Optical Biopsy XXIII: Toward Real-Time Spectroscopic Imaging and Diagnosis

This study employs dynamic laser speckle analysis (DLSA) to differentiate red blood cell (RBC) sedimentation within a cuvette containing non-Newtonian oils, positioned behind medical phantoms made of Intralipid. DLSA revealed significant differences in sedimentation patterns due to the proximity wall effect of veins. This non-invasive method offers a quantitative assessment tool for similar physiological phenomena, with potential clinical applications for detecting RBC sedimentation under tissue.

Our study demonstrates the application of Raman spectroscopy and Raman imaging combined with a relatively cost effective methodology known as reverse stable isotope probing to monitor microbial oil synthesis from hydrophilic (glucose) and hydrophobic (cooking oil & waste cooking oil) carbon sources. This method tracks and maps the microbial oil buildup and distribution over time at the single cell level, offering an extraction-free approach for in-situ microbial metabolite sensing and imaging. This can have a wide range of applications ranging from sensing commercially relevant sustainable microbial metabolite build-up from trash and pollutants.

Early cancer diagnosis is vital for better outcomes and cost savings in healthcare. Biopsy histopathology is key but relies heavily on pathologist expertise. Optical techniques like micro-infrared spectroscopy offer promise by detecting biochemical changes before visible tissue alterations. This method analyzes light interaction to extract data but faces challenges like artifact contamination (water vapor, paraffin) and baseline distortions due to sample heterogeneity. A new mathematical approach was developed using 192 micro-FTIR images of thyroid and mouth tissues, correcting artifacts with principal component regression for contaminants and linear coefficients for baseline variance. Quality tests removed non-histological pixels, and k-means segmentation yielded 90% correlation with hematoxylin-eosin images, validating effectiveness.

Research is underway to use machine learning to identify tumors and tissues from surgical images, but conventional endoscopes only capture visible light, limiting observation to surface layers, making it difficult to identify tissues that are similar in color and shape. Therefore, near-infrared multispectral imaging (NIR-MSI) is gaining attention, but existing device is bulky and long imaging time, hindering clinical use. In this study, we developed NIR-MSI Laparoscope system with excellent portability, high-speed imaging, transmission analysis, and object identification. The light source by combination of LEDs and rotating mechanism can minimizes light loss and enables high-speed wavelength switching. The system performs everything automatically, from data acquisition to neural network analysis and result display. In verification experiments, the system was able to observe the letters below through a 5 mm thick whisper and discriminate four types of transparent resin of the same color within 2 seconds with an average accuracy of 96%.

This pilot study aims to detect and characterize the spectral profile of blood plasma in Normal and DM subjects, using stokes shift spectroscopy (S3). S3 spectra are measured by simultaneously scanning both the excitation and emission monochromators with fixed wavelength interval Δλ=70 nm in spectral range 250─550 nm. Characteristics, highly resolved fluorescence emission peaks owing to tryptophan, NADH, and FAD, with significant spectral differences observed between normal and DM subjects. In order to quantify the observed spectral differences between normal and DM subjects, three potential spectral intensity ratio (SIR) parameters were calculated and the diagnostic sensitivity and specificity were determined. Result of the pilot study demonstrated that S3 spectral changes due to tryptophan, NADH, and FAD, have good diagnostic potentials; therefore, these fluorophores can be used as a native biomarker in discriminating the early stages of DM.

Intraductal carcinoma of the prostate (IDC-P) is a very aggressive histopathological variant of prostate cancer for which there are no accurate biomarkers. In our work, we apply a multimodal nonlinear optical imaging approach which uses second-harmonic generation (SHG) and stimulated Raman scattering (SRS) imaging to distinguish IDC-P from regular prostate cancer and benign prostate tissue. Images from each tissue type were classified using support vector machine (SVM) which classified the images from each region on the basis of first-order statistics and texture-based second-order statistics derived from the gray-level co-occurrence matrix (GLCM) of the images.

Raman spectroscopy is a label-free imaging technique that relies on subtle changes in the metabolic composition of a sample and has been shown to have good early cancer diagnostic potential. Raman measurements can be done on dried sample droplets, which follow the effects of Marangoni and capillary flow as they are drying, causing non-homogenous particle deposition, leading to subtle differences in the Raman spectral measurements. This research investigated these spatial spectral differences by measuring across the area of sample droplets in order to determine if there is an optimal measurement location such that the disease and healthy samples can be distinguished with higher accuracy than with randomly selected locations. The ultimate goal for this research is to discover the location on the samples that produces the best diagnostic results so that the measurements can be collected automatically, without the need for an expert user.

Raman spectroscopy is a widely utilized analytical tool for chemical and structure assessment, material characterization, and biomedical imaging, offering insights into chemical interactions. However, accurate interpretation of Raman spectra is often hindered by various noise sources, differing in nature and temporal evolution. This report presents a novel methodology for spectroscopic data handling aimed at improving signal integrity and analytical precision. We propose and demonstrate a novel cosmic ray suppression routine using a modified Z-score algorithm, a modified baseline correction via asymmetric least squares smoothing, and signal enhancement through data averaging and Savitzky-Golay filtering. Additionally, we highlight an efficient approach for fitting Lorentzian functions to spectral peaks, facilitating precise quantification of peak parameters. The effectiveness of these methods is demonstrated using both synthetic and experimental Raman spectra, with the underlying algorithms available in an open-source GitHub repository. This work provides a framework for researchers and practitioners to enhance the accuracy and reliability of Raman spectroscopy data analysis.

We present cutting-edge, labelfree spectroscopic approaches in the field of biophotonics for precise tumor margin control, offering a valuable solution to the challenges currently encountered in this area. Our objective is to completely remove tumors and enable reliable tumor typing and classification, allowing us to rapidly create individual treatment plans. We utilize artificial intelligence (AI) to achieve this. In the second part of our presentation, we will demonstrate our work on the application of Raman spectroscopy for the successful treatment of infections. This includes the determination of the immune response, the rapid identification of the pathogen and its resistance pattern, as well as the response to treatment.

Raman spectroscopy represents a unique optical vibrational technique capable of harvesting biochemical information about tissue for label-free histopathological assessments. In this work, we present the development of miniaturized fiberoptic Raman spectroscopy techniques which can simultaneously acquire both the fingerprint (FP) and high wavenumber (HW) tissue Raman spectra within 1s, and assess its clinical utility for improving real-time in vivo cancer diagnosis and detection in GI tract and other organs (e.g., bladder and head and neck) during endoscopic examination. We have demonstrated the diagnostic accuracy of >90% for GI cancer and precancer detection with FP/HW Raman spectroscopy. The rapid fiberoptic Raman spectroscopy also allows the delineation of tumor margins for immediate surgical resection as well as enables the rapid assessment of post-treatment efficacy and recurrence in head and neck patients. This work illustrates that fiberoptic Raman endoscopy technique can serve as an objective diagnostic tool for mass screening and surveillance of cancer patients at high risk to significantly improve the cancer management and patients’ quality of life in healthcare.

Bladder cancer (BC) treatment is costly due to high recurrence and revision surgeries. If a resected tumor is diagnosed as high-risk by post-operative histological assessment, revision surgeries are mandatory. We present a fully integrated, clinically compatible SRS imaging device for real-time histological assessment of BC during surgery. We demonstrate excellent agreement between acquired SRS images and classic H&E-stained images. Our device paves the way for classifying BC risk levels, enabling intraoperative therapy stratification. Additionally, our device spans the full Raman spectrum, offering rapid, detailed molecular profiling and potentially reducing reliance on advanced diagnostic tools.

Modern cancer treatments that harness the power of the patient’s own immune system have significantly impacted survival for certain patients. However, cancer clinicians currently lack predictive markers to identify which patients will be responders versus those with severe adverse effects. Here, we present a spatial-omics platform based on surface-enhanced Raman spectroscopy (SERS) to provide rapid, subcellular spatial profiling that enables non-destructive therapeutic response characterization. Our approach combines silicon-based metasurfaces with Raman spectroscopy and machine learning to demonstrate subcellular differentiation and functional state characterization of the TME in melanoma.

We have developed an advanced nonlinear multimodal imaging system that integrates Stimulated Raman Scattering (SRS), multiphoton fluorescence (MPF), and second harmonic generation (SHG) to investigate the intricate relationships between metabolic activities and metabolite distribution in organelles of cells and tissues. This system is enhanced by innovative super-resolution image deconvolution and correlation coefficient mapping algorithms, enabling deeper insights into various metabolic processes within super-resolved nanoscale regions of interest. Additionally, we introduced a state-of-the-art chemical image clustering algorithm to precisely identify signals from specific subcellular organelles. This comprehensive approach holds significant potential for improving disease detection and prognosis, assessing therapeutic outcomes, and advancing our understanding of aging and biomedical science.

Raman spectra of plasma cells obtained from patients with multiple myeloma were classified by a machine learning model. To predict unknown samples, we optimized the model and evaluated its performance toward clinical applications.

In this study, a multimodal imaging system was used to evaluate the efficiency of 5-FU@UIO-66-NH₂ and 5-FU@UIO-66-FA as 5-FU carriers in colorectal cancer（CRC）cells. Two-photon excitation microscopy revealed alterations in the metabolic coenzymes NADH and FAD levels, indicating metabolic shifts occurring in the cells post-nanoparticle treatment. Raman microscopy analyzed CRC cells for alterations in biomarkers and differences in molecular metabolism. The results demonstrated that 5-FU@UIO-66-FA effectively released 5-FU and exhibited a superior inhibitory effect on CRC cell proliferation. This illustrates the potential of multimodal imaging in the analysis of CRC cells with nanoparticle drug delivery, enabling precise molecular and metabolic assessments.

We developed a digital counting method of single enzyme biomarkers using surface-enhanced Raman scattering (SERS) spectroscopy. Single enzymes were stochastically encapsulated in numerous microchamber array decorated with silver nanoparticles. Presence or absence of a single enzyme in each chamber was digitally quantified by SERS signal of reaction products with uniform and reproducible signal amplification. Using this method, enzyme biomarkers such as acetylcholinesterase and butyrylcholinesterase in solution were quantified with fM sensitivity and high molecular selectivity. We also developed a wide-field imaging system for rapid counting of trace enzyme biomarkers in body fluids, demonstrating the potential of SERS-based digital liquid biopsy.

We have used infrared hyperspectral imaging to analyse ~1400 tissue cores from 183 patients who are presenting with prostate cancer. Using machine learning algorithms, we have identified a cohort on nominally low risk patents that have a survival rate almost as poor as patients presenting with metastatic disease. Given the infrared analysis, these patients should be offered more aggressive treatments than currently recommended.

We report lung cancer detection using liquid biopsy. Machine learning-assisted ATR-FTIR spectroscopy of serum achieved 81% specificity and 79% sensitivity. This method requires no labelling or processing beyond isolating serum from blood and takes less than one minute. Samples from 52 participants with confirmed lung cancer and 52 samples from participants matched by age, sex and current smoking status were analysed. ATR-FTIR spectroscopy was used to measure 9 μL samples in liquid form. Specificity and sensitivity were investigated using different machine learning classifiers and figures of merit. Combining measured spectra with medical history and a random forest classifier gave optimum performance. Our strategy for improving sensitivity and sensitivity and for point-of-care integration will be discussed which may lead in the future to use as a triage for patients to determine which patients require further investigation for lung cancer.

Lipid metabolism and signaling are crucial in biology and disease development, yet techniques for lipid bioimaging in tissues are limited. We present the design of a prototype unified instrument combining Raman Spectroscopy (RS), Fourier Transform Infrared Spectroscopy (FTIR), and Desorption Electrospray Ionization-Mass Spectrometry Imaging (DESI-MSI) with 50 µm spatial resolution. Here we demonstrate complex yet complementary relationships between RS/FTIR spectra and MS molecular abundances of tissues. Finally, we show how DESI-MS lipid species can be predicted from Raman data using a machine learning framework.

Esophageal cancer surgery is challenged by anastomotic leakage (AL), affecting 10-20% of patients and leading to severe complications. AL’s main cause is tissue hypoxia. Fluorescence imaging (FI) is increasingly used to quantify tissue perfusion but suffers from drawbacks such as an invasive dye and a lack of signal quantification. To improve surgical outcomes, we evaluate non-invasive multispectral imaging (MSI) for assessing tissue oxygenation during laparoscopic esophagectomy in humans. MSI was performed at video rate using a visible light snapshot camera and was compared with FI. Preliminary results show a critical difference in oxygenation levels between patients with and without AL, highlighting its potential for real-time hypoxia assessment. Moreover, HSI and FI are found complementary, but differences arise when evaluating the anastomotic site. HSI could reduce AL rates and improve surgical outcomes. An ongoing clinical study aims to validate these findings against lactate levels and hypoxia markers.

Schizophrenia and bipolar disorder often present similar clinical features, making accurate diagnosis challenging, and their treatments differ significantly. To address this issue, our study measured brain blood flow changes in healthy individuals, schizophrenia patients, and bipolar disorder patients during a semantic fluency test using near-infrared spectroscopy (NIRS). We combined this data with deep learning and interpretable artificial intelligence techniques to assist doctors in diagnosis. In a two-stage three-class classification, our model achieved over 90% training and testing accuracy for both the control and disease groups, as well as for distinguishing between schizophrenia and bipolar disorder. These results clearly demonstrate that our experimental method effectively utilizes blood oxygen information to differentiate between mental disorders, confirming the feasibility of our approach. This research can help doctors improve diagnostic accuracy, reduce treatment costs, and enhance patient outcomes.

In laparoscopic surgery, there are situations in which important tissues such as nerves and blood vessels that should not be resected are buried deep in the tissue, but it is difficult to recognize them because of their low transparency in visible light. Therefore, we focused on near-infrared hyperspectral imaging (NIR-HSI), which utilizes the biological transparency of over 1000 nm in near-infrared (OTN-NIR) light and its ability to analyze components by absorption spectroscopy, and it might be possible to visualize deep tissues in the body. However, there were no reports of OTN-NIR imaging devices under laparoscopic conditions. Therefore, we realized a system that enables spectral imaging at wavelengths from 490 nm to 1600 nm under laparoscopic conditions. The results of imaging and spectroscopic imaging analysis of a live pig using this system suggest that it is possible to visualize blood vessel in deep tissue.

The mid-infrared (MIR) spectral regime has gained interest in biosensor research serving as a direct diagnostic tool due to the unique excitation of specific molecular vibrational modes of biomarkers and biomolecules with inherent selectivity. Consequently, this presentation focuses on the role of mid-infrared sensing technologies in liquid biopsies emphasizing the crucial requirement for accurate detection and real-time monitoring in early disease diagnosis and monitoring of disease progression.

We developed photoacoustic tomography (PAT) for deep-tissue imaging, offering in vivo functional, metabolic, molecular, and histologic imaging from organelles to entire organisms. PAT combines optical and ultrasonic waves, overcoming the optical diffusion limit (~1 mm) with centimeter-scale deep penetration, high ultrasonic resolution, and optical contrast. Applications include early cancer detection and brain imaging. Additionally, we developed light-speed compressed ultrafast photography (CUP), capable of capturing the fastest phenomena, such as light propagation, in real time. CUP, with a single exposure, captures transient events on femtosecond scales. CUP can be paired with various front optics, from microscopes to telescopes, facilitating diverse applications in fundamental and applied sciences, including biology and cosmophysics. Further, our research extends to quantum entanglement for imaging. Quantum imaging utilizing Heisenberg scaling enhances spatial resolution linearly with the number of quanta, outperforming the standard quantum scaling’s square-root improvement.

Retinal diseases such as age-related macular degeneration (AMD), retinal pigmentary disorders, and diabetic retinopathy are leading causes of vision loss worldwide, with current treatments offering limited efficacy in improving vision. In this study, we present an advanced multimodal imaging system that integrates photoacoustic microscopy (PAM), optical coherence tomography (OCT), and fluorescent imaging into a unique platform for tracking stem cell viability, migration, and replacement of damaged retinal tissue. Additionally, this system was employed to monitor adeno-associated virus (AAV) transfection for gene therapy. Experiments were conducted on 21 rabbits with clinically relevant models such as laser-indced retinal pigment epithelium (RPE) damage. Results demonstrated that the transplanted stem cells successfully grafted onto the rabbit retina and progressively replaced damaged RPE over six months. Furthermore, the efficiency of cell transfection was detectable via OCT and fluorescent imaging approximately two weeks following subretinal injection of AAV serotypes 2, 5, and 8. These findings pave the way for new treatment avenues for various retinal diseases.

Many techniques exist for screening retinal phenotypes in mouse models in vision research, but significant challenges remain for efficiently probing higher visual centers of the brain. The purpose of this study is to implement photoacoustic computed tomography to brain imaging in unanesthetized and unrestrained, i.e. free-moving, mice. A headmount integrating PA illumination and reception, and visual stimulation devices was fabricated. Preliminary results show that the free-moving mice have higher magnitude (n=3,8, p<10-4) and longer sustain in their responses compared to the anesthetized ones.

Vibrational photothermal (VIP) microscopy opens a new window to look at molecules and molecular interactions inside a biological system including virus, bacterium, cell, and organism. In VIP microscopy, a pump beam excites chemical bonds via mid-infrared absorption, stimulated Raman induced absorption, or shortwave infrared absorption. A probe beam measures the local change of refractive index or thermal expansion of a particle induced by the photothermal effect. Since the first demonstration of mid-infrared photothermal (MIP) imaging of living cells and organisms (Science Adv 2016), our team has advanced this pump-probe chemical imaging technology in three ways, namely scanning-based confocal MIP microscopy, camera-based wide-field MIP microscopy, and computation-based MIP tomography. Since its commercialization into a mIRage system in 2018, MIP microscopy has found broad applications, spanning the analysis of functional materials, characterization of environmental microplastics, and structural detection of protein aggregation in neurological diseases. More recently, we developed stimulated Raman photothermal microscopy (Science Adv 2023) and shortwave infrared photothermal micros

The genome is organized non-randomly, in a tissue- and development-specific manner. Structural and molecular changes within the interchromatin regions and alterations of cytoplasm organelles may be identified in buccal cells (cheek swabs) from patients at different stages of Alzheimer Disease (Mathur et al. 2014; Garcia et al. 2017). Infrared and Raman spectra provide complementary, molecular fingerprint information that can facilitate the identification of spectroscopically relevant clinical biomarkers. We will present multimodal fluorescent- and optical- photothermal IR (FL-PTIR, O-PTIR) and Raman spectrochemical imaging of buccal cells, correlated with conventional and superresolution fluorescence imaging of the same cells following staining. Our goal is to develop a method that is not only feasible but will be clinically relevant for a simple, safe, non-invasive, inexpensive, objective measure for staging Alzheimer Disease that will save healthcare costs and optimize patient treatment strategies.

Nano- and microplastics (NMPs) are increasingly raising concerns about their potential health effects. Traditional methods for detection and characterization of NMPs in tissue require elaborate sample preparation, such as labelling or time-consuming tissue digestion. In this contribution we report on the application of optical photothermal infrared spectroscopy (O-PTIR) for localization and identification of plastic particles significantly smaller than 1µm. We show that the method can be applied to routine clinical formalin-fixed paraffin-embedded (FFPE) tissue samples, preserving tissue architecture while accurately locating NMPs. Since this technique is label-free, contactless and non-destructive it can be applied prior to routine histopathology. We present results on various types of clinical samples, among them 3D cell cultures, mouse kidney tissue, human colon tissue, etc.

Our studies utilize label-free, two-photon microscopy to achieve high-resolution, morphofunctional imaging for precision diagnosis and treatment. This technique captures metabolic and morphological data without exogenous labels. Specifically, we characterize mitochondrial organization changes in vitiligo lesions that are abrogated only in patients undergoing succesful treatment. For cervical pre-cancer, metabolic reprogramming indicators distinguish high-grade lesions with high sensitivity and specificity. In vivo and upcoming endoscopic applications demonstrate the technique's potential for non-invasive, real-time diagnostics, improving outcomes in dermatology and oncology.

Using femtosecond pump-probe microscopy in clinical melanoma biopsies, we have developed an optical biomarker that has the potential to evaluate the metastatic potential of the primary cancer lesion and, hence, help guide treatment approaches. We will provide updates on our latest progress in experimental techniques and analysis methods to enhance the performance and reliability of this technique.

Rayleigh wing scattering can probe molecular fluctuations from the dielectric constant of molecular motions such as rotations, translations, redistribution, and structural properties in liquids. We measured the THz Raman spectra, showing peaks associated with structures in the liquid state which is on top of the bell-shaped Debye spectral profile associated with the relaxation phenomena within the liquid. We fitted the Rayleigh wing based on the Debye theory to reveal the relaxation times of the fluctuating motions in the liquids. We observed for the first time peaks in the spectra which were attributed to temporal quasi topological crystal structures in liquids.

The THz-Raman spectra from 5cm-1 to ~333 cm-1 were measured to investigate the inter- and intra-molecular motions that occur in the state of the tissue being normal or diseased. The molecular processes from the THz Raman spectra are used as a biomarker to characterize different tissue states, such as neurological (Alzheimer’s Disease), cancerous, diabetic and metabolism states. In this study, we used the THz-Raman spectroscopy to investigate normal and cancerous tissues of brain and skin melanoma, kidney diseases, key animo acids such as tryptophan, and collagen. The metabolism, aging, and degradation of ex vivo chicken tissues were also studied.

We present a novel low coherence interferometry/fiber Bragg sensing technology for guiding difficult cases of epidurals. The results of an animal study will be presented to demonstrate the suitability of the technology for further in vivo human testing.

This study presents a portable optical coherence tomography imaging system that provides intraprocedural feedback on the precise selection of the biopsy location. It uses a hand-held optical imaging probe supported by imaging processing pipeline based on deep learning neural networks. The preliminary evaluation of the integrated system was performed on an animal model of cancer, where the collected dataset was used to develop the deep learning-based tumor segmentation software. A patient study for technology clinical validation is on-going at MD Anderson Cancer Center and the preliminary results will be presented.

We developed a simple, robust method of customizing the intensity statistics of laser speckles and introducing nonlocal correlations among the speckle grains. The tailored speckle patterns exhibit radically different topologies and varying degrees of spatial order. Our ability of controlling the speckle statistics enabled the customization of speckled illumination patterns for bioimaging applications. One example is parallelized super-resolution microscopy. We design and create special speckle patterns for parallelized nonlinear pattern-illumination microscopy based on fluorescence photoswitching. The spatial resolution is three times higher than the diffraction limit of illumination optics. We further show that tailored speckles vastly outperform standard speckles.

Continuous measurements of blood pressure without a cuff remains a long-standing unmet need in biomedicine. I will present a brief review of current technologies and then present my lab’s recent progress on using speckle contrast optical spectroscopy for highly accurate measurements of cuff blood pressure.

Our work proposes the use of a collinear femtosecond Optical Kerr Gate for the first time to study the temporal behavior of the molecular interactions in different tissues. This study presents the OKE as a new optical biopsy method to differentiate different types of tissues with different degrees of diseases. The main biomarker observed in our study is the doubling in the tissue’s conductivity from the dielectric response time, associated with the conductivity and permittivity observed in different grades of breast cancer tissues. Our finding suggests conductivity from electronic and ions in plasma in a tissue can be used as a new major biomarker for the classification or detection of diseases.

This presentation theoretical focusses on our understanding of the underlying mechanism behind the generation of higher harmonics from electronic self-phase modulation which is driven by the Electric Field E(t) of the optical, NIR, and MIR femtosecond light pulses to produce attosecond laser pulses and providing the direction to produce zeptosecond laser pulses. The cutoff frequency is shown to depend on I/λ2 . The HHG arises from driving the phase of the electric field, E(t) by itself to create HHG in form of odd Bessel functions in time of odd harmonics driven by laser pulse from the nonlinear refractive index of Kerr media, n2 (χ3) and extending to even HHG from n1 (χ2). The three characteristic features of HHG spectrum are initial decreasing harmonics, a plateau region of HHG, and cutoff frequency. We show theoretically in the EM Kerr electronic self-phase modulation (ESPM) model that the distinctive features of HHG arise mainly from the changes in the phase of the E(t) wave driven by the envelope of the laser pulse causing the cosine of the cosine squared for χ3 and the cosine of the cosine for χ2 in time.

Pairing up optical microscopy and computer vision becomes a common strategy adopted in a broad spectrum of biological and biomedical screening applications. The common rationale is to generate the characteristic "fingerprint" profiles of cell morphology that could underpin the cell states/functions, but obscured through visual inspection or even in the molecular assay. However, it remains unachievable or affordable with current technologies to record, integrate, and analyze all relevant cell morphological data. The synergism between ultrafast imaging, microfluidics, and deep learning allows us to overcome some of these current limitations. This talk will highlight a few notable high-throughput, deep-learning-powered imaging techniques and analytical cytometry pipelines over the past few years. These platforms allow us to significantly scale up the single-cell biophysical/mechanical phenotyping throughput (beyond millions of cells per run) for rare-cancer-cell detection, T-cell subtyping and activation, and tumor biopsy analysis. The talk will also discuss the latest strategies for enriching biophysical phenotyping content by integrating with genetic-perturbation assay.

As the final layer of the central dogma, metabolomes directly reflect phenotypic changes in human bodies. However, the cumbersome sample preparation process for mass spectroscopic measurement and the lack of spatial information in tissues limit their wide application in clinical practice. Here, we propose the use of fluorescent metabolomics, which consists of metabolites that can emit fluorescence in blood, saliva, and urine. In comparison to mass spectroscopic measurement, fluorescent metabolomics allows for direct measurement without the need for reagents. It could serve as a continuous monitoring device for detecting the emergence of acute illnesses. Furthermore, when combined with a multiphoton imaging system, spatial metabolomic information can be obtained with sub-cellular resolution. In this lecture, we will demonstrate the use of fluorescent metabolomics in evaluating organoid pharmacokinetics, detecting the presence of acute mesenteric ischemia, sepsis, acute kidney injury, and aiding in the differential diagnosis of prediabetes. This approach provides a new avenue toward time-course molecular diagnosis and advances the precision medicine.

Two photon fluorescent microscopy (TPFM) has shown the capability to produce H&E-like images similar to conventional paraffin embedded diagnosis without the need for long sample preparation times. We present interim results of an ongoing prospective clinical trial performing real-time diagnosis of nonmelanoma skin cancer in a dermatology clinic. Additionally, we will present technical innovations enabling diagnosis of patients as well as coregistration of real-time biopsy images to conventional paraffin sections. We demonstrate that TPFM can rapidly and accurately diagnose patients in dermatology clinic, potentially enabling immediate treatment.

We present an innovative 3D Mueller Matrix (MM) imaging technique for investigating the polycrystalline microstructure of dehydrated blood films. By leveraging the optical anisotropy properties of blood proteins, this approach enables detailed layer-by-layer analysis of blood samples, offering significant potential for early disease detection, particularly in cancer diagnostics. Through the integration of digital holographic reconstruction, we extract critical optical parameters such as linear and circular birefringence and dichroism, providing a quantitative evaluation of blood microstructures. Our findings demonstrate over 90% accuracy in differentiating between healthy and cancerous blood samples, making this a highly promising non-invasive diagnostic tool for early cancer detection. This method's ability to detect subtle structural changes in blood proteins opens new possibilities for advancing biomedical diagnostics and personalized medicine.

This study aims to investigate the antibacterial and anticancer efficacy of pure Zirconium Oxide (ZrO2) and Europium Doped Zirconium Oxide (Eu-ZrO2) nanoparticles (NPs) prepared using solution combustion method and characterized using various standard characterization techniques. X-ray diffraction studies confirmed that the synthesized NPs is of monoclinic structure with P21/a space group. Antibacterial efficacy of pure and Eu doped ZrO2 NPs were tested against gram positive (G+) Staphylococcus aureus stains and gram negative (G-) Escherichia coli bacterial strains and in both cases, enhanced antibacterial activity was observed. In addition to this, anticancer efficacy of synthesized NPs was tested against MCF-7 breast cancer cell lines using MTT cell viability studies and showed significant cytotoxic effect. These results substantiated that 5% Eu doped ZrO2 NPs possesses enhanced antibacterial and anti-cancer efficacy, holding potential as therapeutics for antimicrobial and anticancer applications.

Nonlinear microscopy (NLM) enables rapid evaluation of surgical specimens without the need for physical sectioned. Coupled with careful design of clinical workflow it has the potential to enable many new applications for cancer surgical guidance, reducing rates of repeat surgeries and improving outcomes. This presentation described NLM methods for real time histology and examples applications including breast cancer lumpectomy and prostate cancer surgery.

The formation of advanced glycation end products (AGEs) has been implicated in diabetic pathogenesis. Since some AGEs are fluorescent, fluorescent AGEs may be used as long-term glycemic biomarkers to monitor the extent of diabetic pathogenesis. In the case of wound healing, AGEs may hinder migration and contribute to apoptosis of fibroblasts, thereby disrupt wound healing. In this study, we study the relationship between fluorescence of glycated matrix from different monosaccharides and the migratory behavior of fibroblasts. Our results will provide insights on the use of fluorescent AGEs to monitor wound healing and improvement in monitoring diabetic pathogenesis.

Multiphoton fluorescence microscopy is a powerful tool for revealing complex biological systems with high contrast, minimal out-of-focus bleaching, and deep penetration. However, clinical application faces challenges like the need for stable, tunable laser sources, comprehensive biological information, phototoxicity, and developing universal quality control tools. We developed a new generation label-free super-multiplex multiphoton imaging microscopy, enabling co-registered label-free 4-photon (tryptophan), 3-photon (NAD(P)H), 2-photon (FAD, porphyrin/lipofuscin), second harmonic (collagen), third harmonic (optical heterogeneity), and fluorescence lifetime imaging with single-shot, single-band excitation near 1110 nm. Our laser source features tunable wavelength (950-1150 nm), pulse repetition rate (1-10 MHz), and pulse width (40-400 fs), with stability over 2000 hours. We revealed phototoxicity mechanisms in nonlinear imaging, identified safe imaging parameters, and developed a universal tool for objective imaging performance comparison. This system maximizes fluorescence microscopy information while minimizing harm, paving the way for clinical translation.

We present a novel detection scheme that allows the nonlinear microscope to register the spectral fingerprints of native biomolecules. This tailored instrument leverages a grating to disperse the nonlinear signals, a galvanometric mirror to scan each spectral component, and a high dynamic range photomultiplier tube to detect them. This device co-registers the spectral profiles of each probed sample point, delivering spectra with harmonic and multiphoton absorption fluorescence signals three orders of magnitude faster than a benchmark CCD-based spectrometer. By virtue of spectra, our approach exposes the manifold of biochemical species within the interaction volume, enhancing the granularity needed for investigating biological specimens. This detection scheme offers to nonlinear microscopy a contrast palette with complementary signals and superior analytical power.

Phototoxicity has become a prevailing issue in the super-resolution era when boosted illumination is applied, compromising the physiological relevance of the recorded data. We advocate leveraging chemical approaches to tackle phototoxicity. By exploiting chemical motifs such as triplet state quenchers and biocompatible auxiliaries, we systematically upgrade the commonly used fluorescent markers toward alleviated phototoxicity. These gentle dyes can be directed to various cellular targets spanning mitochondria, DNA, cytoskeleton, insulin granule, and specific proteins, enabling time-lapse super-resolution imaging with minimal photodamage. For example, PK Mito Orange probe is a mitochondrial inner membrane stain that enables 30 frames of STED recording and multi-color imaging of mitochondrial components. PK Zinc dyes enable multiplexed imaging of insulin secretion in isolated islets. These biocompatible probes, with high specificity and gentle behavior under excitation light, promise to offer reliable spatial-temporal information for imaging metabolically important processes.

This pilot study aims to demonstrate that Stokes Shift Spectroscopy (S3) is a specific and sensitive technique for the detection of normal and AD brain tissues taken from the Broadmann's area−17 (BA17) region. S3 signal is recorded by simultaneously scanning the excitation and emission monochromators with a constant wavelength interval of  = 20 nm in the wavelength region between 250 nm and 700 nm. Highly resolved bands of tryptophan, collagen, elastin, FAD, and porphyrin were obtained, with significant spectral differences observed between normal and AD-BA17 brain tissues. To quantify the observed spectral differences between normal and AD-BA17 tissue, four spectral intensity ratios (SIR) parameters were calculated and showed excellent diagnostic accuracy giving 100% specificity and 100% sensitivity for distinguishing normal and AD-BA17. Result of the pilot study demonstrated that S3 spectral changes due to tryptophan, collagen, elastin, FAD, and porphyrin have good diagnostic potentials; therefore, these fluorophores can be used as a native AD biomarker.

Chromophobe renal cell carcinoma (chRCC) and oncocytoma are malignant and benign kidney tumors, respectively, with very different prognoses. Accurate discrimination between the two is important for the clinical management of patients. In this study, multiphoton microscopy images collected from chRCC and oncocytoma were classified using convolutional neural network (CNN), yielding an accuracy of over 70%. The model is also capable of obtaining saliency maps to identify the most influential features, such as collagen structure and cytoplasm. Our current findings demonstrate the immense potential of deep learning for distinguishing chRCC and oncocytoma kidney tumors as well as other cancers in general.