Organ protection and physiology optimization are important goals when taking care of anesthetized patients undergoing surgery. Postoperative cognitive dysfunction and perioperative stroke are unwarranted potential outcomes. Neurovascular coupling, the match between cerebral metabolic demand and substrate supply, should be regarded as the essential cerebral physiology which needs to be monitored during surgery. The brain-targeting near-infrared spectroscopy (NIRS) technology has the potential to fulfill this goal. Proposition of why and how to monitor essential cerebral physiology via advanced NIRS technologies is discussed. We also discussed the limits of the current NIRS technologies which merely measure cerebral tissue oxygen saturation in pooled cerebral arterial, capillary, and venous blood.

The use of near-infrared time-resolved spectroscopy (TRS-20, Hamamatsu Corporation) in two resistance type exercise applications in human subjects is described. First, using isometric flexion of the biceps, we compared the magnitude and relevance of tissue hemoglobin concentration and oxygen saturation (stO2) changes when assuming constant scattering versus continuous measurement of reduced scattering coefficients at three wavelengths. It was found that the assumption of constant scattering resulted in significant errors in hemoglobin concentration assessment during sustained isometric contractions. Secondly, we tested the effect of blood flow restriction (BFR) on oxygenation in a muscle (vastus medialis oblique, VMO) and in the prefrontal cortex (PFC) of the brain. The BFR training technique resulted in considerably more fatigability in subjects, and correlated with reduced muscle stO2 between sets of exertion. Additionally, exercise with BFR resulted in greater PFC deoxygenation than a condition with equivalent work performance but no BFR. These experiments demonstrate novel applications for diffuse optical spectroscopy in strength testing and targeted muscle rehabilitation.

According to American Academy of Orofacial Pain, 75% of the U.S. population experiences painful symptoms of temporo-mandibular joint and muscle disorder (TMJMD) during their lifetime. Thus, objective assessment of pain is crucial for efficient pain management. We used near infrared spectroscopy (NIRS) as a tool to explore hemodynamic responses in the frontal cortex to noxious thermal stimulation of temporomadibular joint (TMJ). NIRS experiments were performed on 9 healthy volunteers under both low pain stimulation (LPS) and high pain stimulation (HPS), using a temperature-controlled thermal stimulator. To induce thermal pain, a 16X16 mm² thermode was strapped onto the right TMJ of each subject. Initially, subjects were asked to rate perceived pain on a scale of 0 to 10 for the temperatures from 41°C to 47°C. For the NIRS measurement, two magnitudes of temperatures, one rated as 3 and another rated as 7, were chosen as LPS and HPS, respectively. By analyzing the temporal profiles of changes in oxy-hemoglobin concentration (HbO) using cluster-based statistical tests, we were able to identify several regions of interest (ROI), (e.g., secondary somatosensory cortex and prefrontal cortex), where significant differences (p<0.05) between HbO responses to LPS and HPS are shown. In order to classify these two levels of pain, a neural-network-based classification algorithm was used. With leave-one-out cross validation from 9 subjects, the two levels of pain were identified with 100% mean sensitivity, 98% mean specificity and 99% mean accuracy to high pain. From the receiver operating characteristics curve, 0.99 mean area under curve was observed.

We report time-domain applications of a new hemodynamic model by Fantini [1] that yields analytic expressions for signals that are measurable with hemodynamic-based neuroimaging techniques such as functional near-infrared spectroscopy (fNIRS) and functional magnetic resonance imaging (fMRI). We show how the model can be used to predict the perturbations in cerebral blood volume (CBV), blood flow (CBF), and metabolic rate of oxygen (CMRO2) that account for the initial dip and post-stimulus undershoot that have been reported in the fMRI and fNIRS literature. Furthermore, we have used data from the literature to perform a comparison between measured fNIRS and fMRI signals and the corresponding signals predicted by the new hemodynamic model. Results showed an excellent agreement between the model predictions and the reported measured data.

We have developed the second generation of our time-domain near-infrared spectroscopy (TD-NIRS) system for baseline and functional brain imaging. The instrument uses a pulsed broadband supercontinuum laser emitting a large spectrum between 650 and 1700 nm, and a gated detection based on an intensified CCD camera. The source laser beam is split into two arms, below and above 776 nm. In each arm, a fast motorized filter wheel enables selection of a bandpass filter at the required wavelength. Each filtered laser beam is then launched into one array of source fibers. The multiplexing through the array of fibers is implemented through a very compact home-made design consisting of two galvanometer mirrors followed by an achromatic doublet. Source fibers are then recombined one-by-one from both arms into the source optodes to be positioned on the head. The detection fibers are all imaged in parallel through a relay lens on an intensified CCD camera. By using detection fibers of different lengths, we introduce optical delays that enable simultaneous recording in different delay windows of the temporal point spread functions. We present the instrumentation and show its preliminary functional imaging capabilities. We also introduce a new probe where we use different fiber lengths on the source and the detector sides in order to record simultaneously both wavelengths from one location through different sets of fibers.

We present a dynamic contrast-enhanced near-infrared (DCE-NIR) technique that is capable of non-invasive quantification of cerebral hemodynamics in adults. The challenge of removing extracerebral contamination is overcome through the use of multi-distance time-resolved DCE-NIR combined with the kinetic deconvolution optical reconstruction (KDOR) analytical method. As proof-of-principle, cerebral blood flow, cerebral blood volume and mean transit time recovered with DCE-NIR are compared with CT perfusion values in an adult pig during normocapnia, hypocapnia, and ischemia. Measurements of blood flow acquired with DCE-NIR were compared against concomitant measurements using CT Perfusion.

We present a near-infrared spectroscopy (NIRS) approach for the optical characterization of two-layered tissuemimicking phantoms. For the data acquisition, we employed a multi-distance frequency-domain system. For the data analysis, we implemented an inversion routine based on a two-layered solution of the frequency-domain diffusion equation as the forward model. Measured quantities were the absorption and reduced scattering coefficients of the first layer (μ_a1, μ’_s1) and the second layer (μ_a2, μ’_s2), and the thickness of the first layer (L). We report measurements on three two-layered liquid phantoms featuring absorption coefficients in the range 0.009-0.017 mm^-1, reduced scattering coefficients in the range 0.69-0.92 mm^-1, and first layer thickness in the range 8-15 mm. Our method yielded measured values of the optical coefficients and first layer thickness (μ_a1, μ’_s1, μ_a2, μ’_s2, and L) that are within 10% of the true values (optical properties measured in the infinite geometry; and the true first layer thickness). These are promising results toward exploring the potential of this two-layered medium approach in the human head, where the two layers would represent extracerebral and cerebral tissue, respectively.

We present frequency-resolved measurements of the amplitude and phase of cerebral hemodynamic oscillations associated with paced breathing and measured with near-infrared spectroscopy on the forehead of a human subject. We have performed measurements at eleven paced breathing frequencies in the range 0.07-0.25 Hz. The resulting frequency spectra of the amplitude and phase of the hemodynamic oscillations were fit with a recently developed hemodynamic model,¹ that was able to reproduce the measured spectra and to determine a number of associated physiological parameters. The parameters measured include the blood transit times in the capillary (0.87 s) and venous (1.0 s) compartments, and the high-pass cutoff frequency of cerebral autoregulation (0.087 Hz). We also illustrate the implications of the new hemodynamic model on the basis of a phasor representation of physiological and hemodynamic oscillations. The combination of frequency-resolved measurements of cerebral hemodynamics and the phasor-based, frequency-domain solution of the new hemodynamic model provides a novel tool for cerebral hemodynamic studies with a potential for assessing a variety of physiological, functional, and pathological conditions.

Near-infrared optical measurements have been shown to offer a promising non-invasive way for monitoring breast neoadjuvant chemotherapy (NAC) and predicting outcome. In this study, we extend optical measurements to capture additional hemodynamic and metabolic biomarkers revealed by dynamically imaging breast tissue during fractional mammographic compression. We are obtaining pre-treatment, day 7 and optional monthly scans in breast cancer patients undergoing NAC. The difference in hemodynamic response to compression between healthy and tumor-bearing breast decreases over the course of neoadjuvant therapy in responders compared to nearly no change in patients not responding to the chemotherapy.

Breast cancer patients often undergo neoadjuvant chemotherapy to reduce the size of the tumor before surgery. Tumors which demonstrate a pathologic complete response associate with improved disease-free survival; however, as low as 10% of patients may achieve this status. The goal is to predict response to anti-cancer therapy early, so as to develop personalized treatments and optimize the patient’s results. Previous studies have shown that tumor response can be predicted within a few days of treatment initiation. We have developed a diffuse optical tomography (DOT) imaging system for monitoring the response of breast cancer patients to neoadjuvant chemotherapy. Our breast imaging system is a continuous wave system that uses four wavelengths in the near-infrared spectrum (765 nm, 808 nm, 827 nm, and 905 nm). Both breasts are imaged simultaneously with a total of 64 sources and 128 detectors. Three dimensional reconstructions for oxy-hemoglobin concentration ([HbO2]), deoxy-hemoglobin ([Hb]) concentrations, and water are performed using a PDE-constrained multispectral imaging method that uses the diffusion approximation as a model for light propagation. Each patient receives twelve weekly treatments of Taxane followed by four cycles of Doxorubicin and Cyclophosphamide (AC) given every other week. There are six DOT imaging time points: baseline, week 3 and 5 of Paclitaxel, before cycle 1 and 2 of AC, and before surgery. Preliminary results show that there is statistical significance for the percent change of [HbO2], [Hb], [HbT], and percent water at week 2 from the baseline between patients with a pathologic response to chemotherapy.

We present diffuse optical mammography images that capitalize on the large optical absorption contrast (two orders of magnitude) between blood vessels and breast tissue, thus displaying breast vasculature. We have found a good correspondence between blood vessels displayed in the optical mammogram and those visible in the x-ray mammogram from the same subject in the same view (craniocaudal). By using broadband spectral information (wavelength range: 650-900 nm), we measured the hemoglobin saturation for the blood vessels displayed in the optical mammogram, for which we obtained an average value of 70%. In order to assess the z-axis depth of optical inhomogneities detected in this planar scanning approach, we have introduced pairs of detector optical fibers that are off-axis with respect to the illumination optical fiber. The spatial shift between the locations of the optical inhomogeneities in the two off-axis images can be translated into a depth measurement through a quasi-universal depth-shift reference curve. We report Monte Carlo simulations that show how this depth-shift reference curve is affected by the distance between the source and detector scanning planes (sample thickness) and to the specific arrangement of the two off-axis detectors. Liquid tissue-like phantoms were used to test this depth assessment approach for an absorbing rod placed at a depth of 33.6 mm. The depth measured with our method over the rod length ranged within 33-35 mm, in excellent agreement with the actual rod depth. The ability to identify blood vessels, measure their hemoglobin saturation, and assess their depth within breast tissue is a valuable feature that can advance optical mammography by providing additional structural and functional information about breast tissue.

A novel Gen-2 hand-held optical imager was developed with capabilities to contour to different tissue curvatures, perform simultaneous illumination and detection and imager large tissue surfaces. Experimental studies using cubical phantoms demonstrated that the imager can detect targets up to 2.5 cm and 5 cm deep via reflectance and transmission measurements, respectively. The target was also localized as regions of high absorption during multi-scan imaging of curved breast phantoms via both reflectance and transmission modes. Preliminary in-vivo breast imaging demonstrated that the target can be detected via varying the pressure applied during imaging, as observed from reflectance-based imaging studies on healthy adults with superficially placed target(s) in the intra-mammary fold.

Multi-diameter single fiber reflectance (MDSFR) spectroscopy enables quantitative measurement of tissue optical properties, including the reduced scattering coefficient and the phase function parameter γ. However, the accuracy and speed of the procedure are currently limited by the need for co-localized measurements using multiple fiber optic probes with different fiber diameters. We here demonstrate the use of a coherent fiber bundle acting as a single fiber with a variable diameter for the purposes of MDSFR spectroscopy. We have previously shown that the spectral reflectance and effective path lengths measured by the fiber bundles are equivalent to those measured by single solid-core fibers for fiber diameters between 0.4 and 1.0 mm (r ≥ 0.997). Based on these results, the coherent fiber bundle was deemed suitable for use as a variable-diameter fiber in clinical MDSFR quantification of tissue optical properties. In this paper we discuss the details of the setup, the limitations of the current experimental design as well as future improvements of the setup.

Current methodologies for obtaining depth-sensitive contrast information from an optically diffusive medium involve complex hardware and software implementations. In turn, such methods typically lead to long acquisition and long reconstruction times, rendering them impractical for real-time use. In this work, we report preliminary proof-of-concept for a hardware-only method capable of providing depth sensitive contrast information without requiring post acquisition image reconstruction and with rapid acquisition. This method, termed Masked Detection of Structured Illumination (MDSI), relies on physically masking, in the detection arm, the point spread function from a collimated beam illuminating a diffusive medium, to isolate the contribution of the photon path lengths of interest. By continuously scanning and integrating the obtained images, MDSI allows, for the first time, optical depth sectioning of a diffusive medium without any processing.

Infantile hemangiomas (IH) are common vascular growths that occur in 5-10% of neonates and have the potential to cause disfiguring and even life-threatening complications. With no objective tool to monitor IH, a handheld wireless device (HWD) that uses diffuse optical spectroscopy has been developed for use in assessment of IH by measurements in absolute oxygenated and deoxygenated hemoglobin concentration as well as scattering in tissue. Reconstructions of these variables can be computed using a multispectral evolution algorithm. We validated the new system by experimental studies using phantom experiments and a clinical study is under way to assess the utility of DOI for IH.

Information about the size and depth of a wound and how it is developing is an important prognostic tool in wound diagnostics. In this study a two-camera vision system has been developed to collect optical properties, shape and volume of chronic skin ulcers as tool for diagnostic assistance. This system combines the functionality of 2D imaging spectroscopy and 3D stereo-photogrammetry. A high resolution hyperspectral camera and a monochromatic video frame camera were mounted on the same scanning system. Stereo images were acquired to obtain information about the wound surface geometry. A Digital Surface Model (DSM) of the wound surface was reconstructed by applying stereophotogrammetric methods. The hyperspectral image was co-registered to the monochromatic frame image and the wound border was extracted by applying spectroscopic analysis (e.g. tissue oxygenation, pigmentation, classification). The resulting DSM of the undamaged surroundings of the wound was used to reconstruct the top surface above the wound and thus the wound volume. The analyses can, if desired, be limited to a certain depth of interest like the wound bed or wound border. Simultaneous analysis of the hyperspectral data and the surface model gives a promising, new, non-invasive tool for characterization of chronic wounds. Future work will concentrate on implementation of real time analysis and improvement of the accuracy of the system.

The design and implementation of a diffuse optical tomography system using wavelength-swept laser is described. Rapid and continuous wavelength change is utilized for high speed spectral scanning from 775 nm to 875 nm optical wavelength. Maximum speed of wavelength repetition is 1 kHz and averaged output power of the wavelength-swept laser is 20 mW. A fiber-optic Sagnac interferometer is incorporated to conduct passive amplitude modulation of the wavelength-swept laser. It is shown that the wavelength-swept laser can be successfully incorporated to the DOT system, and then reduces wavelength-shifting time and hardware complexity in multi-wavelength DOT implementation.

Light from a super continuum source is dispersed into a number of wavelength bands, coded using a spatial light modulator and recombined into an optical fiber that delivers it to living tissue. Diffuse light emanating from the tissue is detected efficiently using an array of high numerical aperture, single-element photo detectors and spectrally resolved by demodulation. This method allows spectral changes in tissues to be imaged with high speed and sensitivity.

We have presented methodology to calibrate data in NIRS/MRI imaging versus an absolute reference phantom and results in both phantoms and healthy volunteers. This method directly calibrates data to a diffusion-based model, takes advantage of patient specific geometry from MRI prior information, and generates an initial guess without the need for a large data set. This method of calibration allows for more accurate quantification of total hemoglobin, oxygen saturation, water content, scattering, and lipid concentration as compared with other, slope-based methods. We found the main source of error in the method to be derived from incorrect assignment of reference phantom optical properties rather than initial guess in reconstruction. We also present examples of phantom and breast images from a combined frequency domain and continuous wave MRI-coupled NIRS system. We were able to recover phantom data within 10% of expected contrast and within 10% of the actual value using this method and compare these results with slope-based calibration methods. Finally, we were able to use this technique to calibrate and reconstruct images from healthy volunteers. Representative images are shown and discussion is provided for comparison with existing literature. These methods work towards fully combining the synergistic attributes of MRI and NIRS for in-vivo imaging of breast cancer. Complete software and hardware integration in dual modality instruments is especially important due to the complexity of the technology and success will contribute to complex anatomical and molecular prognostic information that can be readily obtained in clinical use.

A new optical parallel detection system of both frequency and continuous wave domains was developed to improve the data quality and accuracy in recovery of all breast optical properties. This new system combines frequency domain (FD) measurements using photomultiplier tubes and continuous wavelengths (CW) measurements using photodiode detectors in order to incorporate addition NIR wavelengths up to 948nm. The FD measurements use 6 wavelengths (660, 735, 785, 808, 826, and 849 nm) while the CW use three wavelengths (903, 912, and 948nm). The frequency domain part of the system is described in detail and steps taken to improve signal to noise ratio are discussed. Furthermore, different acquisition procedures were tested in order to reduce the duration of a complete 9 wavelength acquisition.

To overcome the intensive light scattering in the biological tissue, diffuse optical tomography (DOT) in the near infrared range for breast lesion detection usually is combined with other imaging modalities such as ultrasound, X-ray, and MRI, to provide guidance. However, the guided imaging modalities may depend on different contrast mechanics compared to the optical contrast in the DOT. As a result, they can’t provide reliable guidance for diffuse optical tomography because some lesions may not be detectable by a non-optical modality but yet have high optical contrast. An imaging modality which can provide the guidance from optical contrast is desirable for DOT. In this paper, we present a system that combines diffuse optical tomography and photoacoustic tomography (PAT), to detect and characterize the deeply-seated targets embedded in a turbid medium. Photoacoustic tomography utilizes a short-pulsed laser beam to penetrate into tissue diffusively. Upon absorption of the light by the target, photoacoustic waves are generated and used to reconstruct, at ultrasound resolution, the optical absorption distribution that reveals optical contrast. The combined system used in the experiment combines a 64-channel photoacoustic system with a frequency-domain diffused optical system. To further improve the contrast, the exogenous contrast agent, indocyanine green (ICG) is used. Our experiment results show that the combined system can detect a tumormimicking phantom up to 2.5 cm in depth and 10 μM in concentration. Mice experiments also confirmed that the combined system can detect the tumor region and monitor the ICG uptake and washout in the tumor region. This method can potentially improve the accuracy to detect the small breast lesions or any lesions which are sensitive to the reference change, such as the lesions located on the chest wall.

X-ray image pixel intensity and optical scattering are compared for 11 normal subjects to assess the feasibility of using X-ray imaging as a surrogate for optical scattering in NIR spectral tomography. Digital breast tomosynthesis exams, as well as twenty single point reflectance measurements of optical breast scattering are compared for a wide variety of breast sizes and densities to determine if scattering can be accurately predicted based on x-ray attenuation. If implemented, x-ray based scattering estimation will decrease exam time and cost as well as simplify the design of a newly developed integrated near infrared spectral tomography and digital breast tomosynthesis imaging system.

The light transport in biological media is studied theoretically and experimentally. We derived exact analytical solutions of the radiative transfer equation for the semi-infinite scattering medium. Arbitrary rotationally symmetric anisotropic scattering functions can be handled. The exact boundary conditions within the transport theory were implemented including Fresnel-reflection at mismatched interfaces. The analytical solutions of the reflectance from the scattering medium were verified with the Monte Carlo method showing an agreement within the statistics of the simulations. In addition, the analytical solutions were compared to solutions of the diffusion equation and to experiments, e.g. to spatially resolved reflectance measurements. Experiments were performed with optically well-characterized liquid tissue phantoms. The use of the analytical solutions and the Monte Carlo simulations was investigated for determining the optical properties of the scattering media.

Recently developed phase-function corrected diffusion theory is applied to the problem of computing Jacobian matrices in the transport regime. We propose additional approximations that lead to simplified transport-regime expressions for the Jacobian matrices that may be evaluated by Monte Carlo simulations, phase-function-corrected diffusion models, or recently developed analytical solutions to the radiative transport equation.

The normalized Born ratio of steady-state fluorescence measurements in a concave or convex-shaped medium geometry is investigated by analytical and numerical methods. The “concave” geometry refers to a scattering-dominant medium enclosed by a long circular cylindrical applicator, and the “convex” geometry refers to a scattering-dominant medium enclosing a long circular cylindrical applicator. The numerical investigation uses finite element- method, and the corresponding analytical evaluation is based upon a recently developed method of treating steady-state photon diffusion in both concave and convex geometries. The steady-state Born ratio associated with a source and a detector located on the medium-applicator interface is examined for the medium to have a homogeneous distribution of fluorophore, and for the source and detector aligning either azimuthally or longitudinally in both concave and convex geometries. At a given set of optical properties and the line-of-sight source-detector distance, the normalized Born ratio is always smaller in concave and greater in convex geometry respectively when comparing to that in semi-infinite geometry. At a given set of optical properties, the rate of increase of the normalized Born ratio versus the line-of-sight source-detector distance is the greatest along the azimuthal direction in convex geometry among the studied cases. The change of the normalized Born ratio caused by containing a target of either positive or negative contrast of fluorophore in the otherwise homogeneous background of fluorophore is also investigated numerically. The results for both homogenous and heterogeneous fluorophore distribution demonstrate that the normalized Born ratio is a geometry-specific parameter that imposes geometrically-specific sensitivity in measurements.

Based on light propagation theory, the measurements of a contact-free imaging system with generalized optical components can be obtained from a linear transformation of the light intensity distribution on the surface of the imaging object. In this work, we derived the linear measurement operator needed to perform this transformation. Numerical experiments were designed and conducted for validation.

Time-domain near-infrared spectroscopy (TD-NIRS) offers the ability to measure the absolute baseline optical properties of a tissue. Specifically, for brain imaging, the robust assessment of cerebral blood volume and oxygenation based on measurement of cerebral hemoglobin concentrations is essential for reliable cross-sectional and longitudinal studies. In adult heads, these baseline measurements are complicated by the presence of thick extra-cerebral tissue (scalp, skull, CSF). A simple semi-infinite homogeneous model of the head has proven to have limited use because of the large errors it introduces in the recovered brain absorption. Analytical solutions for layered media have shown improved performance on Monte-Carlo simulated data and layered phantom experiments, but their validity on real adult head data has never been demonstrated. With the advance of fast Monte Carlo approaches based on GPU computation, numerical methods to solve the radiative transfer equation become viable alternatives to analytical solutions of the diffusion equation. Monte Carlo approaches provide the additional advantage to be adaptable to any geometry, in particular more realistic head models. The goals of the present study were twofold: (1) to implement a fast and flexible Monte Carlo-based fitting routine to retrieve the brain optical properties; (2) to characterize the performances of this fitting method on realistic adult head data. We generated time-resolved data at various locations over the head, and fitted them with different models of light propagation: the homogeneous analytical model, and Monte Carlo simulations for three head models: a two-layer slab, the true subject’s anatomy, and that of a generic atlas head. We found that the homogeneous model introduced a median 20 to 25% error on the recovered brain absorption, with large variations over the range of true optical properties. The two-layer slab model only improved moderately the results over the homogeneous one. On the other hand, using a generic atlas head registered to the subject’s head surface decreased the error by a factor of 2. When the information is available, using the true subject anatomy offers the best performance.

Different strategies for Monte Carlo simulations are currently used in tissue optics. In this work we analyze and compare four Monte Carlo methods based on different ways to extract the photons’ trajectories. By using theoretical arguments we show that the four methods are statistically equivalent. Afterwards we study the convergence of the four methods both in time and continuous wave domains. Our results show that those Monte Carlo methods based on photons’ annihilation or survival converge faster for continuous wave calculations and at shorter source-detector distances. On the contrary Monte Carlo methods based on weight assignment provide a better representation of the temporal point spread function in time domain.

In this work we develop the discrete form of a correlation measurement density function for a normalised autocorrelation measurement type commonly encountered in ultrasound modulated optical tomography. These sensitivity functions are computed under the assumptions required by a correlation diffusion forward model which we implement by the finite-element method. We utilise these sensitivity functions in the linear reconstruction of an optical absorption perturbation.

We report the use of genetic algorithm-based approach for reconstruction of the optical inhomogeneities embedded in turbid medium using diffuse optical tomography. In the proposed inversion scheme, the task of image reconstruction is formulated as optimization (minimization) problem which is solved using genetic algorithm approach to find the global minimum of the objective function. This approach preserves the full non-linear features of the problem and it can be applied for quantitative reconstruction over a wide range of contrast values where conventional linear and higher order reconstruction approaches show severe limitations. Successful implementation of the proposed scheme has been demonstrated for quantitative reconstruction of absorbing inhomogeneities (single as well as double) embedded in an otherwise homogeneous medium. In order to demonstrate its potential, reconstruction results have been presented for a wide range of parameters including size, location and contrast of the absorbing inhomogeneities embedded in the turbid medium.

X-ray luminescence optical tomography (XLOT) is an emerging hybrid imaging modality in which x-ray excitable particles (phosphor particles) emit optical photons when stimulated with a collimated x-ray beam. XLOT combines the high sensitivity of optical imaging with the high spatial resolution of x-ray imaging. For XLOT reconstruction, we compared two reconstruction algorithms, conventional filtered backprojection (FBP) and a new algorithm, x-ray luminescence optical tomography with excitation priors (XLOT-EP), in which photon propagation is modeled with the diffusion equation and the x-ray beam positions are used as reconstruction priors. Numerical simulations based on dose calculations are used to validate the proposed XLOT imaging system and the reconstruction algorithms. Simulation results show that XLOT can better detect inclusions of particles than x-ray computed tomography (CT) alone. Nanoparticle concentrations reconstructed with XLOT-EP are much less dependent on target depths than those obtained with FBP. Measurements at just two orthogonal projections are sufficient for the XLOT-EP to reconstruct an XLOT image for simple source distributions. The heterogeneity of x-ray dose distribution is included in the XLOT-EP reconstruction and improves the reconstruction accuracy, suggesting that there is a need to calculate the x-ray dose distribution for experimental XLOT imaging. We have built a prototype XLOT imaging system, in which a collimated x-ray bean is used to scan the samples and an electron multiplying charge couple device (EMCCD) is used to collect the optical photons emitted by the x-ray excitable particles. Phantom experiments have been performed to validate the XLOT system and the proposed reconstruction algorithms.

A fundamental problem of separation of absorption and scattering contributions into light attenuation was resolved using time- and frequency-domain systems. In spite of the progress in the development of time- and frequency-domain hardware, continuous-wave systems offer higher signal-to-noise ratios, better portability, and, potentially, lower equipment and service costs. In biological tissues of known composition one can quantify the chromophore concentrations, such as oxy- and deoxyhemoglobin, water, and fat, and also quantify the reduced scattering coefficient by using broadband approach and by employing characteristic spectral features of chromophores. In this submission we report further developments of our hyperspectral technique to measure absolute concentrations of tissue chromophores.

Many studies have found that hypoxia, particularly cycling hypoxia (CH), can lead to enhanced tumor metastasis and resistance to radiation and chemotherapy. It was also reported that tumor total hemoglobin content (THb), which is directly related to tumor angiogenesis, can have significant impact on tumor’s response to radiation and neoadjuvant chemotherapy. There is a growing demand for technologies to measure tumor hypoxia and angiogenesis temporally in vivo. In this paper, a side-firing fiber optic sensor based on a multi-wavelength frequency-domain near infrared spectroscopy (FD-NIRS) instrument was used to quantify tumor oxygenation and hemoglobin concentrations in nude rats bearing human FaDu head and neck (H and N) tumors during normoxia and forced hyperoxia and cyclic hypoxia. Significant increase (with carbogen gas inhalation) or decrease (with reduced O₂ supply) in tumor oxygenation was observed. The studies demonstrated the feasibility of the technology for longitudinal monitoring of H and N tumor’s response to therapy.

Ovarian cancer, the deadliest of all gynecologic cancers, is not often found in its early stages due to few symptoms and no reliable screening test. Optical imaging has a great potential to improve the ovarian cancer detection and diagnosis. In this study we have characterized the bulk optical properties of 26 ex-vivo human ovaries using a Diffuse Optical Tomography system. The quantitative values indicated that, in the postmenopausal group, malignant ovaries showed significantly lower scattering coefficient than normal ones. The scattering parameter is largely related to the collagen content that has shown a strong correlation with the cancer development.

We are developing a method of monitoring treatment progression of interstitial photothermal therapy of focal prostate cancer using transrectal diffuse optical tomography (TRDOT) combined with transrectal 3D ultrasound (3D-TRUS). Measurements of prostate tissue optical properties were made on ex vivo human prostate samples prior to and post coagulation. Interstitial photothermal treatments were delivered to the ex vivo samples and monitored using an interstitial probe near the treatment fiber. After treatment, bulk optical properties were measured on native and coagulated zones of tissue. Changes in optical properties across the boundary between native and coagulated tissues were spatially mapped using a small diffuse reflectance probe. The optical property estimates and spatial information obtained using each method was compared.

We present a dynamic contact-free continuous-wave diffuse optical tomography system for the detection and monitoring of peripheral arterial disease (PAD) in the foot. Peripheral Arterial Disease (PAD) is the narrowing of the functional area of the artery generally due to atherosclerosis. It affects between 8-12 million people in the United States and if untreated this can lead to ulceration, gangrene and ultimately amputation. Contact-Free imaging is highly desirable, due to the presence of ulcerations and gangrene in many patients affected by PAD. The system uses an electron multiplying charge coupled device (EMCCD) camera for detection to achieve a dynamic range of 86 dB with a frame rate of 1 Hz using 20 collimated source fibers and 2 wavelengths. We present first clinical results showing 3D images of total hemoglobin changes in response to a dynamic thigh cuff.

Quantitative measurement oxygen consumption in the muscles is important to evaluate the effect of the exercise. Near-infrared spectroscopy (NIRS) is a noninvasive method for measuring muscle oxygenation. However, measurement results are affected by blood volume change due to changes in the blood pressure. In order to evaluate changes in blood volume and to improve measurement accuracy, we proposed a calculation method of three-wavelength measurement with considering the scattering factor and the measurement with monitoring blood flow for measuring the temporal change of the oxygen concentration more precisely. We applied three-wavelength light source (680nm, 808nm and 830nm) for the continued wave measurement. Two detectors (targeted detector and the reference detector) were placed near the target muscle and apart from it. We measured the blood flow by controlling the intravascular pressure and the oxygen consumption with the handgrip exercise in the forearm. The measured results show that the scattering factor contains the artifact at the surface and the blood flow in the artery and the vein in the same phase. The artifact and the blood flow in the same phase are reduced from the oxygenated and the deoxygenated hemoglobin densities. Thus our proposed method is effective for reducing the influence of the artifact and the blood flow in the same phase from the oxygen consumption measurement. Further, it is shown that the oxygen consumption is measured more accurately by subtracting the blood flow measured by the reference detector.

Interstitial near-infrared laser thermal therapy (LITT) is currently undergoing clinical trials as an alternative to watchful waiting or radical surgery in patients with low-risk focal prostate cancer. Currently, we use magnetic resonance image (MRI)-based thermography to monitor treatment delivery and determine indirectly the completeness of the target tissue destruction while avoiding damage to adjacent normal tissues, particularly the rectal wall. However, incomplete tumor destruction has occurred in a significant fraction of patients due to premature termination of treatment, since the photocoagulation zone is not directly observed. Hence, we are developing transrectal diffuse optical tomography (TRDOT), in combination with transrectal 3D ultrasound (3D-TRUS), to address his limitation. This is based on the large changes in optical scattering expected upon tissue coagulation. Here, we present forward simulations of a growing coagulated lesion with optical scattering contrast, using an established finite element analysis software platform (NIRFAST). The simulations were validated in tissue-simulating phantoms, with measurements acquired by a state-of-the-art continuous wave (CW) TRDOT system and a recently assembled bench-top CW-DOT system, with specific source-detector configurations. Two image reconstruction schemes were investigated and evaluated, specifically for the accurate delineation of the posterior boundary of the coagulation zone as the critical parameter for treatment guidance in this clinical application.

We investigate the feasibility of trans-rectal near infrared (NIR) based diffuse optical tomography (DOT) for early detection of prostate cancer using a transrectal ultrasound (TRUS) compatible imaging probe. For this purpose, we designed a TRUS-compatible, NIR-based image system (780nm), in which the photo diodes were placed on the trans-rectal probe. DC signals were recorded and used for estimating the absorption coefficient. We validated the system using laboratory phantoms. For further improvement, we also developed a hierarchical clustering method (HCM) to improve the accuracy of image reconstruction with limited prior information. We demonstrated the method using computer simulations laboratory phantom experiments.

Peripheral Arterial Disease (PAD) is the narrowing of the functional area of the artery generally due to atherosclerosis. It affects between 8-12 million people in the United States and if untreated this can lead to ulceration, gangrene and ultimately amputation. The current diagnostic method for PAD is the ankle-brachial index (ABI). The ABI is a ratio of the patient’s systolic blood pressure in the foot to that of the brachial artery in the arm, a ratio below 0.9 is indicative of affected vasculature. However, this method is ineffective in patients with calcified arteries (diabetic and end-stage renal failure patients), which falsely elevates the ABI recording resulting in a false negative reading. In this paper we present our results in a pilot study to deduce optical tomography’s ability to detect poor blood perfusion in the foot. We performed an IRB approved 30 patient study, where we imaged the feet of the enrolled patients during a five stage dynamic imaging sequence. The patients were split up into three groups: 10 healthy subjects, 10 PAD patients and 10 PAD patients with diabetes and they were imaged while applying a pressure cuff to their thigh. Differences in the magnitude of blood pooling in the foot and rate at which the blood pools in the foot are all indicative of arterial disease.

We apply the Fourier Transform to absorption and scattering coefficient images of proximal interphalangeal (PIP) joints and evaluate the performance of these coefficients as classifiers using receiver operator characteristic (ROC) curve analysis. We find 25 features that yield a Youden index over 0.7, 3 features that yield a Youden index over 0.8, and 1 feature that yields a Youden index over 0.9 (90.0% sensitivity and 100% specificity). In general, scattering coefficient images yield better one-dimensional classifiers compared to absorption coefficient images. Using features derived from scattering coefficient images we obtain an average Youden index of 0.58 ± 0.16, and an average Youden index of 0.45 ± 0.15 when using features from absorption coefficient images.

We have developed a method to quantify hemoglobin concentration and oxygen saturation within the renal cortex by near-infrared spectroscopy. A fiber optic probe was used to transmit the radiation of three semiconductor lasers at 690 nm, 800 nm and 830 nm to the tissue, and to collect diffusely remitted light at source-detector separations from 1 mm to 4 mm. To derive tissue hemoglobin concentration and oxygen saturation of hemoglobin the spatial dependence of the measured cw intensities was fitted by a Monte Carlo model. In this model the tissue was assumed to be homogeneous. The scaling factors between measured intensities and simulated photon flux were obtained by applying the same setup to a homogeneous semi-infinite phantom with known optical properties and by performing Monte Carlo simulations for this phantom. To accelerate the fit of the tissue optical properties a look-up table of the simulated reflected intensities was generated for the needed range of absorption and scattering coefficients. The intensities at the three wavelengths were fitted simultaneously using hemoglobin concentration, oxygen saturation, the reduced scattering coefficient at 800 nm and the scatter power coefficient as fit parameters. The method was employed to study the temporal changes of renal hemoglobin concentration and blood oxygenation on an anesthetized rat during a short period of renal ischemia induced by aortic occlusion and during subsequent reperfusion.

Prostate cancer diagnosis is based on PSA rate measurement and ultrasound guided biopsy. Recently criticized for its lack of specificity, new approaches are currently investigated: MRI, elastography, TEP, NIRS and Time Resolved (TR) fluorescence tomography. The advantage of TR fluorescence tomography relies on its good complementarity with the standard ultrasound protocol and on the possible localization of prostate tumors marked by specific probes. After a first TR system based on a bulky titanium-sapphire laser, we designed a new one taking advantage of a more compact white pulsed laser (supercontinuum). The improved compactness is now fully compatible with clinical environment. The light, filtered by two linear variable filters to select a 770±20 nm window, is driven to the transrectal probe which also collects the fluorescence light emitted by the marker. The signal is detected by photomultipliers connected to TCSPC boards. A reconstruction algorithm based on intensities and time of flight allows a fast localization of the fluorophore. We compared the performances of the new white laser system to the previous titanium-sapphire on prostate mimicking phantoms. The laser power delivered on the phantom by the new laser appeared to be suitable to fluorescence measurements, just below cutaneous maximum permitted exposure. The new system allowed us to localize fluorescent inclusions of a fluorescent nanoemulsion at fixed positions inside a prostate mimicking phantom.

The presence of metastatic tumor cells in tumor-draining lymph nodes is an important indicator for cancer staging and therapy. Current clinical approaches of assessing lymph node tumor burden require invasive surgery that can be associated with nerve damage and other complications. In this study, a dual-reporter fluorescence molecular imaging approach, previously validated for quantifying targeted reporter binding in various human tumor xenographs, was assessed as a means of quantifying tumor burden in metastatic disease in mice. The utility of the dual-reporter imaging approach to measure tumor burden in sentinel lymph nodes was investigated in a bioluminescent human breast cancer xenograph model in 18 female nude mice. Once the presence of tumor in the lymph node was confirmed by bioluminescent imaging, fluorescently labeled anti-EGFR antibody and an untargeted antibody (labeled with a different fluorophore) were injected intradermally, proximal to the lymph node, and the uptake of the two reporters was imaged simultaneously with a with a flat-panel fluorescent scanner. Preliminary results demonstrated a statistically significant correlation between the dual-reporter measured tumor burden and the bioluminescent measure of tumor burden.

A main problem with tomographic fluorescence recovery is that it can only reliably recover images of high contrast to background ratio, which is a problematic issue when the fluorescent contrast in a region of interest is near a significant source of background contrast, such as organs of filtration. A method is presented here, combining the resolution of structural image guidance with the benefits of using multiple fluorescent tracers, one targeted to the tumor of interest and one untargeted, in order to substantially improve the accuracy of recovered contrast values for targeted tracer concentration. Using the normalized subtraction in the data space, the recovery of lower contrast regions can be dramatically improved by suppressing the effect of larger perturbations which appear in both the targeted and untargeted fluorescence data sets. This methodology has significant potential value when imaging near excretory organs such as liver, lung, kidneys and bladder, depending upon the agent to be imaged.

Fluorescence molecular tomography (FMT) is quite demanding in terms of acquisition/computational times due to the huge amount of data. Different research groups have proposed compression approaches regarding both illumination (wide field structured light instead of raster point scanning) and detection (compression of the acquired images). The authors have previously proposed a fast FMT reconstruction method based on the combination of a multiple-view approach with a full compression scheme. This method had been successfully tested on a cylindrical phantom and is being generalized in this paper to samples of arbitrary shape. The devised procedure and algorithms have been tested on an ex-vivo mouse.

In this paper, we have synthesized a second generation tumor hypoxia targeted 2-nitroimidazole-ICG conjugate using piperazine linker (2-nitro-ICG-p) and validated its performance in in vivo tumor targeting. The results have shown that tumor hypoxia can be targeted with twice higher signal strength beyond three hours post-injection while the un-targeted ICG has completely washed out. The improvement of the second generation 2-nitro-ICG-p dyes is 1.2-1.3 times over the first generation 2-nitro-ICG dyes using ethanol linker beyond 3 hours post-injection which is the optimal time-window for evaluating tumor hypoxia.

Whole body in vivo optical imaging of small animals has widened its applications and increased the capabilities for preclinical researches. However, most commercial and prototype optical imaging systems are camera-based systems using epi- or trans- illumination mode, with limited views of small animals. And for more accurate tomographic image reconstruction, additional data and information of a target animal is necessary. To overcome these issues, researchers have suggested several approaches such as maximizing the detection area or using the information of other highresolution modalities such as CT, MRI or Ultrasound, or using multi-spectral signals. As one of ways to maximizing the detection area of a target animal, we present a new fluorescence and bioluminescence imaging system for small animals, which can image entire surface of a target animal simultaneously. This system uses double mirror reflection scheme and it consists of input unit, imaging unit with two conical mirrors, the source illumination part and the surface scanner, and the detection unit with an intensified CCD camera system. Two conical mirrors are configured that a larger size mirror captures a target animal surface, and a smaller size mirror projects this captured image onto a CCD camera with one acquisition. With this scheme, we could capture entire surface of a target animal simultaneously and improve back reflection issue between a mirror and an animal surface of a single conical mirror scheme. Additionally, we could increase accessibility to an animal for multi-modality integration by providing unobstructed space around a target animal.

The purpose of this study was to develop a dynamic contrast-enhanced (DCE) near-infrared spectroscopy (NIRS) technique to characterize tumor physiology. Dynamic data were acquired using two contrast agents of different molecular weights, indocyanine green (ICG) and IRDye 800CW carboxylate (IRD_cxb). The DCE curves were analyzed using a kinetic model capable of extracting estimates of tumor blood flow (F), capillary transit time (t_c) and the amount of dye that leaked into the extravascular space (EVS) – characterized by the extraction fraction (E). Data were acquired from five nude rats with tumor xenografts (>10mm) implanted in the neck. Four DCE-NIR datasets (two from each contrast agent) were acquired for each rat. The dye concentration curve in arterial blood, which is required to quantify the model parameters, was measured non-invasively by dye densitometry. A modification to the kinetic model to characterize t_c as a distribution of possible values, rather than finite, improved the fit of acquired tumor concentration curves, resulting in more reliable estimates. This modified kinetic model identified a difference between the extracted fraction of IRD_cxb, 15 ± 6 %, and ICG, 1.6 ± 0.6 %, in the tumor, which can be explained by the difference in molecular weight: 67 kDa for ICG since it binds to albumin and 1.17 kDa for IRD. This study demonstrates the ability of DCENIRS to quantify tumor physiology. The next step is to adapt this approach with a dual-receptor approach.

We design a Time-Resolved (TR) instrumentation coupled with a reconstruction method based on Mellin-Laplace Transform (MLT) to accurately assess in depth absorption and diffusion maps of a cylindrical diffusive medium. To deal with experimental large TR dataset, MLT processing is handled without expressing the sensitivity matrix. Moreover an optimization of the TR probe geometry is performed to limit the number of measurements while keeping the sensitivity of the system. Simulations show how to optimize the probe geometry for specific inclusions depth given a background diffusing medium. These results lead to an experimental bench we use to perform experimental validations. This includes a femtosecond laser coupled with an HRI and a CCD camera.

Time-resolved imaging has been well-established as the most powerful imaging paradigm in optical tomography by providing rich information datasets for both functional and fluorescence tomography. The practical implementations of time-resolved imaging platforms however are limited by lengthy acquisition times and demand highly stable instrumentation for collection of robust datasets. In recent years, wide-field imaging strategies have been implemented for time-resolved imaging allowing a fast acquisition of spatially- and temporally- rich datasets within short acquisition times. In this work, we present wide-field illumination and processing strategies which significantly improve the signal-to-noise ratio of time-resolved measurements. First, we demonstrate the impact of temporal and photon noise on timeresolved measurements and compare the performance of various born-normalization schemes designed to improve the robustness of the time-resolved data-types used for reconstruction. Secondly, we present a de-noising algorithm design for time-gated data types which reduce errors arising due to noise and conclude with experimental validation of the approach. The adoption of these strategies alleviates some of the limitations associated with time-resolved imaging, especially when using more advanced data types such as early gates, thus allowing the wider acceptance of timeresolved methods for biomedical applications.

Rapidly pulsed laser excitation with high-rate signal detection allows the dispersion of emitted fluorescent signals through tissue to be measured. This “time-domain” approach to measuring diffuse fluorescence data offers a wide range of advantages over conventional “continuous-wave” fluorescence imaging. While continuous-wave imaging provides a measure of only fluorescence intensity at each source-detector pairing, time-domain approaches provide fluorescence intensity throughout the whole temporal response to the pulsed-laser excitation, adding a whole new dimension to the dataset. This paper introduces the instrumentation required to collect time-domain data, dealing with complications and nuances of system and data calibration, as well as the major current applications of time-domain diffuse fluorescence imaging.

In this work, synthetic time-domain data are generated as if it were collected with a state-of-the-art multi-view experimental optical scanner developed in our group for small animal imaging, and used in a tomographic image reconstruction algorithm. The collected data comprises full time-dependent optical signals leaving the biological medium and acquired all around the medium. The diffuse optical tomography (DOT) algorithm relies on the time dependent parabolic simplified spherical harmonics (TD-pSP_N) equations as the forward model to recover the 3D absorption and diffusion coefficient maps of the medium. The inverse problem is casted and solved as an iterative constrained optimization problem where an objective function determines the accuracy of the forward model predictions at each iteration. Time-dependent adjoint variables are introduced to accelerate the calculation of the gradient of the objective function. A three-dimensional case involving an absorption heterogeneity in a homogeneous medium is presented, reproducing practical situations encountered in our lab. The results support our hypothesis that accurate quantitative 3D maps of optical properties of biological tissues can be retrieved using intrinsic measurements obtained with our experimental scanner along with our DOT algorithm.

A non-contact type time-domain system for the fluorescence diffuse optical tomography was designed. The system is evaluated by a phantom with a fluorescence target. The contamination of the non-specific scattering superimposed on the excitation profile but it could be reduced with closely locating the detection fiber to the surface (~1 mm). Next, we analyzed the contamination in the temporal profiles with an Intralipid solution phantom with a fluorescent target. The contamination to the excitation profile is not clearly observed but that to the fluorescence is strong with a short distance between the excitation source and detection. Finally, we have concluded that a larger distance of source and detector yields better fluorescence sensitivity because the background is limiting the fluorescence detection. On the other hand, the signal quality depends on the statistics and thus the optimum range of the distance comes around 30 mm. Finally, this research gives the idea for the design of the source and detection configuration.

Time-resolved (TR) techniques have the potential to distinguish early- from late-arriving photons. Since light travelling through superficial tissue is detected earlier than photons that penetrate the deeper layers, time-windowing can in principle be used to improve the depth sensitivity of TR measurements. However, TR measurements also contain instrument contributions – referred to as the instrument-response-function (IRF) – which cause temporal broadening of the measured temporal-point-spread-function (TPSF). In this report, we investigate the influence of the IRF on pathlength-resolved absorption changes (Δμ_a) retrieved from TR measurements using the microscopic Beer-Lambert law (MBLL). TPSFs were acquired on homogeneous and two-layer tissue-mimicking phantoms with varying optical properties. The measured IRF and TPSFs were deconvolved to recover the distribution of time-of-flights (DTOFs) of the detected photons. The microscopic Beer-Lambert law was applied to early and late time-windows of the TPSFs and DTOFs to access the effects of the IRF on pathlength-resolved Δμ_a. The analysis showed that the late part of the TPSFs contains substantial contributions from early-arriving photons, due to the smearing effects of the IRF, which reduced its sensitivity to absorption changes occurring in deep layers. We also demonstrated that the effects of the IRF can be efficiently eliminated by applying a robust deconvolution technique, thereby improving the accuracy and sensitivity of TR measurements to deep-tissue absorption changes.

We derive a modification method to simplify the SP_N coupled partial differential equations into some independent equations. The modification leads to significant mathematical simplifications and can be used to calculate the Green’s function of the SP_N equations for infinite and semi-infinite turbid medium. The obtained analytical solutions depend on eigenvectors and eigenvalues. Compared with the derived methods based on coupled equations, the derivation process of our proposed method is general, fast and simple. The derived analytical solutions are successfully verified by comparisons with Monte Carlo simulations.

In this paper, frequency-domain endoscopic diffuse optical tomography image reconstruction algorithm based on dual-modulation-frequency and dual-points source diffuse equation is investigated for the reconstruction of the optical parameters including the absorption and reducing scattering coefficients. The forward problem is solved by the finite element method based on the frequency domain diffuse equation (FD-DE) for dual-points source approximation and multi-modulation-frequency. In the image reconstruction, a multi-modulation-frequency Newton-Raphson algorithm is applied to obtain the solution. To further improve the image accuracy and quality, a method based on the region of interest (ROI) is applied on the above procedures. The simulation is performed in the tubular model to verify the validity of the algorithm. Results show that the FD-DE with dual-points source approximate is more accuracy at shorter source-detector separation. The reconstruction with dual-modulation-frequency improves the image accuracy and quality compared to the results with single-modulation-frequency and triple-modulation-frequency method. The peak optical coefficients in ROI (ROI_max) are almost equivalent to the true optical coefficients with the relative error less than 6.67%. The full width at half maximum (FWHM) achieves 82% of the true radius. The contrast-to-noise ratio (CNR) and image coefficient(IC) is 5.678 and 26.962, respectively. Additionally, the results with the method based on ROI show that the ROI_max is equivalent to the true value. The FWHM can improve by 88% of the true radius. The CNR and IC is improved over 7.782 and 45.335, respectively.

A Gen-2 hand-held optical imager has been developed capable of 2D surface imaging and 3D tomography. In the current work, the capability of the imager to resolve two closely placed targets is assessed via 2D and 3D tomographic studies. Resolution studies have been carried out under various experimental conditions using slab phantoms. Preliminary 2D surface images of reflected measurements have demonstrated the ability of the system to resolve 0.95cm diameter targets placed 0.5cm apart at 2cm depth. Three dimensional tomography reconstructions are currently performed to assess the resolution capacity under different experimental conditions.

Diffuse optical topography remains a valid tool in functional near infrared spectroscopy (fNIRS) since it avoids solving the forward and inverse computational problems, which are encountered in diffuse optical tomography. Topography is particularly useful when a sparse array of optodes is used and depth specificity is not the primary interest. We have developed an easy toolbox for diffuse optical topography (“EasyTopo”) based on a standard template of brain atlas. EasyTopo approximates the cortical layer of the brain as a hemispherical surface. Therefore, the stereotaxic coordinates of the brain surface and the co-registered fNIRS measurements (channels) are converted into the spherical coordinates, where 2D angular interpolation of the channel-wise data is implemented to obtain a topographic image of brain activation in the latitude-longitude space. Then, the interpolated image is projected back onto the brain surface in the original 3D stereotaxic coordinates. Compared with the existing 3D topography methods, EasyTopo is more computationally efficient and does not require any data extrapolation. Another advantage of EasyTopo is that the data between two spatially adjacent channels are interpolated along their included angles (i.e., along the angular direction) rather than along a straight line going under the brain surface. The former geometry in principle matches better with the realistic brain structure than the latter one. EasyTopo has been validated with both simulation and human experiments. Now this toolbox is publically available.

Functional near-infrared spectroscopy (fNIRS) is a well-known technique for non-invasively measuring cerebral blood oxygenation, and many studies have demonstrated that fNIRS signals can be related to cognitive function. However, the fNIRS signal is attenuated by the skin, while scalp blood content has been reported to influence cerebral oxygenation measurements. Mechanical indentation has been shown to increase light transmission through soft tissues by causing interstitial water and blood flow away from the compressed region. To study the effects of indentation on fNIRS, a commercial fNIRS system with 16 emitter/detector pairs was used to measure cerebral blood oxygenation at 2 Hz. This device used diffuse reflectance at 730 nm and 850 nm to calculate deoxy- and oxy-hemoglobin concentrations. A borosilicate glass hemisphere was epoxied over each sensor to function as both an indenter and a lens. After placing the indenter/sensor assembly on the forehead, a pair of plastic bands was placed on top of the fNIRS headband and strapped to the head to provide uniform pressure and tightened to approx. 15 N per strap. Cerebral blood oxygenation was recorded during a breath holding regime (15 second hold, 15 second rest, 6 cycles) in 4 human subjects both with and without the indenter array. Results showed that indentation increased raw signal intensity by 85 ± 35%, and that indentation increased amplitude of hemoglobin changes during breath cycles by 313% ± 105%. These results suggest that indentation improves sensing of cerebral blood oxygenation, and may potentially enable sensing of deeper brain tissues.

Instrument equivalence and quality control are critical elements of multi-center clinical trials. We currently have five identical Diffuse Optical Spectroscopic Imaging (DOSI) instruments enrolled in the American College of Radiology Imaging Network (ACRIN, #6691) trial located at five academic clinical research sites in the US. The goal of the study is to predict the response of breast tumors to neoadjuvant chemotherapy in 60 patients. In order to reliably compare DOSI measurements across different instruments, operators and sites, we must be confident that the data quality is comparable. We require objective and reliable methods for identifying, correcting, and rejecting low quality data. To achieve this goal, we developed and tested an automated quality control algorithm that rejects data points below the instrument noise floor, improves tissue optical property recovery, and outputs a detailed data quality report. Using a new protocol for obtaining dark-noise data, we applied the algorithm to ACRIN patient data and successfully improved the quality of recovered physiological data in some cases.

Indocyanine Green encapsulating liposomes (Lip/ICG) and scVEGF-Lip/ICG liposomes, decorated with site-specifically lipidated engineered single-chain vascular endothelial growth factor (scVEGF) for targeting VEGF receptors were tested as potential tracers for fluorescent tomography. Two groups of experiments were conducted with tumor-bearing mice (n=4 to 6 per group) with tumors placed in a scattering medium at the imaging depths of 1.5 and 2.0 cm. Lip/ICG and scVEGF-Lip/ICG were injected intravenously in the amounts corresponding to 5 nmol of ICG/mouse. We detected kinetics of increase and decline in fluorescent signals in tumors for both imaging depths and for both targeted and untargeted Lip/ICG. Maximum fluorescent signals were approximately 2-fold higher at 1.5 cm vs. 2.0 cm imaging. A signal from untargeted Lip/ICG reached maximum at 15 min post-injection and then rapidly declined with t_1/2 ~15 min. In contrast, a signal from targeted scVEGF-Lip/ICG reached maximum at 30 min post-injection and then slow declined with t1/2 ~60-90 min. Preferential retention of scVEGF-Lip(ICG) vs. Lip(ICG) was confirmed by the analysis of fluorescence in cryosections of corresponding tumors, harvested at 400 min post-injection. Our results suggest that targeted scVEGF-Lip/ICG can provide for significantly better post-injection time window for detection of relatively deeply seated tumors.

The hemodynamic changes during natural human sleep are still not well understood. NIRS is ideally suited for monitoring the hemodynamic changes during sleep due to the properties of local measurement, totally safe application and good tolerance to motion. Several studies have been conducted using NIRS in both normal subjects and patients with various sleep disorders during sleep to characterize the hemodynamic changing patterns during different sleep stages and during different symptoms such as obstructive apneas. Here we assessed brain and muscle oxygenation changes in 7 healthy adults during all-night sleep with combined polysomnography measurement to test the notion if hemodynamic changes in sleep are indeed brain specific. We found that muscle and brain showed similar hemodynamic changes during sleep initiation. A decrease in HbO₂ and tissue oxygenation index (TOI) while an increase in HHb was observed immediately after sleep onset, and an opposite trend was found after transition with progression to deeper slow-wave sleep (SWS) stage. Spontaneous low frequency oscillations (LFO) and very low frequency oscillations (VLFO) were smaller (Levene’s test, p<0.05) during SWS compared to light sleep (LS) and rapid-eye-movement (REM) sleep in both brain and muscle. Spectral analysis of the NIRS signals measured from brain and muscle also showed reductions in VLFO and LFO powers during SWS with respect to LS and REM sleep. These results indicate a systemic attenuation rather than local cerebral reduction of spontaneous hemodynamic activity in SWS. A systemic physiological mechanism may exist to regulate the hemodynamic changes in brain and muscle during sleep.

Near infrared (NIR) diffuse optical imaging (DOI) are increasingly used to detect hemodynamic changes in the cerebral cortex induced by brain activity. For the sake of capturing the dynamic changes in real-time imaging applications, such as brain imaging, digital lock-in detection technique could be applied. Using particular modulation and sampling constraints and averaging filters, one can achieve optimal noise reduction and discrimination between sources in different modulation frequencies. In this paper, we designed and developed a compact dual-wavelength continuous wave DOI system based on the single photon counting digital lock-in detection technique. According to the frequency division multiplexing light source coding technique, sine waves with different frequencies are generated so as to amplitude-modulate two laser sources with different wavelengths. The diffuse light is detected by photomultiplier tubes (PMTs) and the data is collected by the detection channels simultaneously. A digital lock-in detection circuit for photon counting measurement module and a DDS (Direct Digital Synthesizer) signal generation module were separately implemented in two FPGA development platforms. To validate the feasibility and functionality of the developed system, a series of experimental tests were performed. Preliminary results show that the system could be used to reconstruct the absorption coefficient and could separate the response of the dual wavelength sources which were modulated by sine signals of different frequencies effectively. In addition, several imaging experiments were performed on the semi-infinite solid phantom to find the “best imaging position” for a given source-detector placement.

As a new non-invasive medical imaging technology, diffuse optical tomography (DOT) has received considerable attention that can provide vast quantities of functional information of tissues. The reconstruction problem of DOT is highly ill-posed, meaning that a small error in the measurement data can bring about drastic errors of the reconstruction optical properties. In this paper, the shape-based image reconstruction algorithm of DOT is proposed for reducing the ill-poseness under the assumption that the optical properties of target region distribute uniformly. Since some human organs and tumors can be simplified as an ellipsoid, in this paper, the shape of the inhomogeneity is described as an ellipsoid. In the forward problem, the boundary element method (BEM) is implemented to solve the continuous wave diffusion equation (DE). By the use of the ellipsoid parametric method, the description of the shape, location and optical properties of the inhomogeneity, and the value of the background could be realized with only a small number of parameters. In the inverse calculation, a Levenberg-Marquardt algorithm with line searching is implemented to solve the underlying nonlinear least-squares problem. Simulation results show that the algorithm developed in this paper is effective in reducing the ill-poseness and robust to the noise.

In the non-invasive brain imaging with near-infrared light, precise head model is of great significance to the forward model and the image reconstruction. To deal with the individual difference of human head tissues and the problem of the irregular curvature, in this paper, we extracted head structure with Mimics software from the MRI image of a volunteer. This scheme makes it possible to assign the optical parameters to every layer of the head tissues reasonably and solve the diffusion equation with the finite-element analysis. During the solution of the inverse problem, a semi-3D reconstruction algorithm is adopted to trade off the computation cost and accuracy between the full 3-D and the 2-D reconstructions. In this scheme, the changes in the optical properties of the inclusions are assumed either axially invariable or confined to the imaging plane, while the 3-D nature of the photon migration is still retained. This therefore leads to a 2-D inverse issue with the matched 3-D forward model. Simulation results show that comparing to the 3-D reconstruction algorithm, the Semi-3D reconstruction algorithm cut 27% the calculation time consumption.

A region-based approach of image reconstruction using the finite element method is developed for diffuse optical tomography (DOT). The method is based on the framework of the pixel-based DOT methodology and on an assumption that different anatomical regions have their respective sets of the homogeneous optical properties distributions. With this hypothesis, the region-based DOT solution greatly improves the ill-posedness of the inverse problem by reducing the number of unknowns to be reconstructed. The experimental validation of the methodology is performed on a solid phantom employing a multi-channel DOT system of lock-in photon-counting mode, as well as compared with the traditional pixel-based reconstruction results, demonstrate that the proposed DOT methodology presents a promising tool of in vivo reconstructing background optical structures with the aid of anatomical a priori.

In this paper, we constructed a continuous wave non-contact diffuse optical tomography (DOT) system for the dense sampling of both the illumination and detection by using a laser raster scanning and a CCD-based data acquisition. For dealing with the large size of measurement data obtained from the non-contact system, a fast tomographic image reconstruction scheme for reconstructing the absorption coefficient of a slab is developed. The proposed algorithm is carried out with the spatial-frequency encoding in both the measurement and the image spaces, and involves a strategy for selecting the useful spatial frequency based on the transfer function of tissue. Dense sampling offers an effective way of improving the image reconstruction performances and the developed algorithm is expected to considerably reduce the calculation time for reconstruction whilst retain the quality of the reconstructed images. Reconstructions from the experimental data show that the inversion scheme developed in this paper can get an absorption image within 20s, which has higher quality than those reconstructed in several hours by using the conventional reconstruction method.

Diffuse optical tomography was recognized as one of the most potential methods to in-vivo imaging due to its advantages of non-invasiveness, high sensitivity and excellent specificity etc. This modality aims at portraying the concentration distribution of oxy-hemoglobin and deoxy-hemoglobin statically or dynamically by resolving the optical properties at multiple wavelengths. To further improve the instantaneity and sensitivity of the method, we have developed a continuous-wave diffuse optical tomography system based on lock-in photon-counting technique, which can perform dual-wavelength measurement simultaneously at ultra-high sensitivity. The system was configured by modulating the laser sources at different wavelengths with different frequencies and adopting a single photon-counting block based on the digital lock-in detection for the data demodulation. Phantom experiments were conducted to evaluate the capability of the method. Results have shown that the absorption contrast can be commendably reconstructed, and the system we proposed provides a promising tool for in-vivo imaging.

A non-contact near-infrared spectroscopy (NIRS) scanning system with a phosphor cell placed on the skin for in vivo measurement of biological tissue was developed and evaluated. Because the phosphor is excited by the light that propagates in the tissue, and the excitation light is cut by optical filters, the light that propagates in the tissue is selectively detected. The non-contact system was extended to create a scanning system that can flexibly change source positions with a galvano scanner. The optical scanning system was used for non-contact measurement of the human forearm muscle, and the dependence of optical-density change (ΔOD) caused by the upper-arm occlusion and release on source-detector distance was observed. The obtained ΔOD demonstrates the effectiveness of using this system for multi-distance human-forearm measurement. Furthermore, a human forehead was measured with the system. To extract a deep-layer signal, a surface-layer subtraction method with short-distance regression was applied to measured data. On the basis of the correlation with a simultaneously measured laser-Doppler flowmetry signal, it was confirmed that the deep-layer signal was successfully extracted. The extraction result demonstrates that the optical scanning system can be used as a multi-distance NIRS system for measuring the human brain activity at the forehead.

To estimate the depth distribution of the absorption coefficient in a turbid medium, a new nonlinear inversion technique was developed. It can resolve important shortcomings of the nonlinear error of conventional linear inversion techniques. First, a turbid medium is divided into imaginary layers with arbitrary thickness. The spatial pathlength distribution (SPD) is obtained for each layer in the Monte Carlo simulation as a function of source-detector distance. In the integral operation using SPDs, we can obtain the absorption coefficient of each layer, or the depth distribution of absorption. This inversion process is based on the assumption that the light attenuation is linear with respect to the small pathlength of a photon. However, if we consider the variance of pathlength of many photons in each layer, then this assumption produces nonlinear error. We developed a technique to solve this problem in the following three steps. First, the initial values of absorption coefficient of each layer are obtained using a conventional linear inverse matrix assuming that variance of pathlength of many photons does not exist. Then, improved absorption coefficients are obtained with these initial values and the pathlength variance of many photons using the same matrix. In this way, the nonlinear error is corrected. Repeating this process, the improved absorption coefficient approaches a true value. The effectiveness of the proposed technique was confirmed in the Monte Carlo simulation. The effect of measurement noise was analyzed in the simulation.

Diffuse reflectance spectroscopy using fiber optic probe is one of most promising technique for evaluating optical properties of biological tissue. We present a method determining the reduced scattering coefficients μ_s’, the absorption coefficients μ_a, and tissue oxygen saturation StO₂ of in vivo brain tissue using single reflectance fiber probe with two source-collector geometries. In this study, we performed in vivo recordings of diffuse reflectance spectra and the electrophysiological signals for exposed brain of rats during the cortical spreading depression (CSD) evoked by the topical application of KCl. The time courses of μ_a in the range from 500 to 584 nm and StO₂ indicated the hemodynamic change in cerebral cortex. Time courses of μ_s’ are well correlated with those of μ_a in the range from 530 to 570 nm, which also reflect the scattering by red blood cells. On the other hand, increases in μ_s’ at 500 and 584 nm were observed before the profound increase in μ_a and they synchronized with the negative DC shift of the local field potential. It is said that the DC shift coincident with a rise in extracellular potassium and can evoke cell deformation generated by water movement between intracellular and extracellular compartments, and hence the light scattering by tissue. Therefore, the increase in μ_s’ at 500 and 584 nm before the profound increase in μ_a are indicative of changes in light scattering by tissue. The results in this study indicate potential of the method to evaluate the pathophysiological conditions of in vivo brain.

Working memory (WM) is fundamental for a number of cognitive processes, such as comprehension, reasoning and learning. WM allows the short-term maintenance and manipulation of the information selected by attentional processes. The goal of this study was to examine by time-resolved fNIRS neural correlates of the verbal and visual WM during forward and backward digit span (DF and DB, respectively) tasks, and symbol span (SS) task. A neural dissociation was hypothesised between the maintenance and manipulation processes. In particular, a dorsolateral/ventrolateral prefrontal cortex (DLPFC/VLPFC) recruitment was expected during the DB task, whilst a lateralised involvement of Brodmann Area (BA) 10 was expected during the execution of the DF task. Thirteen subjects were monitored by a multi-channel, dual-wavelength (690 and 829 nm) time-resolved fNIRS system during 3 minutes long DF and DB tasks and 4 minutes long SS task. The participants’ mean memory span was calculated for each task: DF: 6.46±1.05 digits; DB: 5.62±1.26 digits; SS: 4.69±1.32 symbols. No correlation was found between the span level and the heart rate data (measured by pulse oximeter). As expected, DB elicited a broad activated area, in the bilateral VLPFC and the right DLPFC, whereas a more localised activation was observed over the right hemisphere during either DF (BA 10) or SS (BA 10 and 44). The robust involvement of the DLPFC during DB, compared to DF, is compatible with previous findings and with the key role of the central executive subserving in manipulating processes.

A multichannel (16 sources and 8 detectors) time-domain fNIRS medical device is presented. The system was extensively characterized on tissue phantoms. Preliminary in vivo measurements on muscle and brain cortex are reported to test the ability of the system to noninvasively measure tissue hemodynamics.

We have previously presented an inverse Monte Carlo algorithm based on a three-layer semi-infinite skin model for analyzing diffuse reflectance spectroscopy (DRS) data. The algorithm includes pre-simulated Monte Carlo data for a range of physiologically relevant epidermal thicknesses and tissue scattering levels. The simulated photon pathlength distributions in each layer are stored and the absorption effect from tissue chromophores added in the post-processing. Recorded DRS spectra at source-detector distances of 0.4 and 1.2 mm were calibrated for the relative intensity between the two distances and matched to simulated spectra in a non-linear optimization algorithm. This study evaluates the DRS spectral fitting accuracy and presents data on the main output parameters; the tissue fraction of red blood cells and local oxygenation (S_O2). As a reference, the microcirculatory perfusion (Perf) was measured simultaneously in the same probe using laser Doppler Flowmetry. Data were recorded on the volar forearm of three healthy subjects in a protocol involving a 5 min systolic occlusion. The DRS spectra were modeled with an rms-error < 2%. In two subjects, S_O2 decreased during occlusion to <10%, and increased to above baseline after hyperemia, while Perf increased >7 times compared to baseline. In the third subject the S_O2 decreased less during occlusion and increased to baseline values at hyperemia with only a 2-fold increase in Perf. The observed difference could be due to different microvascular beds being probed. It is concluded that integrating DRS and LDF enables new possibilities to deduce microcirculation status.

For improvement of diffuse optical tomography, one needs a numerical model to compute photon propagation accurately and efficiently. Thus, in the paper, we construct a hybrid model based on the radiative transfer equation (RTE) and diffusion equation (DE) in the time domain under the refractive-index-mismatching at boundary. At first, a fictive interface between the RTE and DE regions is determined, which is characterized by a length scale ρ_DA. By comparing the fluence rate calculated based on the RTE to that on the DE, we estimate ρ_DA at ~10/μ_t, where μ_t denotes the extinction coefficient. Next, we determine a coefficient representing the reflectivity in the Robin boundary condition of the DE by analyzing the fluence rates in the domain outside ρ_DA. Then, the hybrid model is constructed by using the determinations. The fluence rates based on this hybrid algorithm is consistent with those on the RTE in a whole range of the medium and computational costs are reduced efficiently.

This PDF file contains the front matter associated with SPIE Proceedings Volume 8578, including the Title Page, Copyright Information, Table of Contents, and the Conference Committee listing.