to 16% was observed between weeks 4 and 6.

Identifying the microbial species present in a sample is critical in healthcare, industry, ecology and even national security. The traditional diagnosis process involves the growth of colonies of microorganisms on a solid-state medium in a Petri dish. This step provides great amounts of biological material, allows spatial separation of the different microorganisms and can help to visually discriminate potentially harmful species. However, it does not allow a precise identification and thereby requires further sample preparation and analysis to offer a proper diagnosis. Here, a new optical-based Petri dish analysis technology is discussed. This technique, called lensless multispectral mid-infrared imaging, relies on the acquisition of images at nine wavelengths corresponding to relevant chemical functions. It provides both morphological and discrete spectral data, which allows to discriminate even closely related species. The instrument is simply made of a laser source (four quantum cascade lasers) and a microbolometer array as the imager. A total of 1050 colonies belonging to three Staphylococcus species and two strains of Staphylococcus epidermidis have been acquired. After feature extraction and classification by Support Vector Machine, the tenfold cross-validation test yields a correct identification rate between 93 % and 96 %, with only a 5 % confusion between the two strains of S. epidermidis. Further work on the data analysis algorithms could dramatically improve these already promising results. Therefore, considering its label-free and non-destructive aspects, as well as its absence of secondary sample preparation, this technology has a great potential to offer precise Petri-dish based diagnosis.

Cell mechanical properties may undergo changes as a response to external stimuli or altered environment. Being able to map the mechanical properties over an entire cell in a contact-free and label-free manner is crucial to avoid affecting the measurements. Brillouin imaging is a non-invasive all-optical modality that can provide a distribution map of high-frequency elastic modulus and viscosity. In this report we first provide an overview of prior Brillouin cell imaging studies. We then demonstrate viscoelastic 2D and 3D cell imaging capabilities of a molecular filter-based Brillouin imaging system, with a first-time imaging application of multi-excitation (ME) Brillouin elastography.

Rapid detection and identification of novel viruses, such as SARS-CoV-2, is critical to treat, isolate, or hospitalize those infected, ultimately, to curb the spread of the virus. Diagnostic assays, such as enzyme-linked immunosorbent assay (ELISA) and polymerase chain reaction (PCR), are considered the gold standard for testing, but are labor-intensive and/or involve creating probe molecules specific to the virus. We propose Raman spectroscopy as an alternative method of detection because it can be label-free and can offer identifying information on many analytes within a sample. Specifically, we are developing a vacuum ultraviolet (VUV) Raman spectrometer using an incoherent excitation source that emits the hydrogen Lyman-α line radiation at 121.57 nm. The main benefit of using a VUV source is that the Raman scattering cross-section is inversely proportional to the fourth power of wavelength, which means that VUV photons would yield several orders of magnitude higher scattering cross-sections than excitation in the visible or near-infrared. In addition, the 121.57 nm wavelength transmits up to 2 cm in dry air, so the analyte can be oriented in standard testing formats. We have been successful in producing strong and consistent HLA emission using both a dielectric-barrier discharge (DBD) plasma source and a commercial source. In the near future, we expect to produce consistent Raman signatures in solid and liquid media. HLAS will transform diagnostic medicine and several other industries through its powerful capabilities of detecting real-time infections and important health markers.

Nailfold capillaroscopy is a technique for imaging the capillary bed in the finger nailfold, which is used in the diagnosis of scleroderma. Knowledge of the capillary oxygenation profile would be advantageous in disease evaluation due to the suspected involvement of hypoxia in causing fibrosis. A hyperspectral nailfold capillaroscopy system with narrowband illumination provided by a supercontinuum laser and acousto-optic tuneable filter was created to enable spectral analysis of the nailfold. Hyperspectral imaging (HSI) data from 500-595 nm was analysed to extract image quality metrics, which suggested HSI oximetry holds promise for understanding how rheumatic diseases affect oxygenation of the nailfold.

Pain is currently assessed using subjective measurements, often not aligning with clinical symptoms. Therefore, objective pain level assessments, using minimally-invasive and molecular methods, are needed to assess disease activity and response to treatment in osteoarthritis and rheumatoid arthritis. We report sophisticated quantitative biochemical “signatures” from the label-free hyperspectral imaging (HSI) of cartilage tissue for the characterization of molecular composition, structure and functional status. Further study on sinuvium tissue provides evidence that HSI could be used as a novel technique to delineate disease state. Additionally, HSI could be used to objectively separate individuals based on pain severity providing molecular correlates of pain.

We explored a Mask Region-based Convolutional Neural Network (Mask R-CNN) to detect macrophages in quantitative phase images, which were acquired by digital holographic microscopy (DHM), an interferometry-based variant of quantitative phase imaging (QPI). The Mask R-CNN deep learning architecture is capable to detect and segment single macrophage cells in quantitative phase images and allows to perform both tasks in a multi-stage process. Our results show that the combined detection and segmentation of cells through Mask R-CNN-based automated evaluation prospects a fast and robust screening in label-free high throughput microscopy.

Reactive oxygen species (ROS) play an essential role as cellular messengers, functioning as redox regulators under normal physiological conditions. However, the excessive production of ROS in cells and organs due to disorders such as diabetes mellitus, inflammation, cardiovascular, cancer, and neurodegenerative disease leads to oxidative stress, which may be an early indication of progressive pathology . Currently, ROS -specific indicators which requires labelling are used for ROS quantification. Therefore, a label-free imaging technique is desirable for assessing the level of ROS. In this work, we introduced a novel imaging method to quantitatively identify ROS in cells and tissues named autofluorescence multispectral imaging (AFMI). This technique involved a custom-built spectral imaging system with 18 spectral channels with distinctive excitation and emission wavelength spanning specific excitation (365 nm-495 nm) and emission (420 nm-700 nm) wavelength ranges. Such a system can extract rich spectral information related to the ROS present in cells and tissue. We correlated the spectral information obtained from AFMI to the level of ROS acquired from CellROX imaging, which served the reference ROS value in this study. Further, our analyses were repeated for UV sensitive applications where an excitation spectrum less than 400nm was avoided. Quantitative analysis of the spectral images showed a strong multispectral signature correlating the spectral variable with the ROS level in the cells and tissue. Our results showed that ROS levels can be determined non-invasively using AFMI, which potentially can be translated to future clinical applications where ROS are known to correlate with progressive disease.