We present a preclinical study that demonstrates that deep photodynamic therapy using a minimally-invasive wireless light-emitting implant with orally-administered 5-ALA can reduce tumor volume in an aggressive triple-negative model of breast cancer.

Frequency-domain near-infrared spectroscopy (FD-NIRS) is a noninvasive in vivo sensing and imaging technique that is used to quantify tissue composition and oxygen metabolism in the brain, muscle, and other tissues. However, the size and complexity of FD-NIRS instrumentation have largely limited its use to laboratory and clinical research settings. To expand the use of FD-NIRS into continuous monitoring applications and in naturalistic environments, we report a novel real-time multi-frequency (50-350 MHz) wearable FD-NIRS system based on a custom application-specific integrated circuit (ASIC). The wearable device includes 685 nm and 850 nm laser diodes and a silicon-photomultiplier (SiPM) detector. The system has an optical property accuracy of 0.0007/mm^-1 for absorption and 0.08/mm^-1 for reduced scattering at 14.7 Hz measurement rate, evaluated using tissue-simulating phantoms, and is capable of capturing single frequency measurements up to 1 kHz with both wavelengths. Without a battery, it weighs 37 grams and measures less than 7 x 3 x 3 cm in size. For proof-of-concept, we demonstrate measurement of an arteriovenous occlusion of the human forearm. Overall, this work demonstrates the feasibility of quantitative FD-NIRS tissue optical spectroscopy for wearable health monitoring.

Quantitative diffuse optical spectroscopy (DOS) is used in clinical research to non-invasively characterize deep tissue composition and metabolism by measuring its near-infrared optical properties. However, it has been observed that absorption from epidermal melanin can reduce signal strength and decrease the accuracy of absolute optical property measurements. With the ultimate goal of developing techniques to correct for these effects, we first needed a reproducible phantom system that mimics epidermal melanin absorption. Here we report the fabrication and evaluation of skin phantoms based upon widely available polyurethane adhesive dressing films with tunable near-infrared absorption introduced by desktop laser printing. These easy-to-fabricate films reproduced physiological melanin absorption in the range of 600 to 1000 nm and exhibited good reproducibility (<8% inter-sample variation over 20 replicates). We applied films of varying opacity to homogenous tissue-simulating phantoms to demonstrate their effect on absolute optical properties measurements by frequency-domain near-infrared spectroscopy (FD-NIRS).

The use of noninvasive near-infrared spectroscopy (NIRS) has grown significantly over the past few decades for use in the characterization of tissue optical properties and derived physiological parameters such as tissue oxygen saturation and hemoglobin composition to support human health and wellness. Wearable NIRS devices have been developed in recent years that enable real-time continuous monitoring. These devices have been enabled by advances in the dense integration of miniature integrated circuits, light sources and detectors, embedded processing, and wireless technologies. The most attractive aspect of wearable NIRS devices is that they enable real-time sensing both inside and outside of laboratory and clinical settings. This review paper focuses on recent hardware developments related to fully integrated wearable NIRS devices, including continuous wave, time domain, and frequency domain NIRS techniques, and their enabling technologies. We also review their application in biomedical fields such as neuroscience, musculoskeletal physiology, and others. We provide our perspective on the future technology research opportunities and direction of the field.

Skin pigmentation reduces the signal-to-noise ratio of diffuse optical spectroscopy (DOS) measurements of tissue. This increases the error in optical property estimation of the underlying tissue, e.g. if a single-layer geometry model is used, leading to errors in the derived concentration of tissue absorbers. Recent efforts have modeled skin pigmentation with polyurethane-and silicone-based skin-mimicking phantoms, in order to observe the effect of skin pigmentation on the estimation of the optical properties of the target internal tissue. We employ previously-reported skin-mimicking, polyurethane-film phantoms of various opacity levels on a breast-tissue silicone phantom. We report on hybrid multi-frequency frequency-domain and broadband DOS measurements processed with either single-distance or multi-distance methods. The optical property estimations of the underlying breast-tissue phantom erroneously increased with skin phantom opacity using the single-distance method, while they retained +/-10% accuracy when obtained with the multi-distance method. This underlines the potential of the multi-distance method to process DOS measurements to accurately measure the optical properties of internal tissue across a population with various skin-pigmentation levels.

The goal of this prospective clinical study was to determine if quantitative broadband (600-1000nm) diffuse optical spectroscopy (DOS) can differentiate suspicious benign and malignant breast lesions and compare various optical biomarkers. To accomplish this goal, N=68 subjects were measured using a custom probe-based hybrid broadband continuous-wave and frequency domain diffuse optical spectroscopic imaging device. Lesion concentrations of oxyhemoglobin, deoxyhemoglobin, methemoglobin, water, lipid, fat, and collagen were estimated from tissue optical absorption. We found that absolute concentrations of collagen were significantly elevated (p=0.037) in malignant lesions versus normal tissue, while elevated methemoglobin approached significance (p=0.058). In addition, elevated water concentrations were a significant differentiator (p=0.039) of malignant versus benign lesions when applying Z-score normalization and a previously published model of malignancy probability.

The design and evaluation of an integrated circuit (IC) that provides gain and phase measurements for a broad modulation bandwidth frequency-domain near-infrared spectroscopy (fd-NIRS) system is described. This circuit includes a voltage-controlled oscillator that provides modulating frequencies from 50 MHz up to 1 GHz, as well as analog measurements of amplitude and phase. The phase detector is based on an N-path filter and allows for accurate phase measurement. The IC works with a laser diode and an avalanche photodiode to implement a complete fd-NIRS spectroscopy system. The IC-based system was used to characterize tissue-simulating phantoms with varying absorption and reduced scattering properties with modulation frequencies of 105-405 MHz with the upper frequency limited by the APD used in the system. Mean optical property accuracy was within 0.00064 mm-1 and 0.054 mm-1 for absorption and reduced scattering, respectively, compared to a reference system. These results show a promising path toward scalable and wearable optical sensing systems based on fd-NIRS.

The purpose of this pilot observational clinical study was to investigate early differences in cerebral oxygenation in newborn babies with prenatal opioid exposure (POE) compared to healthy controls. Three (3) term-born infants with a POE and twelve (12) age-matched controls were assessed for cerebral oxygenation in the frontal lobe using frequency-domain near-infrared spectroscopy (FD-NIRS) within 24 hours of birth. Infants with POE demonstrated significantly less frontal lobe cerebral oxygenation compared to age-matched controls. Incidentally, we found evidence of skin color bias among healthy controls. Our data shows that quantitative diffuse optical spectroscopy has the potential to provide bedside measurements of cerebral physiology in neonates with POE and warrants further investigation. However, corrections for the potential effects of skin color bias need to be explored.

Light-mediated healthcare has several very attractive features, including minimally invasive administration, reduced possibility of side effects, and potential for patient-specific treatment. However, the shallow penetration of light into skin and tissue is an inherent physical limitation that restricts the use of phototherapy to superficial sites of disease. To mitigate this major problem, researchers and clinicians are developing miniature light devices that can be implanted at strategic deep-tissue locations within a living subject. The work is leveraged by emerging new classes of light-absorbing drugs and light-activated drug delivery methods. This Perspective summarizes the major concepts and recent technical advances in this interdisciplinary subfield of advanced phototherapeutics. Fundamental factors are described, such as the wavelength dependence of light penetration and scattering through tissue, and the required power levels for the desired phototherapeutic effect. The implanted devices can be powered by an attached wire or otherwise powered wirelessly, and each approach has its own engineering challenges to produce desired performance features. A forward-looking conclusion section envisions future phototherapies that use implanted miniature light devices with the capacity to deliver phototherapy and monitor therapeutic response using sensors that provide immediate feedback and dose control.

Deep-Learning based Diffuse Optical Tomography enables real-time clinical breast imaging, but current models are scanning-pathway-specific. Here, a transformer architecture is proposed to encode arbitrary scanning pathways, enabling flexible real-time imaging with a single trained model. (c) 2025 The Author(s)

This study employs broadband Diffuse Optical Spectroscopy (DOS) to study the temporal haemodynamic changes in the breast during pregnancy. Findings reveal variations in tissue oxygenation and haemoglobin concentration, offering deeper insights into breast development.

We present a wirelessly powered implantable device that emits 573 nm visible light for deep tissue photodynamic therapy. This device can generate sufficient singlet oxygen to effectively eliminate cancer cells in vitro.

We explore the use of Monte-Carlo based optical modeling to predict light propagation, sensitivity, and optical interrogation volume (photon hitting density) of a miniature tissue-implantable optical sensor system. There are a limited number of Monte Carlo tools available that allow for the direct import of 3D models of complex optoelectronic systems. We therefore investigate the use of TracePro, a commercial Monte Carlo-based ray-tracing software package, to guide the design of a needle-injectable optical sensor designed for tumor monitoring. We first validated the use of TracePro to model light propagation in multiple scattering tissue by modeling simple infinite geometry systems and comparing light propagation to the known analytical diffusion approximation solutions. We also analyzed ray-tracing history (i.e., photon paths) and observed agreement to analytical models of the optical interrogation volume, varying by an average of 12% across source-detector separations ranging from 5 to 15mm. Finally, we describe how this approach was used to analyze and guide the design of the implantable optical sensor for tumor monitoring. Overall, TracePro provides a straightforward, easy-to-use, and accurate approach to importing and analyzing complex diffuse optical sensing systems.

Welcome to the 2024 Feature Issue on Diffuse Optical Spectroscopy: Technology and Applications in Biomedical Optics Express! This feature issue provides an exemplary sample of established and emerging DOS technologies as well as their biomedical applications through 27 contributed research papers and 1 invited review article. DOS researchers are inherently multidisciplinary, advancing topics spanning the basic theory of light-tissue interactions, computational modeling, technique and system development and preclinical and clinical applications. You will find this full range of topics represented in this feature issue.

We present a miniature dual-wavelength fd-NIRS device for wearable real-time physiological monitoring. Based on a custom CMOS circuit, wearable device demonstrates accurate phase and amplitude measurements for modulation frequencies up to 300 MHz.

Frequency-domain near-infrared spectroscopy (FD-NIRS) is a noninvasive method for quantitatively measuring optical absorption and scattering in tissue. This study introduces structured interrogation (SI) as an interference-based approach for implementing FD-NIRS in order to enhance optical property estimation in multilayered tissues and sensitivity to deeper layers. We find that, in the presence of realistic noise, SI accurately estimates properties and chromophore concentrations with less than a 5% error. Particularly noteworthy, the phase-only component of SI FD-NIRS can quantify both the optical absorption and reduced scattering in homogeneous tissues and shows a 20% improved sensitivity to absorption changes in deeper tissues compared to conventional methods. We show that this enhanced sensitivity is promising for improving the accuracy of functional brain monitoring in the cortex of an infant with less superficial contamination.

Breastfeeding provides widely recognized advantages for infant and maternal health. Unfortunately, many women experience trouble with breastfeeding. Nevertheless, few suitable imaging modalities are available to study human lactation and determine the possible causes of breastfeeding problems. In this study, we apply broadband, quantitative diffuse optical spectroscopy (DOS) for this purpose. We present a study of fourteen lactating and eight similarly aged, premenopausal, non-lactating women to investigate the feasibility of DOS to study the optical and physiological differences between 1) lactating and non-lactating breasts, 2) the areolar and non-areolar region within the breast, and 3) lactating breasts before and after milk extraction. Our study shows that i) the median total hemoglobin concentration [tHb] of the lactating breast is 51% higher than for the non-lactating breast. ii) the median [tHb] of the lactating breast is 37% higher in the areolar region compared to the non-areolar region. iii) lactating breasts exhibit a positive median difference of 8% in [tHb] after milk extraction. Our findings are consistent with the expected physiological changes that occur during the lactation period. Importantly, we show that DOS provides unique insight into breast tissue composition and physiology, serving as a foundation for future application of the technique in lactation research.

Frequency-domain near-infrared spectroscopy (fdNIRS) has been shown to be a promising tool for the diagnosis and monitoring of breast cancer treatment in point-of-care settings. However, current fdNIRS embodiments suffer from poor scalability and high complexity that has slowed their clinical translation. For the first time, we present a handheld, fully-wireless, multi-detector, multi-wavelength, fdNIRS system capable of real-time quantitative noninvasive measurements of optical properties and tissue chromophore concentrations at >10 kHz. High spatial resolution 2D topography images are displayed in real-time on a mobile platform with motion tracking. We characterize the system against prior generations, as well as in-vivo performance in human subjects.

Achieving negative surgical margins, defined as no tumor found on the edges of the resected tissue, during lumpectomy for breast cancer is critical for mitigating the risk of local recurrence. To identify nonpalpable tumors that cannot be felt, pre-operative placements of wire and wire-free localization devices are typically employed. Wire-free localization approaches have significant practical advantages over wired techniques. In this study, we introduce an innovative localization system comprising a light-emitting diode (LED)-based implantable device and handheld system. The device, which is needle injectable and wire free, utilizes multiple wirelessly powered LEDs to provide direct visual guidance for lumpectomy. Two distinct colors, red and blue, provide a clear indication of tissue depth: blue light is absorbed strongly in tissue, visible within a close range of <1 cm, while red light remains visible through several centimeters of tissue. The LEDs, integrated with an impedance-matching circuit and receiver coil, are encapsulated in biocompatible epoxy for injection with a 12 G needle. Our findings demonstrate that the implant exhibits clearly perceivable depth-dependent color changes and remains visible through >2 cm of ex vivo chicken breast and bovine muscle tissue using less than 4 W of transmitted power from a handheld antenna. These miniaturized needle-injectable localization devices show promise for improving surgical guidance of nonpalpable breast tumors.

This study develops reliable and reproducible imaging practices using a free-form raster scan based on fdNIRS. We propose a study and investigate the reproducibility of images, finding a minimum correlation coefficient of 0.789 between scans.