High resolution optical imaging modalities such as optical coherence tomography and multiphoton microscopy have excellent specificity and sensitivity for early detection of cancer, however they may have very limited field of view. This small field of view makes visualizing the entirety of organs time-consuming or impractical. Wide field of view optical modalities such as narrowband or multispectral fluorescence imaging can be combined with high resolution imaging to identify suspicious areas for further investigation. Ovarian and gastric cancer both have poor prognosis unless they are detected early. We have utilized the dual field of view technique for examining the fallopian tubes and stomach, in order to create a viable screening method. Sub-mm to few-mm diameter endoscopes are configured with dual optical modalities, either combined in the same device or with a parent/babyscope design.

Esophageal adenocarcinoma (EAC), which can arise from Barrett’s esophagus (BE), has a 5-year survival rate of < 20%. Unfortunately, the severe sampling limitations associated with conventional histology may limit the sensitivity for detecting EAC and dysplasia (a precursor lesion to EAC) through regular endoscopic screening of BE patients. We have developed a non-destructive 3D pathology workflow to provide comprehensive evaluation of whole biopsies and a deep learning-based computational triage method that automatically segments potentially neoplastic regions (dysplasia or EAC) to guide pathologist review. A preliminary clinical validation study shows that our AI-assisted 3D workflow enables neoplasia to be identified with higher sensitivity on a per-biopsy level than conventional slide-based 2D histology.

Cancer research recently revealed that anticancer therapies can cause cell senescence instead of death, a phenotype governing tumor relapse. Developing safe, quick, and precise tools to spot such therapy-induced senescence (TIS) is an urgency. We present multimodal coherent Raman and multiphoton nonlinear optical microscopy as powerful to unveil TIS, via a home-built microscope including forward-detected Stimulated Raman Scattering, forward and epi-detected Coherent Anti-Stokes Raman Scattering, Two-Photon Excited Fluorescence and Second-Harmonic Generation modalities. We exposed early TIS in human cancer cells, confirmed comparing diverse signals during therapy period. We consider our findings will strongly influence anticancer practices, helping prevent tumor recurrence.

Cameras with extreme speeds are enabling technologies in both fundamental and applied sciences. However, existing ultrafast cameras cannot cope with extended three-dimensional (3D) scenes. To address this unmet need, we developed a new category of computational ultrafast imaging technique, light field tomography (LIFT), which can perform 3D snapshot transient (time-resolved) imaging at an unprecedented frame rate with full-fledged light field imaging capabilities, including depth retrieval, post-capture refocusing, and extended depth of field. As a niche application, we demonstrated real-time non-line-of-sight imaging of fast-moving hidden objects, which was previously impossible without the presented technique. Moreover, we showcased 3D imaging of fiber-guided light propagation along a twisted path and the capability of resolving extended 3D objects. I will also talk about our recent work on applying LIFT to image transient biological events at microscopic scales in this talk.

There is a need for an accurate, non-invasive, and real-time measurement of physiological hydration, in which we need to directly measure changes in water content and dynamics and changes in its molecular environment. Both sweat at the surface of your skin and deeper tissue properties are easily accessible non-invasively to Raman spectroscopy. Complexities in biomolecular systems require a fundamental understanding of the dynamics present in order to accurately assess changes in hydration using Raman spectroscopy. AIMD simulations will be made to compute changing vibrational spectra corresponding to the major hydration related changes in physiological molecular environments and compared to experiment.

In Raman-based diagnostic applications, principal component analysis (PCA) has often been used to distinguish different cell types or abnormalities. The performance of PCA greatly depends on the baseline adjustment of the measured spectra. Hence, the effect of erroneous baseline fitting on PCA requires to be addressed. Thus, we investigate the impact of baseline error for Raman spectra on PCA through the application of polynomial function with different orders in the fingerprint region (~600-1800cm-1). We found that the third order polynomial baseline fitting generated the fitted spectra closest to the mean spectrum and provided more precise PCA results.

Advances in computational image analysis and machine learning in the field of digital pathology seem poised to make transformational changes in disease diagnosis and prognosis, yet currently 99% of all histopathology slides generated in clinical practice are never digitized. We seek to incorporate digital imaging into the standard glass-slide clinical pathology workflow using relatively inexpensive modifications to traditional clinical microscopes. Here we will introduce PathCAM, a pathology computer-assisted microscope for "ambient" digitization of histology slides during the standard pathology workflow. PathCAM leverages high-frequency spatiotemporal sampling and human-machine interaction to deliver complete and accurate digital records of clinical glass-slide analysis. We will discuss the development of the PathCAM system and technical approach, and demonstrate the unique datasets and capabilities afforded by this hybrid digital-analog pathology imaging and analysis platform.

Biological imaging studies are often limited by a low amount of data, decreasing the reliability of typical machine learning methods. We attempted to address this by creating large numbers of synthetic second harmonic generation images that can be tuned to reflect properties of different disease classes. To start, we collected collagen images from a variety of healthy and diseased specimens. These were analyzed with a modified generative adversarial network (StyleGAN) combined with an encoder. After training, we were able to produce images that accurately reflected our samples. These results can be applied to increase the accuracy of classification algorithms and models of extracellular matrix tissue.

Circulating tumor cell clusters (CTCCs) are extremely rare events (<4 events per 7.5 mL of blood) found in the bloodstream of metastatic cancer patients. Despite their scarcity, they represent an increased risk for metastasis. Detection and isolation of CTCCs remain a priority for oncologists to improve cancer patients' diagnosis, stratification, and treatment. This study aims to demonstrate that confocal backscatter and fluorescence flow cytometry (BSFC) with deep learning-based signal analysis can enable sensitive detection of CTCCs in whole blood in vitro and in vivo without using exogenous labeling.

Fluorescence microscopy acquires multicolor images using fluorescence dyes that label specific structures or compounds in biological samples. Excitation and emission spectra for such dyes are often broad and multiband, which requires complex hardware or difficult decomposition methods and complicates the choice of labels to avoid spectral overlap. We propose a deep learning algorithm for channel decomposition. We trained two supervised pixel-to-pixel generative adversarial networks (GAN) to generate separate decomposed nuclear and cytoplasmic images from a single input image containing overlapping fluorescence from both dyes, achieving high similarity between ground truth and GAN images via SSIM and MSE.

Hyperspectral reflectance imaging can be used to develop tissue classification algorithms, which is often based on machine learning. To improve the performance of classification algorithms, pre-processing is often used to remove variations in data not related to the tissue itself. In hyperspectral imaging, these variations are the result of reflections from the tissue surface (glare) and height variations within and between tissue samples. We investigated and quantified the performance of 8 commonly used pre-processing algorithms to reduce differences in spectra due to glare and height differences, while retaining contrast between tissues with different optical properties on simulated and clinical datasets.

The altered placental function which is characteristic of preeclampsia is challenging to characterize with existing imaging techniques. We have developed ultrasound-guided spectral photoacoustic imaging of placental function in an in vivo rat model of preeclampsia, and investigated the impact of therapy on preeclampsia and placental function. Additionally, we have studied the acute vascular response during pregnancy – determining vascular selectivity and response time to vasoactive compounds — using photoacoustic tomography. We integrate our photoacoustic imaging methods with contrast-enhanced ultrasound to provide measures of placental vascular flow. By combining these hybrid optical-acoustic imaging techniques, we characterize placental pathologies and their responses to treatment.

Margin status directly correlates with patient survival in many types of cancer. To improve the identification of positive margins, we report a rapid surgical guidance workflow including tissue staining, open-top light-sheet (OTLS) microscopy imaging, and post-processing to generate en face histologic images of fresh tissue surfaces. Compared with conventional frozen section analysis (FSA) in which a few vertical sections of tissue are sampled, our technology could comprehensively image large margin surfaces in a non-destructive manner and achieve superior image quality. We provide examples showing that the image quality generated by our rapid surface-histology method approaches that of gold-standard archival histology.

Breast cancer is the most common cancer worldwide. Mammography screening and biopsy procedures for abnormal mammograms are the gold standard detection method, however, it is invasive, time-consuming, and labor-intensive. RNA biomarkers in circulating blood may be an alternative to the gold standard. We have adopted a multi-modal approach using Fourier transform infrared (FTIR) spectroscopy and Raman spectroscopy that provides a complete characterization of RNA biomarker fingerprints in the full spectrum. We have measured 99 patients’ serum samples and achieved a higher accuracy, specificity, and sensitivity using the multi-modal approach combined with machine learning analysis than using the individual techniques alone.

Multi-photon excited intensity and lifetime fluorescence images relying on endogenous contrast can be analyzed to quantify contributions from key metabolic co-enzymes and associated metabolic function and mitochondrial organization metrics. The high spatio-temporal resolution and context of these non-destructive measurements can be used to provide important insights related to a wide range of samples, conditions and disease models. Corresponding images are acquired from mitochondria, engineered tissues, excised and in vivo human tissues. Recent studies highlight the value of multi-parametric, label-free, metabolic assessments to improve our understanding of traumatic brain injury, (pre)cancer development, and vitiligo lesions.

We have recently developed a novel, cost-effective, versatile, and practical frequency-domain (FD) FLIM implementation capable of simultaneous multi-wavelength excitation, simultaneous multispectral detection, and sub-nanosecond to nanosecond fluorescence lifetime estimation. This novel technology has been adopted to implement an FD-FLIM endoscopy system for label-free metabolic imaging of epithelial tissues. The FD-FLIM endoscopy system is capable of simultaneous excitation at two wavelengths (375-nm for NADH, 445-nm for FAD). Preliminary results indicate that established metabolic autofluorescence biomarkers of epithelial pre-malignancy and malignancy can be clinically imaged with this novel FD-FLIM endoscopy technology. Future work will focus on exploring clinical applications of label-free metabolic imaging using this novel FD-FLIM endoscopy technology.

Multi-photon microscopy was used to determine the differential response to radiation at multiple time points between radiation-resistant and -susceptible human head and neck cancer tumor xenografts grown in a cohort of 60 athymic mice. Mice were randomly divided into groups for cell line, treatment or control, and time points relative to treatment — baseline, 1 hour-, 24 hours-, and 48 hours-post-treatment. We quantified the optical redox ratio, lifetime of NAD(P)H, and heterogeneity in ORR and lifetime endpoints. We saw a significant effect of cell line, treatment, and the interaction of all three factors — line, treatment, and time point.

Frequency Domain Diffuse Optical Spectroscopy (FD-DOS) can measure tissue optical properties and Diffuse Correlation Spectroscopy (DCS) can measure tissue blood flow. We have combined a custom FD-DOS and a custom DCS system to simultaneously measure oxy and deoxy hemoglobin, blood flow, and muscle metabolism of the sternocleidomastoid during inspiratory threshold loading in healthy volunteers. Additionally, we have developed Monte Carlo based multilayer look up tables for both FD-DOS and DCS that take into account both the skin tone and adipose tissue thickness. In the future, this system may assist in monitoring patients under mechanical ventilation, especially during weaning.

This conference presentation was prepared for the Multiscale Imaging and Spectroscopy IV conference at SPIE BiOS, SPIE Photonics West 2023.

Phytoplankton community constitutes the basis of aquatic food chain and rapidly change in response to ecological factors. Tracking health and diversity of phytoplankton is crucial in monitoring nutrient availability and environmental impacts in aquaculture industry. Herein we use NIR Raman spectroscopy as highly specific, label free method for monitoring phytoplankton. The differences in the composition of complex mixtures of bioactive molecules were analyzed using multivariate statistical data analysis techniques (PCA, PARAFAC and PLS-DA). Our results demonstrate high cross-validation and prediction accuracy for different growth phases, taxonomic groups and cell viability in three different species of phytoplankton.

Speckle contrast optical spectroscopy (SCOS) allows for simultaneous monitoring of blood flow and volume changes within each cardiac pulse. SCOS measurements were collected from 13 subjects and blood flow and volume pulse waveforms (PWFs) were extracted. The correlations between features extracted from the PWFs and blood pressure before and after an exercise activity were investigated. We found that the time delay between the blood flow and volume peaks was strongly correlated with changes in blood pressure (R = -0.73), suggesting that the combination of blood flow and volume information may improve blood pressure estimation.

Tissue-mimicking phantoms are vital for the calibration of imaging systems. Blood oxygenation is an important physiological parameter that spectral imaging devices can measure, motivating the need for blood phantoms with variable oxygenation. We present a soft lithography method for fabricating blood flow phantoms that enable control of channel size and blood oxygenation. Channel size was varied during the fabrication process. Oxygenation was varied by chemical oxygenation and deoxygenation of horse blood flowing in the phantom. The spectral properties of the phantoms were evaluated using a narrowband nailfold capillaroscopy system. The resulting phantoms yield a flexible approach for spectral imaging calibration.

Data from our recent clinical trial of patients with adnexal/ovarian lesions revealed that total hemoglobin concentration (HbT) and blood oxygen saturation (%sO2) obtained from photoacoustic imaging (PAI) are important predictors of malignancy. A model utilizing the co-registered US and PAI HbT and %sO2 has achieved superior performance with the area under the receiver operating curve of 0.97 (95% CI: 0.932-1).

Optical coherence microscopy (OCM) facilitates imaging of biological processes in living cells with no need for fluorescence labeling and at power levels that do not harm the cells. In this paper, we present a compact full-field OCM system coupled with a commercial fluorescence microscope that allows for label-free imaging of living cells with high lateral and axial resolution. The dynamic imaging allows for the internal cell structures to be characterized based on their light scattering potential and motion dynamics. Additionally, we demonstrate that the internal motion of the cytoplasm effectively reduces the speckle noise resulting in high-contrast images.

In Spatial Frequency Domain Imaging (SFDI) a sinusoidal intensity pattern is projected onto tissue, and the reflectance is a function of the projected spatial frequency and the tissue optical properties. For low spatial frequencies measurements are in the diffuse regime and a model by Cuccia et al. is currently used for this regime. For high spatial frequencies measurements are in the subdiffuse regime and no model exists. Here, we derive a comprehensive model that describes the reflectance as a function of tissue optical properties and spatial frequencies from the diffuse to the subdiffuse regime.

Peritoneal metastases are characterized by significant disruptions in the extracellular matrix. Hence, the scattering cross sections of malignant and benign lesions and surrounding background tissues are distinct. In this work, Monte Carlo based regression was used to develop an empirical relation to extract the scattering power of tissue based on co- and cross-polarized RGB reflectance images of tissue. The empirical relation improved the sensitivity of lesion detection, and discrimination accuracy of malignant and benign lesions. The proposed empirical equation is both accurate and simple, paving the way for real-time diagnostic applications.

Women’s health problems ranging from pregnancy complications to cancer do not have sufficient physiologic understanding nor clinical interventions, resulting in poor outcomes. Light-based technologies can be specifically designed to gain new insights into women’s health. They can be used at the point of care, be non-contact and re-usable, provide objective and real-time results, generate label-free or exogenous molecular contrast, and have significantly lower costs than most non-optical modalities. Successful application of optical technologies for furthering our understanding of cervical change during pregnancy and improving breast cancer surgery outcomes and access will be presented.

We built a fibre optic illuminated deep ultraviolet (DUV) fluorescence microscope. Our microscope was designed to be compact and portable, sitting on a 30cm by 30cm breadboard. Multimode fibres were magnetically butt coupled to 2 DUV light emitting diodes, which provided oblique epi-illumination of biological samples with short working distance objectives. The stage was motorised by stepper motors and controlled by a microcontroller which allowed for high-resolution imaging of large, fresh, and unsectioned tissue with high-throughput. We demonstrated the fluorescence microscope across a wide range of biological samples from single cells and spheroids to large tumours and organ samples.

We have developed a nanoscale strurctue measment system based on a digital micromirror device (DMD) based structured illumination microscopy (SIM) capable of wide-range and high-resolution imaging. The DMD based SIM yields the widefield images with the 20 different illumination patterns,a laser source with 349 nm wavelgnth, and a high sensitivity sCMOS camera the UV range. The developed image acquistion and reconstruct process can enhnace the imaging resolution under 150 nm and maximize the imaging area. In order to find practical applications of the SIM system, we prepared metallic nano pattern samples fabricated using FIB milling process and PET film samples fabricated using nano imprinting process that sample has nanohole structure with a diameter of 150 nm. The DMD based SIM system can reconstruct the nanostructure image with an large area and the obtained SIM images were compared with the SEM result to verify the improved resolution.

Structured Illumination microscopy is a super-resolution imaging technique based on sample fluorescence excitation with a spatially modulated light pattern. The pattern properties as well as the capability to shift it over sample determine the quality of the final images. At the current state of the art, pattern generation and translation require bulky and non-trivial optical setups. Here we propose an integrated optical device for the versatile generation and translation of the light pattern. This device can be used as light source for a standard microscope, upgrading it to a super-resolution system.

Non-invasive imaging with high resolution deep within biological materials without the use of harmful ionizing radiation is of great interest in the field of medical imaging. Second harmonic generation is an excellent mechanism to circumvent this issue by providing outstanding contrast and optical sectioning. In general, these signals are weak and prone to scattering which introduce great challenges when imaging deep within turbid media. We will discuss recently demonstrated Epi-SHG holography, which can detect very weak backscattered SHG signals and enables the application of recently developed techniques which utilize the phase information to allow diffraction limited imaging within deep tissue.