Double-integrating-sphere (DIS) measurement is a common method for characterizing phantom and tissue optical properties, but can display low accuracy. To investigate the sources of errors, a digital twin based on Monte Carlo simulations and sphere corrections was built. We compared simulation and measurement results of phantoms with known optical properties, and identified error sources. After minimizing these sources, the average errors were reduced to -1% and 2.44% in the absorption and reduced scattering coefficient estimates respectively, highlighting the potential to achieve high accuracy in optical property estimation using a relatively low-cost measurement approach.

Deep learning is a powerful tool for image analysis and medical applications. However, due to their intricate black-box nature, comprehending deep learning model predictions are often challenging. Oral cancer is globally prevalent, necessitating reliable AI algorithms for screening, especially for low-income regions. Interpretability is crucial for reliable AI. Visual explanation, generating attention maps highlighting decision-influencing regions, aids interpretability and also helps guide AI focus. Elevating AI reliability involves assessing decision confidence as well. Quantifying model output certainty helps identify uncertain cases, which need additional examination. Dataset quality is also pivotal for reliable AI development. Methods to evaluate and enhance the data and label/annotation quality will also be essential.

This study aims to develop a high-fidelity prediction model based on artificial neural networks to quantify changes in blood oxygen saturation of the internal jugular vein (IJV) (ΔSijvO2) from the pulsatile component of diffuse reflectance spectra measured non-invasively from the neck surface above the IJV. Training and testing data are generated using a surrogate model, which is millions of times faster than the original Monte Carlo simulations. We have investigated the model’s resilience to measurement noise, changes in surrounding tissue’s oxygen saturation, and fluctuations in IJV’s depth and size due to respiration. Results of validating the prediction model by simulated data have exhibited root mean square errors of less than 4%. Finally, validation of the prediction model on healthy subjects performing the Valsalva maneuver in vivo has demonstrated agreements between predicted results and expected physiological responses.

This study explores the development of a multimodal imaging system for disease assessment. Various experiments were conducted to evaluate performance in terms of power density, illumination uniformity, and fluorescence emission properties, comparing the handheld setup to the benchtop system. Test samples included phantom gels and oral cancer samples. Preliminary results indicate that the compact LED ring illuminator provided sufficient power for detectable emission signals and improved emission distribution due to sample scattering. The presentation also discusses solutions for achieving a more uniform illumination field and provides insights into imaging in oral epithelial neoplasia with the compact widefield system, along with considerations for translating from a benchtop test system to a compact handheld multimodal system.

The NIH RADx Tech initiative was designed to accelerate the development, scale-up, and deployment of innovative point-of-care (POC) and over the counter (OTC) technologies, with a goal of significantly expanding COVID-19 test accessibility and capacity. Over a period of three years, from its April 29, 2020 launch through April 2023, RADx Tech-supported lab, POC, and OTC platforms received 55 Food and Drug Administration (FDA) Emergency Use Authorizations (EUA) and increased U.S. testing capacity by more than 7.8 billion tests. RADx Tech helped drive a paradigm shift from predominantly laboratory-based COVID-19 testing in the early stages of the pandemic, to widely available home and point of care rapid tests. This talk describes key elements of the RADx process and ongoing activities to accelerate the development of more powerful, and accessible diagnostics, across a spectrum of diseases and conditions, that patients can use in home POC settings.

Cardiovascular disease is one of the leading causes of death in the United States and over 40% of the population lives with cardiovascular conditions including hypertension. We have developed a dynamic pulsatile phantom for PPG signal evaluation towards creating a non-invasive, wearable, point-of-care health device for monitoring cardiovascular health. This phantom has the capabilities to generate and control a pressurized pulsatile waveform ranging from 30mmHg to 200mmHg. The phantom manufacturing process allows control of the optical properties, interfaces and geometrical form factors which are known to change due to both skin tone and obesity to influence the captured PPG signature.

Recent articles reporting racial bias in pulse oximeter performance in clinical studies have highlighted the need for well-validated, objective approaches for assessing skin pigmentation. Highly effective objective measurements of skin color will require not only the use of well-validated melanometers, but also optimized methods for implementation. Therefore, we identify Best Practices for melanometer use that can be widely applied in pulse oximeter clinical studies. Additionally, given the variety of acquisition, processing approaches and metrics currently employed in melanometers, we discuss standardization of device performance and outputs as well as use of pigmented tissue mimicking phantoms for performance evaluation.

Objective: We examined the impact of skin color on transcutaneous bilirubin (TcB) measurements in vitro. Methods: Using layered neonatal skin mimicking phantoms with varying dermal bilirubin levels and epidermal melanosome volume fractions, the relationship between skin color and TcB precision is systematically investigated. TcB measurements were performed with a commercially available bilirubin meter (JM-105, Draeger Medical, Lubeck, Germany). Results: Epidermal melanosome volume fractions affected TcB measurements, leading to larger underestimations at higher melanosome volume fractions and bilirubin levels. In this in vitro setting, underestimations ranged from 30 to 131 µmol/L at a TcB value of 250 µmol/L. Conclusion: Skin pigmentation affects TcB accuracy, with greater underestimations observed in darker skin tones and higher bilirubin levels. Our results highlight the need for improved TcB meter design and cautious interpretation of TcB readings on newborns with dark skin.

Measurement accuracy and performance reliability of medical devices determines the quality of medical diagnoses and therapies. The National Institute of Standards and Technology has developed tissue-mimicking materials for the proficiency testing and validation of optical medical imaging techniques, including diffuse optical imaging, axial dimension imaging standards, and photoacoustic microscopy.

Point-of-care measurement of blood oxygen saturation is commonly performed using pulse oximetry-based methods, yet clinical evidence indicates that these devices may exhibit racial disparities in accuracy. We are working to develop tissue-mimicking phantoms for performance comparison and standardization of pulse oximeters. In this study, we have evaluated the use of standard silicone modified by reducing curing agent content to reduce material hardness to biologically relevant levels. A new silicone formulation was identified which provides low hardness without modification. Measurements of compliance – channel diameter as a function of fluid pressure – indicate that both materials have potential for use in pulse oximeter testing.

Understanding pigmentation’s effect on pulse oximetry is critical amid evidence that pulse oximetry is less accurate for patients with pigmented skin. Optical phantoms can help validate oximeters, but commercial phantoms do not vary pigmentation. We develop a resin-based 3D printing method that generates mechanically flexible phantoms with tunable optical properties and <100 µm diameter channels. Using a reflectance-mode Maxim 86171 pulse oximeter, we evaluate how photoplethysmogram waveforms change as phantom pigmentation increases, and test an algorithm for estimating pigmentation from waveforms alone. 3D-printed phantoms can provide a platform for testing pulse oximeter performance across the spectrum of human pigmentation.

Retinal images technology such as optical coherence tomography (OCT), OCT-angiography (OCTA) and fluorescein angiography (FA) are an important for diagnosing eye diseases. As retinal imaging technology continues to evolve, standard test methods and optical phantoms for demonstrating image quality and performance also need to be developed and advanced. We used mixture solution of polydimethylsiloxzne (PDMS) and different titanium dioxide (TiO2) to develop a more realistic retinal phantom. The retinal phantom consists of two multilayer membranes, a base plate with retinal curvature, and three microfluidic channels. In addition, we developed retinal phantom for fluorescence imaging using micro-fluorescence beads. We could obtain OCT, OCTA, and fluorescence images.

We propose a novel and accessible approach for fabricating thin phantoms with controllable absorption properties in terms of magnitude, spectral shape, and spatial distribution. The method involves a standard laser printer to print on thin polyurethane films emulating optical properties of biological tissue. We characterize the thin phantoms in terms of optical properties, thickness, microscopic structure, and reproducibility of the printing process. We argue that these thin phantoms hold potential for a wide range of biomedical applications and will discuss their potential application in transcutaneous bilirubinometry performance studies.

Porcine skin possesses many of the morphological and functional features of human skin and is considered an appropriate surrogate for human skin in a wide range of preclinical animal studies. Nevertheless, there are differences in structure that may affect optical measurements of the tissue health. It is, therefore, important to understand these differences in optical properties for translation of swine data and experiments to human clinical studies. Here, we compare and contrast in-vivo measurements of the optical properties of normal and burned human and porcine skin obtained using two commercial SFDI systems.

Assessments of cellular metabolic function non-invasively, at high spatiotemporal resolution can provide important insights regarding cell state and fate relevant for the understanding, diagnosis and monitoring of numerous diseases and treatments. Methods that rely on the fluorescence intensity and lifetime characteristics of NAD(P)H and FAD, two co-enzymes involved in a number of key metabolic pathways, are becoming important tools for this purpose. This presentation will highlight efforts by several members of the biomedical optics community to establish guidelines on reporting of associated measurements to facilitate broad adoption and comparison of results from different studies. Specifically, I will describe recommended means of reporting evaluations of the: a) NAD(P)H and FAD fluorescence intensity-based optical redox ratio, and b) the NAD(P)H fluorescence lifetime-associated metrics based on time-domain and frequency (phasor)-domain analysis.

Tumor spheroids represent valuable tools for investigating avascular solid tumors under controlled experimental conditions. In this work we investigated alterations in cellular metabolism, lipid storage, and their interactions with the extracellular matrix (ECM) using various human breast adenocarcinoma cell lines and a non-malignant breast epithelial cell line as spheroids. Three imaging methodologies, were used. The Phasor FLIM, of NADH was used to analyze metabolic signatures. Hyperspectral imaging was used to image signatures from two orthogonal solvatochromic probes: ACDAN for assessing intracellular polarity and Nile Red for detecting lipid composition. SHG images were collected through multiphoton excitation and analyzed via PLICS (Phasor analysis of Local Image Correlation Spectroscopy) to assess the cellular microenvironment and collagen remodeling within breast cancer spheroids. Notable differences emerged in NADH and lipid signatures across cell lines, spheroids, and time. Invading MDA-MB231 cells exhibited elevated hydration and lipid droplets, suggesting adaptation for invasion. These methods shed light on metastatic mechanisms and potential therapies.

Fluorescence lifetime imaging measures the time a fluorophore remains in the excited state before returning to the ground state. Fluorescence lifetime measurements provide environmental information about fluorophores. For the endogenous metabolic co-enzymes reduced nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FAD), the fluorescence lifetime is different for free and protein-bound molecules; thus, lifetime imaging provides metabolic information about cells in a label-free manner. However, fluorescence lifetime analysis via traditional decay fitting methods is time-consuming and requires expertise. Furthermore, direct relationships between lifetime metrics and cell metabolism or functions are difficult to define. Recent advances in the combination of machine learning and fluorescence lifetime imaging allow accelerated image analysis and cell phenotype identification. Guidelines for ensuring rigor and transference of machine learning models will be discussed in the context of the development and testing of machine learning models that identify metabolism pathway use of individual cells from autofluorescence lifetime images.

Autofluorescence FLIM provides unique insight into metabolic activities in living cells with minimal perturbation of the native cell state. However, quantitative analysis of FLIM images can introduce variability in cell parameters that obscure the molecular changes in the cell. This talk will discuss how phasor and fit analysis of autofluorescence FLIM images can provide complementary information that maintains quality, tracks reproducibility, and improves confidence in quantitative FLIM parameters.

Research and development of optical phantoms is a growing and popular field, but it suffers from the same inequities that exist in our overall healthcare system- the lack of research into optical phantoms that represent different skin tones, genders, and body types. We are developing a library of multilayered liquid and solid phantoms with vastly different light absorbing and light scattering properties at differing wavelengths, focusing particularly on the optical properties of the uterine tissue. Our multilayer phantoms mimicking the pelvic region where the uterus lies also incorporate various properties that replicate different thicknesses of adipose tissue and skin pigmentation.

We provide an overview of NIST efforts related to the development of optical medical imaging standards. Specifically, we introduce the planar NIST-certified standard reference material (SRM) appropriate for calibrating depth-resolving 3D optical systems such as optical coherent tomography. We also present the development of an optical measurement facility for the characterization of the optical properties of materials used in such standards such as an integrating sphere system for broadband measurements of the absorption coefficients and the reduced scattering coefficients of potential phantom materials. Lastly, we report on our progress toward the development of a realistic, validated 3D retinal eye phantom.

Hyperspectral dark-field microscopy (HSDFM) and analysis algorithms demonstrate classification of various tissue types, including carcinoma in human post-lumpectomy breast tissues. Performance of the HSDFM is evaluated by comparing the backscattering intensity spectra of polystyrene microbead solutions with Monte Carlo simulations of the experimental data. For classification algorithms, two approaches, a supervised spectral angle mapper (SAM) algorithm and an unsupervised K-means algorithm, are used. The manually extracted endmembers of known tissue types were determined by the histopathology reading of the hematoxylin and eosin (H&E)-stained slides. Their associated threshold spectral correlation angles from the SAM algorithm for supervised classification make a good reference library that validates endmembers from the unsupervised algorithm. For unsupervised classification, a K-means algorithm, with no a priori information, produces abundance maps with dominant endmembers of various tissue types, including carcinoma subtypes of invasive ductal carcinoma and invasive mucinous carcinoma. The endmembers extracted by the two methods agree with each other within less than 2% residual