This PDF file contains the front matter associated with SPIE Proceedings Volume 7896, including the Title Page, Copyright information, Table of Contents, and the Conference Committee listing.

To facilitate the design and optimization of instruments for time-domain optical brain imaging within the European project "nEUROPt", the performance of various instruments is assessed and compared. This type of instruments relies on picosecond lasers with high repetition rates, fast detectors and time-correlated single photon counting. The first step of the assessment included a number of basic tests that are related to parameters of the source, to the differential nonlinearity of the timing electronics and to the temporal instrument response function (IRF). An additional test has been devised to measure the responsivity of the detection system, i.e. the overall efficiency to collect and detect light emerging from tissue. Dedicated solid slab phantoms have been developed and quantitatively spectrally characterized to provide sources of known radiance with nearly Lambertian angular characteristics. The wavelength-dependent transmittance factor of these phantoms was of the order of 10²⁰/(W s m²sr). Measurements of the responsivity of the detection systems of three time-domain optical brain imagers tested yielded similar values of the order of 0.1 mm²sr.

Cerebrovascular reactivity (CVR) reflects the compensatory dilatory capacity of cerebral vasculature to a dilatory stimulus and is an important indicator of brain vascular reserve. fMRI has been proven to be an effective imaging technique to obtain the CVR map when the subjects perform CO2 inhalation or the breath holding task (BH). However, the traditional data analysis inaccurately models the BOLD using a boxcar function with fixed time delay. We propose a novel way to process the fMRI data obtained during a blocked BH by using the simultaneously collected near infrared spectroscopy (NIRS) data as regressor¹. In this concurrent NIRS and fMRI study, 6 healthy subjects performed a blocked BH (5 breath holds with 20s durations intermitted by 40s of regular breathing). A NIRS probe of two sources and two detectors separated by 3 cm was placed on the right side of prefrontal area of the subjects. The time course of changes in oxy-hemoglobin (Δ[HbO]) was calculated from NIRS data and shifted in time by various amounts, and resampled to the fMRI acquisition rate. Each shifted time course was used as regressor in FEAT (the analysis tool in FSL). The resulting z-statistic maps were concatenated in time and the maximal value was taken along the time for all the voxels to generate a 3-D CVR map. The new method produces more accurate and thorough CVR maps; moreover, it enables us to produce a comparable baseline cerebral vascular map if applied to resting state (RS) data.

Structural abnormalities in brain microvasculature are commonly associated with Alzheimer's Disease and other dementias. However, the extent to which structural microvascular abnormalities cause functional impairments in brain circulation and thereby to cognitive impairment is unclear. Non-invasive, near-infrared spectroscopy (NIRS) methods can be used to determine the absolute hemoglobin concentration and saturation in brain tissue, from which additional parameters such as cerebral blood volume (a theoretical correlate of brain microvascular density) can be derived. Validating such NIRS parameters in animal models, and understanding their relationship to cognitive function is an important step in the ultimate application of these methods to humans. To this end we applied a non-invasive multidistance NIRS method to determine the absolute concentration and saturation of cerebral hemoglobin in rat, by separately measuring absorption and reduced scattering coefficients without relying on pre- or post-correction factors. We applied this method to study brain circulation in folate deficient rats, which express brain microvascular pathology1 and which we have shown to develop cognitive impairment.2 We found absolute brain hemoglobin concentration ([HbT]) and oxygen saturation (StO2) to be significantly lower in folate deficient rats (n=6) with respect to control rats (n=5) (for [HbT]: 73±10 μM vs. 95±14 μM; for StO2: 55%±7% vs. 66% ±4%), implicating microvascular pathology and diminished oxygen delivery as a mechanism of cognitive impairment. More generally, our study highlights how noninvasive, absolute NIRS measurements can provide unique insight into the pathophysiology of Vascular Cognitive Impairment. Applying this method to this and other rat models of cognitive impairment will help to validate physiologically meaningful NIRS parameters for the ultimate goal of studying cerebral microvascular disease and cognitive decline in humans.

We have previously reported an optical response in human subjects occurring at 100 ms following electrical stimulation of peripheral nerves. In the present study, an animal model has been created to directly investigate the myogenic components of the signal. In addition, experiments have been performed in human subjects to investigate the signal's neuroanatomical specificity, sensitivity to muscle motion, and spatial and spectral features. The results of this work suggest that the observed optical signal derives from stimulus-induced motion associated with muscle contraction and likely contains myological information of clinical value.

Near-infrared spectroscopy (NIRS) can be used to assess the cerebrovascular response to breath hold. We measured eight healthy subjects during voluntary end-expiratory breath hold to study inter- and intraindividual variability of the deoxy- (HbR) and oxyhemoglobin (HbO₂) response curves for the scalp and cerebral cortex. Although cortical [HbO₂] behaves qualitatively similarly in all subjects, there is large inter- and intraindividual variability, and in the case of [HbR] also qualitative variability. However, the linearity of [HbO₂] increase during the breath hold has encouraging measurement repeatability, and it may even indicate an individual's CO₂ tolerance. This result may help understand why breath hold duration varies between subjects more than the total [HbO₂] increase during breath hold.

The use of near-infrared spectroscopy (NIRS) is increasingly being investigated in critical care settings to assess cerebral hemodynamics, because of its potential for guiding therapy during the recovery period following brain injury. Cerebral blood flow (CBF) can be quantified by NIRS using indocyanine green (ICG) as an intravascular tracer. However, extracting accurate measurements from complex tissue geometries, such as the human head, is challenging and has hindered the clinical applications. With the development of fast Monte Carlo simulations that can take into account a priori anatomical information (e.g. near-infrared light propagation in tissue from MRI or CT imaging data), it is now possible to investigate signal contamination arising from the extracerebral layers, which can confound NIRS-CBF measurements. Here, we present a theoretical model that combines Monte Carlo simulations of broadband time-resolved near-infrared measurements with indicator-dilution theory to model time-dependent changes in light propagation following ICG bolus injection. Broadband, time-resolved near-infrared spectroscopy measurements were simulated for three source-detector positions. Individual simulations required 56 seconds for 5x10⁸ photons, and a set of simulations consisting of baseline measurements at 40 wavelengths, and single-wavelength measurements at 160 time-points required on average 3.4 hours. To demonstrate the usefulness of our model, the propagation of errors associated with varying both the scalp blood flow and the scalp thickness was investigated. For each simulation the data were analyzed using four independent approaches-simple-subtraction blood flow index (ΔBFI_SS), time-resolved variance time-to-peak (ΔTTP_TR), and absolute and relative CBF with depth-resolved NIRS (CBF_DR and ΔCBF_DR)-to assess cerebral hemodynamics.

Continuous bedside monitoring of cerebral blood flow (CBF) in patients recovering from brain injury could improve the detection of impaired substrate delivery, which can exacerbate injury and worsen outcome. Diffuse correlation spectroscopy (DCS) provides the ability to monitor perfusion changes continuously, but it is difficult to quantify absolute blood flow - leading to uncertainties as to whether or not CBF has fallen to ischemic levels. To continuously measure CBF, we propose to calibrate DCS data using a single time-point, time-resolved near-infrared (TR-NIR) technique for measuring absolute CBF. Experiments were conducted on newborn piglets in which CBF was increased by raising the arterial tension of CO₂ (40-62 mmHg) and decreased by carotid occlusion. For validation, values of CBF measured by TR-NIR were converted into blood flow changes and compared to CBF changes measured by DCS. A strong correlation between perfusion changes from the two techniques was revealed (slope = 0.98 and R² = 0.96), suggesting that a single time-point CBF measurement by TR-NIR can be used to convert continuous DCS data into units of CBF (ml/100g/min).

We present a study of the relative phase of oscillations of cerebral oxy- and deoxy-hemoglobin concentrations in the low-frequency range, namely 0.04-0.12 Hz. We have characterized the potential contributions of noise to the measured phase distributions, and we have performed phase measurements on the brain of a human subject at rest, and on the brain of a human subject during stage I sleep. While phase distributions of pseudo hemodynamic oscillations generated from noise (obtained by applying to two independent sets of random numbers the same linear transformation that converts absorption coefficients at 690 and 830 nm into concentrations of oxy- and deoxy-hemoglobin) are peaked at 180º, those associated with real hemodynamic changes can be peaked around any value depending on the underlying physiology and hemodynamics. In particular, preliminary results reported here indicate a greater phase lead of deoxy-hemoglobin vs. oxy-hemoglobin low-frequency oscillations during stage I sleep (82º ± 55º) than while the subject is awake (19º ± 58º).

The use of the spectral derivative method in Near Infrared optical spectroscopy and tomographic imaging is presented, whereby instead of using discrete measurements around several wavelengths, the difference between nearest neighboring spectral measurements is used. The proposed technique is shown to be insensitive to the unknown tissue and fiber contact coupling coefficients providing substantially increased accuracy as compared to more conventional techniques. The self-calibrating nature of the spectral derivative techniques increases its robustness in clinical applications, as is demonstrated based on simulated results.

In this work we introduce the finite volume (FV) approximation to the simplified spherical harmonics (SP_N) equations for modeling light propagation in tissue. The SP_N equations, with partly reflective boundary conditions, are discretized on unstructured grids. The resulting system of linear equations is solved with a Krylov subspace iterative method called the generalized minimal residual (GMRES) algorithm. The accuracy of the FV-SP_N algorithm is validated through numerical simulations of light propagation in a numerical phantom with embedded inhomogeneities. We use a FV implementation of the equation of radiative transfer (ERT) as the benchmark algorithm. Solutions obtained using the FV-SP_N (N > 1) algorithm are compared to solutions obtained with the ERT and the diffusion equation (SP₁). Compared to the SP₁, the SP₃ solutions obtained using the FV-SP_N algorithm can better approximate ERT solutions near boundary sources and in the vicinity of void-like regions. Solutions using the SP₃ algorithm are obtained 9.95 times faster than solutions with the ERT-based algorithm.

Fluorescence-enhanced optical imaging/tomography (FEOI/FEOT) could play an important role in drug discovery and clinical diagnostics. In recent years, improvement in spatial resolution and quantitative biological information has been increasingly promised. Time- and frequency-resolved FEOT has great potential to obtain improved reconstruction information compared with continuous wave (CW) mode due to the acquisition of more measurement information. In this paper, a phase-only FEOT(PO-FEOT) is proposed and evaluated. In PO-FEOT, a finite element-based linear relationship between the unknown fluorophore variables and phase information obtained from time- and frequency-domain boundary measurements is used in FEOT reconstruction. Synthetic data shows that, compared to the amplitude and phase-based reconstruction, the reconstruction time cost is remarkably reduced when phase is used alone. The reconstructed results show that PO-FEOT with high frequency information can acquire preferable reconstruction quality.

A spectroscopic imaging device to study brain activation, without any scan nor contact, is under construction. The entire instrument will be assembled in a unique setup and will use light emitted by picosecond laser diodes, a frontal light distributor and a time-gated intensified camera. The instrument is controlled by an FPGA based module which generates the pulse sequences for laser diodes and for the photocathode of the micro-channel-plate intensifier, and for the trigger of the CCD camera. A time resolved 3D simulation study, using the Finite Element Method, was performed in order to evaluate the proposed method for brain activation imaging. It is based on the widely used Brainweb digital brain phantom, where the tissues of the whole head were distributed into 10 classes, for which optical absorption and scattering coefficients were determined accordingly to the literature. Simulation data were calibrated thanks to timeresolved experiments and results will be presented with special attention on the sensitivity and accuracy for detection of optical absorption changes due to brain activation.

One of the major challenges in diffuse optical tomography (DOT) is attributed to the severe decay of sensitivity along depth. In conventional reconstruction method using regularized inversion, it yields significant depth distortion in the reconstructed image as a cortical activation is always projected into the skull. Recently we developed a depth compensation algorithm (DCA) to minimize the depth localization error in DOT, which introduces a depth-variant weight matrix to counterbalance the severe sensitivity decay of A-matrix. The DCA algorithm has been previously validated in both laboratory phantom experiments and an in vivo human study. In this study, we first present a comprehensive analysis on how DCA alters the depth localization and spatial resolution in DOT. It reveals that DCA greatly improves the transverse resolution in sub-cortical region. Second, we present a quantification approach for DCA. By forming a spatial prior directly from the reconstructed image, this approach greatly improves the quantification accuracy in DOT.

In MRI-guided diffuse optical tomography of the human brain function, three-dimensional anatomical head model consisting of up to five segmented tissue types can be specified. With disregard to misclassification between different tissues, uncertainty in the optical properties of each tissue type becomes the dominant cause of systematic error in image reconstruction. In this study we present a quantitative evaluation of image resolution dependence due to such uncertainty. Our results show that given a head model which provides a realistic description of its tissue optical property distribution, high-density diffuse optical tomography with cortically constrained image reconstruction are capable of detecting focal activation up to 21.81 mm below the human scalp at an imaging quality better than or equal to 1.0 cm in localization error and 1.0 cm³ in FVHM with a tolerance of uncertainty in tissue optical properties between +15% and -20%.

Stroke, due to ischemia or hemorrhage, is the neurological deficit of cerebrovasculature and is the third leading cause of death in the United States. More than 80 percent of stroke patients are ischemic stroke due to blockage of artery in the brain by thrombosis or arterial embolism. Hence, development of an imaging technique to image or monitor the cerebral ischemia and effect of anti-stoke therapy is more than necessary. Near infrared (NIR) optical tomographic technique has a great potential to be utilized as a non-invasive image tool (due to its low cost and portability) to image the embedded abnormal tissue, such as a dysfunctional area caused by ischemia. Moreover, NIR tomographic techniques have been successively demonstrated in the studies of cerebro-vascular hemodynamics and brain injury. As compared to a fiberbased diffuse optical tomographic system, a CCD-camera-based system is more suitable for pre-clinical animal studies due to its simpler setup and lower cost. In this study, we have utilized the CCD-camera-based technique to image the embedded inclusions based on tissue-phantom experimental data. Then, we are able to obtain good reconstructed images by two recently developed algorithms: (1) depth compensation algorithm (DCA) and (2) globally convergent method (GCM). In this study, we will demonstrate the volumetric tomographic reconstructed results taken from tissuephantom; the latter has a great potential to determine and monitor the effect of anti-stroke therapies.

In this work a generalization of the approach allowing time-domain (TD) excitation and fluorescence data to be generated using a finite element model (FEM) is introduced. This new functionality allows simulation of temporal point-spread functions (TPSF) for a heterogeneous scattering and absorbing media of arbitrary geometry. In the first part of this paper, the approach used to develop a computationally efficient model for solving the time-dependent diffusion equation for excitation and fluorescence data is presented. In the second part, a detailed theoretical evaluation of the method is given by comparing the developed FEM simulations with analytical and Monte Carlo data. The total fluence (intensity data), shows qualitative match whereas meantime of flight is almost identical among the three models for both excitation and emission data. The results show that the model is reliable and warrants its use for future TD applications where diffusion modelling can be used.

We predict the phenomenon of "spiral-planar equivalence" for steady-state photon diffusion associated with a cylindrical applicator. Recently we have derived a unified theory of steady-state photon diffusion in a homogenous medium bounded either externally (referred to as a concave geometry) or internally (referred to as a convex geometry) by an infinitely long circular cylindrical applicator [JOSAA, 27(3): 648-662 (2010)]. Despite the idealization of the geometry by assuming an infinite length of the applicator, the analytic prediction withholds the quantitative examinations based on experimental measurements, and finite-element solution of photon diffusion. An interesting finding is that the decay of photon fluence in a concave boundary is smaller in the azimuth direction but greater along the longitudinal direction, in comparison with that in a semi-infinite geometry along a straight line, for the same line-of-sight distance between the source and the detector. Conversely, the decay of photon fluence in a convex boundary is greater in the azimuth direction but smaller along the longitudinal direction, in comparison with that in a semi-infinite geometry along a straight line, for the same line-of-sight source-detector distance. These findings suggest that on the cylindrical applicator interface there should exist a spiral direction (oblique to both the azimuthal and longitudinal directions), along which the rate of photon fluence decay follows that along a straight line on a planar semi-infinite interface---which is called the "spiral-planar equivalence". The "spiral-planar equivalence" is derivable analytically, and subject to quantitative evaluations. Validating the "spiral-planar equivalence" not only enriches the understanding of photon diffusion in cylindricalinterface geometry, but also provides unique semi-infinite-based imaging application in trans-lumenal diffuse optical sensing. The "spiral-planar equivalence" may be applicable to time-resolved photon-diffusion.

We introduce here a transport-theory-based PDE-constrained multispectral imaging algorithm for direct reconstruction of the spatial distribution of chromophores in tissue. The method solves the forward and inverse problems simultaneously in the framework of a reduced Hessian sequential quadratic programming method. The performance of the new algorithm is evaluated using numerical and experimental studies involving tumor bearing mice. The results show that the PDE-constrained multispectral method leads to 15-fold acceleration in the image reconstruction of tissue chromophores when compared to the unconstrained multispectral approach and also gives more accurate results when compared to the traditional two-step method.

We present a multi-laboratory comparison of several independent forward solvers used for photon migration through layered media. Two main categories of forward solvers are presented: Monte Carlo procedures and solutions of the diffusion equation for the time domain. For Monte Carlo we have included four independent codes. For the solutions of the diffusion equation, we have presented: two semi-analytical approaches based on the Green's function method and one solution obtained with the finite element method. The comparisons between the different time-dependent solutions were performed for a two-layer medium.

Local fine representation of the fluorescence map on the standard mesh can be redundant in the sense that the reconstruction resolution is usually limited in such a severely ill-posed problem. Using global characteristic shape functions that can approximately capture the major structural information, we study fluorescence tomography with a new shape-guided representation based on some underlying mesh. Moreover, the proposed method can naturally enforce the prior coexistence of fluorescence yield and lifetime when fluorescence maps are formulated in complex sources. The simulation results suggest that, compared with standard pixel-wise representation, the shape-guided representation offers better localization of inclusions with improved quantitative accuracy, particularly in the case with inclusions of low fluorescence contrast, such as 2:1 inclusion-to-background ratio, and is more robust to the initial guess and the noise.

Optical spectroscopy has been shown to be an effective method for detecting neoplasia. Guided Therapeutics has developed LightTouch, a non invasive device that uses a combination of reflectance and fluorescence spectroscopy for identifying early cancer of the human cervix. The combination of the multispectral information from the two spectroscopic modalities has been shown to be an effective method to screen for cervical cancer. There has however been a relative paucity of work in identifying the individual spectral components that contribute to the measured fluorescence and reflectance spectra. This work aims to identify the constituent source spectra and their concentrations. We used non-negative matrix factorization (NNMF) numerical methods to decompose the mixed multispectral data into the constituent spectra and their corresponding concentrations. NNMF is an iterative approach that factorizes the measured data into non-negative factors. The factors are chosen to minimize the root-mean-squared residual error. NNMF has shown promise for feature extraction and identification in the fields of text mining and spectral data analysis. Since both the constituent source spectra and their corresponding concentrations are assumed to be non-negative by nature NNMF is a reasonable approach to deconvolve the measured multispectral data. Supervised learning methods were then used to determine which of the constituent spectra sources best predict the amount of neoplasia. The constituent spectra sources found to best predict neoplasia were then compared with spectra of known biological chromophores.

New small-scale optical probes designed for endoscopic near-infrared imaging, can have source-detector distances considerably less than 1 cm, pushing the limits for which the diffusion approximation is valid for quantifying optical properties. Scalable Monte Carlo algorithms have been successful in recovering optical properties in this realm for homogeneous tissue; however, these techniques are unsuitable for heterogeneous media. The purpose of this study was to implement a fast Monte Carlo algorithm to retrieve optical properties in media comprised of two distinct layers (a geometry similar to that which would be seen in endoscopic prostate imaging where the rectal wall is superficial to the prostate). A two-layer phantom was constructed from two materials with known optical properties (one 5-mm thick slab was placed on a 3-cm thick slab). The phantom was raster-scanned using a reflectance-based system with a 3-mm source-detector separation. Data was analyzed using Monte Carlo eXtreme as a forward model in an iterative fitting procedure. On average, 125 iterations per pixel were required for convergence (3.8 minutes of computational time on a single nVidia GTX480 GPU card). The errors in recovered absorption coefficients were 0.58% and 0.39% in the top and bottom layers, respectively. This work demonstrates the promise of ultrafast Monte Carlo algorithms for applications within an iterative fitting routine for geometries where the diffusion approximation fails.

Relying on deeper penetration of light in the tissue, Diffuse Optical Tomography (DOT) achieves organ-level tomography diagnosis, which can provide information on anatomical and physiological features. DOT has been widely used in imaging of breast, neonatal cerebral oxygen status and blood oxygen kinetics observed by its non-invasive, security and other advantages. Continuous wave DOT image reconstruction algorithms need the measurement of the surface distribution of the output photon flow inspired by more than one driving source, which means that source coding is necessary. The most currently used source coding in DOT is time-division multiplexing (TDM) technology, which utilizes the optical switch to switch light into optical fiber of different locations. However, in case of large amounts of the source locations or using the multi-wavelength, the measurement time with TDM and the measurement interval between different locations within the same measurement period will therefore become too long to capture the dynamic changes in real-time. In this paper, a frequency division multiplexing source coding technology is developed, which uses light sources modulated by sine waves with different frequencies incident to the imaging chamber simultaneously. Signal corresponding to an individual source is obtained from the mixed output light using digital phase-locked detection technology at the detection end. A digital lock-in detection circuit for photon counting measurement system is implemented on a FPGA development platform. A dual-channel DOT photon counting experimental system is preliminary established, including the two continuous lasers, photon counting detectors, digital lock-in detection control circuit, and codes to control the hardware and display the results. A series of experimental measurements are taken to validate the feasibility of the system. This method developed in this paper greatly accelerates the DOT system measurement, and can also obtain the multiple measurements in different source-detector locations.

We present a multi-wavelength DC system using Light Emitting Diode (LED) sources of four wavelengths in the near infrared range. These LEDs are commercially available, much cheaper than laser diodes, and have adequate power to probe deeply seated lesions. In our system, 8 groups of LEDs of four wavelengths were deployed on a hand-held probe and 10 PMT detectors were fiber coupled to the probe. A co-registered ultrasound (US) array located in the middle of the probe provided lesion location and morphology, which were used for assisting near infrared imaging reconstruction. Experiments evaluated the performance of the LED based DC system using phantom targets.

A handheld photoacoustic tomography-guided diffuse optical tomography system for imaging deeply-seated targets in scattering media is presented. This hybrid imager consists of a probe with an ultrasound transducer in the center and straddled by two optical fibers for taking photoacoustic images. The diffuse optical tomography component comprises of 9 light-source fibers for delivering light to the imaged tissue, and 14 detector fibers for collecting the light. Single- and two-phantom targets of high and low optical contrasts were immersed in a scattering intralipid solution to depths of up to 3cm and imaged. The reconstructed absorption coefficients of the targets with guidance from photoacoustic tomography were compared to those obtained with a-priori depth-only information, and no a-priori information. The reconstructed absorption maps yielded as much as 2.6-fold improvement in the quantification accuracy compared to the cases with no guidance from photoacoustic tomography.

We present and develop an approach using optical interference and heterodyne technology to investigate the light migration in highly scattering media. The theoretical model is based on diffusion approximation in steady-state frequency domain. The model incorporates pair-photon dipole source in order to satisfy the emulsion boundary condition and is suitable for either refractive index matched or mismatched surface. The experimental results showed that the diffusion theory applies in this study. Under the appropriate boundary and interference condition, the study accurately estimates the optical parameters of the medium.

To improve the light quantification of clustered lesions, a multi-zone reconstruction algorithm guided by co-registered ultrasound image was investigated using simulations and phantoms. The performance of the algorithm was demonstrated using clinical examples.

Due to the exponential decay of the photon density wave, the accuracy of quantifying deeper portion of the large lesions by diffuse optical tomography in reflection geometry is much lower than that of the top portion. In this study, we introduce a modified depth correction method that incorporates the target depth information provided by co-registered ultrasound. Phantom experiments and clinical example will be presented to demonstrate the utility of the method in improving light quantification of large lesions.

A prototype time-domain fluorescence diffusion optical tomography (FDOT) system using near-infrared light is presented. The system employs two pulsed light sources, 32 source fibers and 32 detection channels, working separately for acquiring the temporal distribution of the photon flux on the tissue surface. The light sources are provided by low power picosecond pulsed diode lasers at wavelengths of 780 nm and 830 nm, and a 1×32-fiber-optic-switch sequentially directs light sources to the object surface through 32 source fibers. The light signals re-emitted from the object are collected by 32 detection fibers connected to four 8×1 fiber-optic-switch and then routed to four time-resolved measuring channels, each of which consists of a collimator, a filter wheel, a photomultiplier tube (PMT) photon-counting head and a time-correlated single photon counting (TCSPC) channel. The performance and efficacy of the designed multi-channel PMT-TCSPC system are assessed by reconstructing the fluorescent yield and lifetime images of a solid phantom.

High false positives and over diagnosis is a major problem with management of prostate cancer. A non-invasive or a minimally invasive technique to accurately distinguish malignant prostate cancers from benign tumors will be extremely helpful to overcome this problem. In this paper, we had used three different fluorescence spectroscopy techniques viz., Fluorescence Emission Spectrum (FES), Stokes' Shift Spectrum (SSS) and Reflectance Spectrum (RS) to discriminate benign prostate tumor tissues (N=12) and malignant prostate cancer tissues (N=8). These fluorescence techniques were used to determine the relative concentration of naturally occurring biomolecules such as tryptophan, elastin, NADH and flavin which are found to be out of proportion in cancer tissues. Our studies show that combining all three techniques, benign and malignant prostate tissues could be classified with accuracy greater than 90%. This preliminary report is based on in vitro spectroscopy analysis. However, by employing fluorescence endoscopy techniques, this can be extended to in vivo analysis as well. This technique has the potential to identify malignant prostate tissues without surgery.

Depth-resolved localization and quantification of fluorescence distribution in tissue, called Fluorescence Molecular Tomography (FMT), is highly ill-conditioned as depth information should be extracted from limited number of surface measurements. Inverse solvers resort to regularization algorithms that penalize Euclidean norm of the solution to overcome ill-posedness. While these regularization algorithms offer good accuracy, their smoothing effects result in continuous distributions which lack high-frequency edge-type features of the actual fluorescence distribution and hence limit the resolution offered by FMT. We propose an algorithm that penalizes the total variation (TV) norm of the solution to preserve sharp transitions and high-frequency components in the reconstructed fluorescence map while overcoming ill-posedness. The hybrid algorithm is composed of two levels: 1) An Algebraic Reconstruction Technique (ART), performed on FMT data for fast recovery of a smooth solution that serves as an initial guess for the iterative TV regularization, 2) A time marching TV regularization algorithm, inspired by the Rudin-Osher-Fatemi TV image restoration, performed on the initial guess to further enhance the resolution and accuracy of the reconstruction. The performance of the proposed method in resolving fluorescent tubes inserted in a liquid tissue phantom imaged by a non-contact CW trans-illumination FMT system is studied and compared to conventional regularization schemes. It is observed that the proposed method performs better in resolving fluorescence inclusions at higher depths.

Estrogen induced proliferation of mutant cells is widely understood to be the one of major risk determining factor in the development of breast cancer. Hence determination of the Estrogen Receptor[ER] status is of paramount importance if cancer pathogenesis is to be detected and rectified at an early stage. Near Infrared Fluorescence [NIRf] Molecular Optical Imaging is emerging as a powerful tool to monitor bio-molecular changes in living subjects. We discuss pre-clinical results in our efforts to develop an optical imaging diagnostic modality for the early detection of breast cancer. We have successfully carried out the synthesis and characterization of a novel target-specific NIRf dye conjugate aimed at measuring Estrogen Receptor[ER] status. The conjugate was synthesized by ester formation between 17-β estradiol and a hydrophilic derivative of Indocyanine Green (ICG) cyanine dye, bis-1,1-(4-sulfobutyl) indotricarbocyanine-5-carboxylic acid, sodium salt. In-vitro studies regarding specific binding and endocytocis of the dye performed on ER+ve [MCF-7] and control [MDA-MB-231] adenocarcinoma breast cancer cell lines clearly indicated nuclear localization of the dye for MCF-7 as compared to plasma level staining for MDA-MB-231. Furthermore, MCF-7 cells showed ~4.5-fold increase in fluorescence signal intensity compared to MDA-MB-231. A 3-D mesh model mimicking the human breast placed in a parallel-plate DOT Scanner is created to examine the in-vivo efficacy of the dye before proceeding with clinical trials. Photon migration and florescence flux intensity is modeled using the finite-element method with the coefficients (quantum yield, molar extinction co-efficient etc.) pertaining to the dye as obtained from photo-physical and in-vitro studies. We conclude by stating that this lipophilic dye can be potentially used as a target specific exogenous contrast agent in molecular optical imaging for early detection of breast cancer.

Quantitative measurements of fluorescent parameters have merited great interest lately for near-infrared fluorescence diffuse optical tomography - the efficient small animal imaging tool. We present a two-dimensional image reconstruction method for time-domain fluorescence diffuse optical tomography, which employs the analytical solution to the Laplace-transformed time-domain photon-diffusion equation to construct the inverse model and introduces a pair of real-domain transform-factors to effectively separate the fluorescent yield and lifetime parameters from the algebraic reconstruction technique solutions to the resultant linear inversions. By use of a specifically designed a multi-channel time-correlated single photon counting system and a normalized Born formulation for the inversion, the proposed scheme in a circular domain is experimentally validated using small-animal-sized cylindrical phantoms that embed several fluorescent targets made from 1%-Intralipid solution and differently contrasting fluorescent agents, where the time-resolved excitation and fluorescence signals are measured on the boundary. The results show that the approach retrieves the positions and shapes of the targets with a reasonable accuracy and simultaneously achieve quantitative reconstruction of the fluorescent yield and lifetime.

This article aims at the development of the fast inverse Monte Carlo (MC) simulation for the reconstruction of optical properties (absorption coefficient μ_s and scattering coefficient μ_s) of cylindrical tissue, such as a cervix, from the measurement of near infrared diffuse light on frequency domain. Frequency domain information (amplitude and phase) is extracted from the time domain MC with a modified method. To shorten the computation time in reconstruction of optical properties, efficient and fast forward MC has to be achieved. To do this, firstly, databases of the frequency-domain information under a range of μ_a and μ_s were pre-built by combining MC simulation with Lambert-Beer's law. Then, a double polynomial model was adopted to quickly obtain the frequency-domain information in any optical properties. Based on the fast forward MC, the optical properties can be quickly obtained in a nonlinear optimization scheme. Reconstruction resulting from simulated data showed that the developed inverse MC method has the advantages in both the reconstruction accuracy and computation time. The relative errors in reconstruction of the μ_s and μ_s are less than ±6% and ±12% respectively, while another coefficient (μ_s or μ_s) is in a fixed value. When both μ_s and μ_s are unknown, the relative errors in reconstruction of the reduced scattering coefficient and absorption coefficient are mainly less than ±10% in range of 45< μ_s <80 cm^-1 and 0.25< a μ <0.55 cm^-1. With the rapid reconstruction strategy developed in this article the computation time for reconstructing one set of the optical properties is less than 0.5 second. Endoscopic measurement on two tubular solid phantoms were also carried out to evaluate the system and the inversion scheme. The results demonstrated that less than 20% relative error can be achieved.

An algorithm to solve the diffuse optical tomography (DOT) problem is described which uses the anatomical information from x-ray CT images. These provide a priori information about the distribution of the optical properties hence reducing the number of variables and permitting a unique solution to the ill-posed problem. The light fluence rate at the boundary is written as a Taylor series expansion around an initial guess corresponding to an optically homogenous object. The second order approximation is considered and the derivatives are calculated by direct methods. These are used in an iterative algorithm to reconstruct the tissue optical properties. The reconstructed optical properties are then used for bioluminescence tomography where a minimization problem is formed based on the L1 norm objective function which uses normalized values for the light fluence rates and the corresponding Green's functions. Then an iterative minimization solution shrinks the permissible regions where the sources are allowed by selecting points with higher probability to contribute to the source distribution. Throughout this process the permissible region shrinks from the entire object to just a few points. The optimum reconstructed bioluminescence distributions are chosen to be the results of the iteration corresponding to the permissible region where the objective function has its global minimum. This provides efficient BLT reconstruction algorithms without the need for a priori information about the bioluminescence sources.

We propose the use of a retrieval operator for biomedical applications in near-infrared spectroscopy. The proposed retrieval operator is based on the "Optimal Estimation" method. The main characteristic of this method relates to the possibility to include prior information both on target and on forward model parameters of the inversion procedure. The possibility of the retrieval operator to elaborate a-priori information can in principle be a benefit for the whole retrieval procedure. This means that a larger number of target parameters can be retrieved, or that a better accuracy can be achieved in retrieving the target parameters. The final goal of this inversion procedure is to have an improved estimate of the target parameters. The procedure has been tested on time-resolved simulated experiments obtained with a Monte Carlo code. The results obtained show that an improved performance of the inversion procedure is achieved when prior information on target and forward model parameters is available. With the use of a priori information on target parameters we have in average a lower difference between the retrieved values of the parameters and their true values, and the error bars determined by the inversion procedure on the retrieved parameters are significantly lower. At the same time a good estimate of the errors on the forward model parameters can significantly improve the retrieval of the target parameters.

In this work, based on our previously proposed perturbation theory for the diffusion equation, we present new theoretical results in time and frequency domains. More specifically, we have developed a fourth order perturbation theory of the diffusion equation for absorbing defects. The method of Padé Approximants is used to extend the validity of the proposed theory to a wider range of absorbing contrasts between defects and background medium. The results of the theory are validated by comparisons with Monte Carlo simulations. In the frequency domain, the discrepancy between theoretical and Monte Carlo results for amplitude (AC) data are less than 10% up to an absorption contrast of Δμa ≤ 0.2 mm^-1, whereas the discrepancy of phase data is less than 1° up to Δμa ≤ 0.1 mm^-1. In the time domain, the average discrepancy is around 2-3% up to Δμa ≤ 0.06 mm^-1. The proposed method is an effective and fast forwardproblem solver that has the potential to find general applicability in a number of situations.

We report first a new derivation of the scattering phase theorem and provide, for the first time, the correct relation between the variance of phase gradient and the reduced scattering coefficient. The scattering-phase theorem is then applied to investigate bulk light scattering property from the phase map of thin slices of tissue phantoms measured by a differential phase interference (DIC) microscope using the intensity propagation equation approach. The scattering coefficient, the reduced scattering coefficient, and the anisotropy factor of the sample obtained with this approach is compared to known scattering property of the bulk samples.

Introduction: Two major disadvantages of currently available oxygenation probes are the need for contact with the skin and long measurement stabilization times. A novel oxygenation imaging device based on spatial frequency domain and spectral principles has been designed, validated preclinically on pigs, and validated clinically on humans. Importantly, this imaging system has been designed to operate under the rigorous conditions of an operating room. Materials and Methods: Optical properties reconstruction and wavelength selection have been optimized to allow fast and reliable oxyhemoglobin and deoxyhemoglobin imaging under realistic conditions. In vivo preclinical validation against commercially available contact oxygenation probes was performed on pigs undergoing arterial and venous occlusions. Finally, the device was used clinically to image skin flap oxygenation during a pilot study on women undergoing breast reconstruction after mastectomy. Results: A novel illumination head containing a spatial light modulator (SLM) and a novel fiber-coupled high power light source were constructed. Preclinical experiments showed similar values between local probes and the oxygenation imaging system, with measurement times of the new system being < 500 msec. During pilot clinical studies, the imaging system was able to provide near real-time oxyHb, deoxyHb, and saturation measurements over large fields of view (> 300 cm²). Conclusion: A novel optical-based oxygenation imaging system has the potential to replace contact probes during human surgery and to provide quantitative, wide-field measurements in near real-time.

Diffuse optical tomography (DOT) and Fluorescence mediated tomography (FMT) are powerful in-vivo optical imaging techniques but they are affected by long acquisition and computational times. Recently, the use of structured light has been proposed in order to reduce the acquisition time and also the computational time of the inverse problem. Additionally, it has been proposed to compress the measured data set to reduce the reconstruction time. Here we present our experimental approach, describing the instrument for structured illumination and wide field detection and we discuss the advantages to use a finite elements based approach. Then, we introduce the use of spatial wavelets. Our method is based on the projection of a small number of wavelet patterns (Haar and Battle-Lemarie wavelets). The detected images are wavelet transformed and the information content is compressed to achieve fast 3D reconstruction. Experimental results are presented, showing fast reconstruction of complex absorbing/fluorescent objects in thick diffusive samples. Implications for fast small animal imaging are discussed.

We investigate the feasibility of rapid near infrared diffuse optical tomography by spectrally-encoded sequential light delivery using wavelength-swept source. The wavelength-swept light beam is dispersed by a spectrometer to form "swept-spectral-encoded" light beam which scans linearly across the exit window of the spectrometer and delivers sequential illumination to linearly bundled source fibers. A data acquisition rate of 0.5 frame/second is reached from a 4mW 830nm swept-source and a 20mm-diameter transverse-imaging intra-lumenal applicator with 7 source and 8 detector channels placed in a liquid phantom. Higher rate of data acquisition is achievable with more powerful wavelength-swept source or in a smaller imaging regime. This new configuration is intended for being implemented in rapid fluorescence diffuse optical tomography by enabling sequential source-channel-encoded excitations of fluorophores.

Interferometric based imaging has a number of advantages over direct detection of photons, including very high time resolution, shot noise limited sensitivity, and reasonable cost. In spite of these advantages, diffuse optical techniques have almost exclusively used direct detection. In order to explore the feasibility of interferometrically detected multiply scattered light, we constructed a multiply scattered low coherence interferometry (MS/LCI) system. Using angle resolved detection and a novel illumination scheme, we demonstrate direct imaging through nearly 94 mean free scattering paths.

We developed a time-resolved diffuse optical tomography system that enables performing noncontact 3D-DOT measurements of irregular shapes which is appropriate for small animal imaging. To retrieve the surface mesh, a noncontact holographic setup using a sensor and an XY optical scanning system was used. We present a noncontact modeling approach that consists in computing the temporal intensity distribution of detected photons taking into account the free space propagation from, and to, the fibers disposed around the studied object at some distance from its surface. The optimization was performed once on the time-weighted moments then on some points of the temporal profiles.

We recently applied time domain near infrared diffuse optical spectroscopy (TD-NIRS) to monitor hemodynamics of the cardiac wall (oxy and desoxyhemoglobin concentration, saturation, oedema) on anesthetized swine models. Published results prove that NIRS signal can provide information on myocardial hemodynamic parameters not obtainable with conventional diagnostic clinical tools.1 Nevertheless, the high cost of equipment, acquisition length, sensitivity to ambient light are factors limiting its clinical adoption. This paper introduces a novel approach, based on the use of wavelength and code division multiplexing, applicable to TD-NIRS as well as diffuse optical imaging systems (both topography and tomography); the approach, called WS-CDM (wavelength and space code division mltiplexing), essentially consists of a double stage intensity modulation of multiwavelength CW laser sources using orthogonal codes and their parallel correlation-based decoding after propagation in the tissue; it promises better signal to noise ratio (SNR), higher acquisition speed, robustness to ambient light and lower costs compared to both the conventional systems and the more recent spread spectrum approach based on single modulation with pseudo-random bit sequences (PRBS).2 Parallel acquisition of several wavelengths and from several locations is achievable. TD-NIRS experimental results guided Matlab-based simulations aimed at correlating different coding sequences, lengths, spectrum spreading factor, with the WS-CDM performances on such tissues (achievable SNR, acquisition and reconstruction speed, robustness to channel inequalization, ...). Simulations results and preliminary experimental validation confirm the significant improvements that WS-CDM could bring to diffuse optical imaging (not limited to cardiac functional imaging).

Herein we report on hardware development and evaluation for frequency-domain photon migration (FDPM) technique that is miniaturized for incorporation into a micro-CT gantry for hybrid CT/NIR/PET imaging. Immunity to endogenous optical properties and enhanced contrast associated with fluorophore lifetime is inherent to the FDPM measurements and enables unique opportunities for quantitative tomography when compared to the time independent (continuous wave) approach. A miniaturized radiofrequency (rf) circuitry has been developed in our laboratory for homodyne FDPM measurements that makes use of a single 100MHz oscillator to simultaneously launch optically modulated excitation light into a small animal as well as to modulate an NIR sensitive image intensifier for collection of fluorescent signals. The use of a single oscillator not only eliminates signal drift that otherwise results from the use of multiple oscillators individually driving both source and detector, but also reduces the circuit footprint for incorporation into the CT gantry. Herein, overall system performance parameters of signal-to-noise ratio, measurement precision, spatial resolution, modulation depth (ac/dc), excitation light rejection, and clinically relevant data acquisition times are presented for mouse phantom data. Image reconstruction of phantom data and integration of circuitry for hybrid CT/NIR/PET imaging is also presented towards the ultimate validation of NIR optical tomography using PET imaging as a "gold-standard" for quantification.

We developed an endorectal time-resolved optical probe aiming at an early detection of prostate tumors targeted by fluorescent markers. Optical fibers are embedded inside a clinical available ultrasound endorectal probe. Excitation light is driven sequentially from a femtosecond laser (775 nm) into 6 source fibers. 4 detection fibers collect the medium responses at the excitation and fluorescence wavelength (850 nm) by the mean of 4 photomultipliers associated with a 4 channel time-correlated single photon counting card. We also developed the method to process the experimental data. This involves the numerical computation of the forward model, the creation of robust features which are automatically correctly from numerous experimental possible biases and the reconstruction of the inclusion by using the intensity and mean time of these features. To evaluate our system performance, we acquired measurements of a 40 μL ICG inclusion (10 μmol.L^-1) at various lateral and depth locations in a phantom. Analysis of results showed we correctly reconstructed the fluorophore for the lateral positions (16 mm range) and for a distance to the probe going up to 1.5 cm. Precision of localization was found to be around 1 mm which complies well with precision specifications needed for the clinical application.

In this work, we introduce an optical technique to measure sound. The technique involves pointing a coherent pulsed laser beam on the surface of the measurement site and capturing the time-varying speckle patterns using a CCD camera. Sound manifests itself as vibrations on the surface which induce a periodic translation of the speckle pattern over time. Using a parallel speckle detection scheme, the dynamics of the time-varying speckle patterns can be captured and processed to produce spectral information of the sound. One potential clinical application is to measure pathological sounds from the brain as a screening test. We performed experiments to demonstrate the principle of the detection scheme using head phantoms. The results show that the detection scheme can measure the spectra of single frequency sounds between 100 and 2000 Hz. The detection scheme worked equally well in both a flat geometry and an anatomical head geometry. However, the current detection scheme is too slow for use in living biological tissues which has a decorrelation time of a few milliseconds. Further improvements have been suggested.

It is well acknowledged that treatment efficacy could be increased and unnecessary toxicities reduced if a rapid assessment strategy were available to allow individual tailoring of cancer therapy. In this work we focus on using optical tomographic imaging to detect tumor response to anti-angiogenic treatment within the first 5 days of therapy. For this study we used two models with well-characterized and divergent responses to inhibition of vascular endothelial growth factor (VEGF). SK-NEP and NGP cells were implanted intrarenally into NCR nude mice and the resulting tumors were monitored until a threshold of 1-2 g was reached. Optical tomographic imaging with a dual-wavelength (λ = 765nm and 830nm) continuous wave system, was performed prior to the first treatment with the anti-VEGF bevacizumab (BV), as well as 1, 3, and 5 days later. We found that the SK-NEP tumor model, known to be responsive to BV treatment, shows a decrease in hemoglobin levels over the 5 days. Mice implanted with the NGP tumor model, known to be less responsive to treatment, do not show such decreases. These results were further validated with histopathological findings that showed a decrease in tumor vascularization in treated SK-NEP mice. These results suggest that optical tomography is a promising tool for monitoring early tumor response to drug therapy.

The dynamic vascular change in the breast induced by prospectively targeted and sustained end-tidal partial pressures of oxygen and carbon dioxide are imaged by a fast frame rate frequency domain tomographic system. The results of a particular normal subject case show that under the inhaled gas stimuli the maximum changes of deoxy-hemoglobin, oxygen saturation, oxy-hemoglobin and total hemoglobin in the breast are 20%, 9%, 7% and 3%, respectively.

We report the development of an instrument for diffuse spectral imaging of the human breast operating over the wavelength range 650-900 nm. This instrument images the slightly compressed human breast in a planar geometry by performing a tandem scan, over the x-y plane, of a 3 mm illumination optical fiber and a 5 mm collection optical fiber that are collinear and located on opposite sides of the breast. An edge-correction algorithm accounts for breast thickness variability over the x-y plane, a second-derivative imaging algorithm enhances the display of optical inhomogeneities, and a paired-wavelength spectral method yields oxygenation maps. We report our results of oxygenation mapping in eighteen human subjects, two of which are breast cancer patients, one with a ductal carcinoma in situ, the other with an invasive ductal carcinoma.

Hand-held optical imagers are currently developed toward clinical imaging of breast tissue. However, the hand-held optical devices developed to are not able to coregister the image to the tissue geometry for 3D tomography. We have developed a hand-held optical imager which has demonstrated automated coregistered imaging and 3D tomography in phantoms, and validated coregistered imaging in normal human subjects. Herein, automated coregistered imaging is performed in a normal human subject with a 0.45 cm³ spherical target filled with 1 μM indocyanine green (fluorescent contrast agent) placed superficially underneath the flap of the breast tissue. The coregistered image data is used in an approximate extended Kalman filter (AEKF) based reconstruction algorithm to recover the 3D location of the target within the breast tissue geometry. The results demonstrate the feasibility of performing 3D tomographic imaging and recovering a fluorescent target in breast tissue of a human subject for the first time using a hand-held based optical imager. The significance of this work is toward clinical imaging of breast tissue for cancer diagnostics and therapy monitoring.

Continuous wave optical tomography is non-ionizing, uses endogenous contrast, and can be performed quickly and at low cost which makes it a suitable imaging modality for breast cancer screening. Using multiple wavelengths of light to illuminate the breast at various angles, three-dimensional images of the distribution of chromophores such as oxy- and deoxy-hemoglobin can help identify cancerous tissue. Dynamic optical imaging can provide additional insight into cancer characteristics such as angiogenesis and metabolism. Here we present the first clinical data acquired with our novel digital breast imaging system. This system is based upon a Digital Signal Processor (DSP) architecture that leverages the immediate digitization of acquired analog data to reduce noise and quickly process large amounts of data. Employing this new instrument we have investigated the dynamic changes due to a breath hold and its potential for use in breast cancer screening. Over the course of a breath hold, images have been collected simultaneously from both breasts at a rate of 1.7 frames per second with 32 sources and 64 detectors per breast and four wavelengths of light at 765, 805, 827, and 905nm. Initial results involving one healthy volunteer and one breast cancer patient support the potential use of dynamic imaging for breast cancer detection.

A combined tomosynthesis and diffuse spectroscopy system may provide both spatial and physiological information about breast tissue. Using patient tomosynthesis images and simulated near infrared measurements, it is possible to accurately reconstruct for hemoglobin, water and lipid concentrations. This study utilizes both frequency domain and continuous wave components, given the constraints of a projection geometry but with broadband illumination in order to accurately recover chromophore concentrations from the segmented tissue regions. This analysis will assist in determining the optimum hardware components for a combined system that is currently being built, in order to achieve the best possible accuracy in quantifying tissue properties. Comparing several different configurations including variations in the number of wavelengths, number of regions reconstructed as well as reconstruction methods within the frequency domain components of the system may decrease the complexity, cost and examination time without significant decreases in accuracy.

Spectral measurements were performed from 900 to 1300 nm, using a fully automated set-up for time domain optical spectroscopy. Spectrally selected picosecond pulses emitted from a supercontinuum fiber source were used for illumination. The detection of re-emitted pulses was achieved using a photomultiplier tube with InP/InGaAsP photocathode, followed by a PC board for time-correlated single photon counting. To allow the estimate of tissue composition at long wavelengths, the optical characterization of collagen type I powder was extended up to 1300 nm. A marked absorption peak was detected around 1200 nm, which could prove useful for collagen quantification from in vivo optical data. In vivo spectral measurements of breast tissue were performed for the first time from 900 to 1300 nm in reflectance geometry. The sensitivity of the detector was very low above 1200 nm, still it allowed us to reveal a long-wavelength range (1000-1300 nm) potentially interesting for applications. A dominant absorption peak is present around 1200 nm. All major tissue constituents (i.e., water, lipid, and collagen) contribute to it. Thus, it is potentially interesting for the assessment of tissue composition, but it might cause exceeding attenuation in some practical cases. However, slightly shorter wavelengths (i.e. 1100-1150 nm) corresponding to the raising edge of the peak, might allow an accurate estimate of tissue composition, with the advantage of much lower attenuation.

A near-infrared (NIR) tomography system with spectral-encoded sources at two wavelength bands was built to quantify the temporal contrast at 20 Hz bandwidth, while imaging breast tissue. The NIR system was integrated with a magnetic resonance (MR) machine through a custom breast coil interface, and both NIR data and MR images were acquired simultaneously. MR images provided breast tissue structural information for NIR reconstruction. Acquisition of finger pulse oximeter (PO) plethysmogram was synchronized with the NIR system in the experiment to offer a frequency-locked reference. The recovered absorption coefficients of the breast at two wavelengths showed identical temporal frequency as the PO output, proving this multi-modality design can recover the small pulsatile variation of absorption property in breast tissue related to the heartbeat. And it also showed the system's ability on novel contrast imaging of fast flow signals in deep tissue.

Tumor hypoxia is an important indicator of tumor metabolism and tumor response to various forms of therapy. Currently, no imaging modality exists that can directly map tumor hypoxia non-invasively. We present an ultrasound guided diffuse optical imaging technique for precisely measuring the tumor oxygenation. The approach employs ultrasound structural information as a-prior knowledge for diffuse optical imaging. Hypoxia mapping is achieved using endogenous chromophores such as oxy- and deoxy- hemoglobin in the tissue. Because oxy- and deoxy- hemoglobin respond differently at different wavelengths, four different laser diodes of wavelengths 740 nm, 780 nm, 808 nm and 830 nm were used for mapping tumor hypoxia by diffuse optical imaging. Hypoxia model experiments were performed using phantoms at different oxygenation conditions (Hemoglobin oxygen saturation: 14%-92%) representing the hemoglobin oxygenation range in tissue. Targets of different sizes mimicking different development stages of breast tumors, 1.0 cm to 2.5 cm diameter in 0.5 cm steps, were tested to validate the oxygen saturation measurement accuracy with target size. The absolute deviations between the estimated hemoglobin oxygen saturations from absorption maps and oxygen measurements obtained using a pO₂ electrode were less than 8% over the measured range of oxygen saturations (14% - 92%). An inhomogeneous cocentric blood phantom of deoxygenated center core and oxygenated outer shell was imaged and deoxy- and oxy- hemoglobin maps revealed corresponding distributions which correlate with inhomogeneous deoxy- and oxy- distributions frequently seen in advanced breast cancers located in the depth range of 1-3 cm.

Hyperspectral excitation-resolved fluorescence tomography (HEFT) exploits the spectrally-dependent absorption properties of biological tissue for recovering the unknown three-dimensional (3D) fluorescent reporter distribution inside tissue. Only a single light source with macro-illumination and wavelength-discrimination is required for the purpose of light emission stimulation and 3D image reconstruction. HEFT is built on fluorescent sources with a relatively broad spectral absorption profile (quantum dots) and a light propagation model for strongly absorbing tissue between wavelengths 560 nm and 660 nm (simplified spherical harmonics - SPN, - equations). The measured partial current of fluorescence light is cast into an algebraic system of equations, which is solved for the unknown quantum dot distribution with an expectation-maximization (EM) method. HEFT requires no source-detector multiplexing for 3D image reconstruction and, hence, offers a technologically simple design.

We are reporting the first trial image reconstruction of a implanted fluorescent target into a live rat abdomen. We use a simplified algorithm for fluorescence diffuse optical tomography (FDOT), so-called the Total-light algorithm to obtain the absorption image of the target from the measured mean-transit time (MTT). We reconstructed two absorption images with and without a fluorescence target. It is difficult to identify something in the absorption images. However, the difference image between the two images highlights the target. This suggests that our algorithm is robust to the artifacts in the images in the real situation of in vivo measurements.

We present tumor hypoxia mapping by diffuse optical fluorescence tomography. A novel 2-nitroimidazole bis-carboxylic acid indocyanine dye conjugate has been developed for tumor-targeted hypoxia fluorescence imaging. The hypoxia probe has been evaluated in-vitro using 4T1 tumor cell lines and in-vivo tumor targeting in mice. In-vivo tumor targeting in six mice demonstrated that a measured half-life of 2-nitroimidazole-indocyanine dye wash out in the tumor was significantly longer (112±32.37 minutes) than that of bis-carboxylic acid indocyanine dye (69.75±14.01 minutes). The bis-carboxylic acid indocyanine dye was completely washed out from the tumor site within 3-5 hours post-injection, but 2-nitroimidazole-ICG remained for up to 21 hours in the tumor site. Near infrared fluorescence images of mice tumors showed a 2.6-fold contrast of dye uptake with hypoxic conjugate injection (7.46±1.68 μM) compared to that with indocyanine dye injection (2.9±0.60 μM). The in-vitro cell studies were performed to assess fluorescence labeling comparing hypoxia to normoxia conditions. A fluorescence emission ratio of 2.5-fold was found between the cells treated with the 2-nitroimidazole-indocyanine dye and incubated under hypoxia compared to the cells in normoxia condition. Hypoxia specificity was also confirmed when compared to cells treated with unconjugated indocyanine dye alone. Fluorescence images acquired using a Li-COR scanner from harvested tumor samples support the in vivo monitoring and imaging results.

Fluorescence imaging in diffusive media locates tumors tagged by injected fluorescent markers in NIR wave-lengths. For deep embedded markers, natural autofluorescence of tissues comes to be a limiting factor to tumor detection and accurate FDOT reconstructions. A spectroscopic approach coupled with Non-negative Matrix Factorization source separation method is explored to discriminate fluorescence sources according to their fluorescence spectra and remove unwanted autofluorescence. We successfully removed autofluorescence from acquisitions on living mice with a single subcutaneous tumor or two capillary tubes inserted at different depths.

The superior sensitivity and dynamic range of photodetection using time-correlated single-photon counting (TCSPC) technology can extend the diameter of specimens that can be imaged with fluorescence molecular tomography (FMT) and extend the minimum fluorescence concentration that can be reconstructed. To test these limits, a multi-modal system combining microcomputed tomography (microCT) and time-domain TCSPC FMT was used to image a 5 cm-diameter tissue-simulating cylindrical phantom containing an inclusion of various concentrations of AlexaFluor 647, mixed with 1% intralipid in water. This was repeated at lower concentrations for a 2.5 cm-diameter cylindrical phantom. Fluorescence tomography images were reconstructed using the structural information obtained from the microCT (outer surface of the interrogated sample) as prior information for light transport modeling. Results are presented showing that the location of the fluorescent inclusion can be reconstructed down to nanomolar concentrations in the 5 cm phantom and down to sub-nanomolar concentrations in the 25 mm phantom. The ability to reconstruct fluorescence images of these phantoms demonstrates that an unprecedented level of sensitivity can be achieved with time-domain TCSPC fluorescence tomography allowing this technology to be used for applications involving animals larger than mice as well as for applications where limited contrast is available.

Quantum dots (QDs) are widely used in fluorescence tomography due to its unique advantages. Despite the very high quantum efficiency of the QDs, low fluorescent signal and autofluorescence are the most fundamental limitations in optical data acquisition. These limitations are particularly detrimental to image reconstruction for animal imaging, e.g., free-space in vivo fluorescence tomography. In animals studies, fluorescent emission from exogenous fluorescent probes (e.g. QDs) cannot be effectively differentiated from endogenous broad-spectral substances (mostly proteins) using optical filters. In addition, a barrow-band fluorescent filter blocks the majority of the fluorescent light and thus makes signal acquisition very inefficient. We made use of the long fluorescent lifetime of the QDs to reject the optical signal due to the excitation light pulse, and therefore eliminated the need for a fluorescent filter during acquisition. Fluorescent emission from the QDs was excited with an ultrafast pulsed laser, and was detected using a time-gated image intensifier. A tissue-simulating imaging phantom was used to validate the proposed method. Compared to the standard acquisition method that uses a narrow-band fluorescent filter, the proposed method is significantly more efficient in data acquisition (by a factor of >10 in terms of fluorescent signal intensity) and demonstrated reduction in autofluorescence. No additional imaging artifact was observed in the tomographic reconstruction.

Model based analysis of calibrated diffuse reflectance spectroscopy can be used for determining oxygenation and concentration of skin chromophores. This study aimed at assessing the effect of including melanin in addition to hemoglobin (Hb) as chromophores and compensating for inhomogeneously distributed blood (vessel packaging), in a single-layer skin model. Spectra from four humans were collected during different provocations using a twochannel fiber optic probe with source-detector separations 0.4 and 1.2 mm. Absolute calibrated spectra using data from either a single distance or both distances were analyzed using inverse Monte Carlo for light transport and Levenberg-Marquardt for non-linear fitting. The model fitting was excellent using a single distance. However, the estimated model failed to explain spectra from the other distance. The two-distance model did not fit the data well at either distance. Model fitting was significantly improved including melanin and vessel packaging. The most prominent effect when fitting data from the larger separation compared to the smaller separation was a different light scattering decay with wavelength, while the tissue fraction of Hb and saturation were similar. For modeling spectra at both distances, we propose using either a multi-layer skin model or a more advanced model for the scattering phase function.

Peripheral Artery Disease (PAD) affects over 10 million Americans and is associated with significant morbidity and mortality. While in many cases the ankle-brachial index (ABI) can be used for diagnosing the disease, this parameter is not dependable in the diabetic and elderly population. These populations tend to have calcified arteries, which leads to elevated ABI values. Dynamic optical tomography (DDOT) promises to overcome the limitations of the current diagnostic techniques and has the potential to initiate a paradigm shift in the diagnosis of vascular disease. We have performed initial pilot studies involving 5 PAD patients and 3 healthy volunteers. The time traces and tomographic reconstruction obtained from measurements on the feet show significant differences between healthy and affected vasculatures. In addition, we found that DOT is capable of identifying PAD in diabetic patients, who are misdiagnosed by the traditional ABI screening.

Implanted Cardioverter Defibrillator (ICD) provides one of the most effective therapies for the prevention of sudden cardiac death, but also delivers some high voltage shocks inappropriately, causing morbidity and mortality. Implanted near-infrared spectroscopy (NIRS) may augment ICD arrhythmia detection by monitoring skeletal muscle perfusion. A two-wavelength, single-distance, continuous-wave implanted NIRS has been evaluated in-vivo. A weighted difference of the changes in attenuation at two wavelengths, across the isobestic point of the hemoglobin spectra, was taken to be the microvascular oxygenation trend indicator (O2 Index). Although the exact weight depends on the local vascular distribution and their oxygen levels, the hypothesis that a constant weight may be adequate for hemodynamic trending during short arrhythmic episodes, was tested. The sensor was implanted subcutaneously both on fresh tissue and inside scar tissue that formed around a pre-existing implant, in 3 animals each. Attenuations were recorded at 660 and 890 nm during normal sinus rhythm (NSR) and induced ventricular fibrillation (VF). The slope of the O2 Index over 10 seconds was computed for 7 NSR and 8 VF episodes in fresh and 13 NSR and 15 VF episodes in scar tissue pockets. The mean O2 Index slope was significantly different (p<0.0001) between NSR and VF rhythms for both the fresh and scar tissue pockets. Therefore implanted NIRS may be useful for preventing inappropriate detection of VF during electromagnetic interference, double counting of ECG T-wave as an R-wave, ICD lead failure, electrocardiographic aberrancy etc.

The angular filter array (AFA) is a silicon micro-machined optical collimator, which only accepts photons propagating within a narrow solid angle. It can be used to select photons exiting an imaging sample along a specific direction. This paper describes a novel Angular Domain Spectroscopic Imaging (ADSI) technique that utilizes deep illumination from the front surface of the sample and a camera with an AFA to image features embedded inside a turbid medium. This approach permitted spectroscopic imaging of turbid samples too thick to be imaged in a trans-illumination setup. The tissue-mimicking test phantom contained three groups of Indocyanine Green doped inclusions at depths from 1 to 3 mm embedded within an Intralipid^TM/agarose gel. The sample was scanned across the AFA and the intensity of the back scattered light along the direction normal to the surface was acquired as a function of location and wavelength. The resultant spectral images were captured and analyzed. The experiments demonstrated that ADSI could detect subsurface features that differed in wavelength-dependent absorption and/or scattering properties from the surrounding medium with the deep illumination configuration. Deep illumination ADSI may be useful as a non-invasive tissue imaging tool.

Broad-band light reflectance spectroscopy (LRS) of tissue with sub-millimeter fiber optic probes in the visible and nearinfrared range has shown its utility in differentiation of tissue types, identification of cancer, and measurement of stimulus-induced physiological responses. So far, single point measurement set-up has been widely employed to determine local optical properties of tissue. However, it may often be of interest to obtain a 2D map of a surface area of a tissue under investigation, rather than a single point reading, as in case of cancer margin detection or intraoperative perfusion measurement. It is thus imperative to expand the LRS technique to multipoint measurement covering a larger surface area. Here we describe the two methods that we use to quantify the hemoglobin derivatives and scattering of tissue under investigation, and then utilize two bifurcated fiber optic probes with different fiber diameters and different source-detector separations, to demonstrate the 2D imaging capability of LRS technique. In this study, we constructed a tissue phantom, simulating tissue and blood vessel, and used 2D scanning to determine the spatial resolution and depth resolution using two different probe geometries. Our results suggest that the depth sensitivity of these probes was limited to sub-millimeter for hemoglobin derivatives, whereas scattering changes could be observed up to 2mm deep. It was also found that the lateral resolution was affected, and the scattering signal became more diffuse, as a function of depth.