We introduce a set of corrections to the integral equations of scattering theory within the diffusion approximation to the radiative transport equation. We use this result to obtain an image reconstruction algorithm for optical tomography with spatial resolution below the transport mean free path.

A new algorithm based on multi-static data and vector subspace classification to eigenvectors of a round-trip matrix is introduced for optical imaging and localization of objects embedded in a turbid medium. The transport of light from multiple sources through excitation of the embedded objects to the array of detectors is represented by a response matrix that can be constructed from experimental data. The 'round-trip (RT) matrix’ is constructed by multiplying the response matrix by its transpose for continuous-wave (adjoint matrix for frequency domain) illumination. Mathematically, the RT matrix is equivalent to transfer of light from the sources via the embedded objects to the array of detectors and back, and is similar to the time-reversal matrix used in the general area of array processing for acoustic and radar time-reversal imaging. The eigenvectors with leading non-zero eigenvalues of the RT matrix correspond to embedded objects, which are orthogonal to the vectors in the noise subspace. The vector subspace method along with Green’s functions calculated from an appropriate model for light propagation through turbid media is then used to determine the locations of the embedded objects. We tested this algorithm in simulation for light transmitting through a 50 l _tr thick (l _tr ~ 1 mm is transport mean free path) parallel slab turbid medium with up to six embedded absorptive objects. The method was able to globally locate all six objects with surprising accuracy. This “round-trip tomographic imaging” approach is fast, applicable to different geometries and to different forward models.

Polarized light scattering spectroscopy (LSS) is sensitive to the cell nuclear morphological changes in the various forms of epithelial dysplasia. Extensive studies illustrate it is a promising in situ technique to detect precancerous and early cancerous changes in the epithelial tissue. To determine the density and size distribution of cell nuclei with spectra, generally, Mie theory-based inverse model is adopted. This model is of multiple parameters, multiple extreme values and nonlinear. The determination of all unknown parameters needs to solve a nonlinear inverse problem. Other than least-square fitting used by previous studies, in this paper, we developed a novel method - float genetic algorithm (FGA) to determine the particle size distribution and refractive index for LSS. Our results showed that, relative errors of three estimated statistical quantities: diameter, standard deviation and refractive index are less than 5% for different additive Gauss noise levels with 70 iteration epochs. The errors gradually decrease with iteration epoch increases. Moreover, comparing with Newton-type iteration method coupled with a Marquardt-Tikhanov regularization scheme, FGA avoids the problems of local extreme value and selection of initial value and regularization parameters, thus obtains the advantages of high precision, stability and robustness.

A number of studies indicate that compartmental modeling of indocyanine green (ICG) pharmacokinetics, as measured by near infrared (NIR) techniques, may provide diagnostic information for tumor differentiation. However, compartmental parameter estimation is a highly non-linear problem with limited data available in a clinical setting. Furthermore, pharmacokinetic parameter estimates show statistical variation from one data set to another. Thus, a systematic and robust approach is needed to model, estimate and quantify ICG pharmacokinetic parameters. In this paper, we propose to model ICG pharmacokinetics in extended Kalman filtering (EKF) framework. EKF effectively models multiple-compartment and multiple-measurement systems in the presence of measurement noise and uncertainties in model dynamics. It provides simultaneous estimation of pharmacokinetic parameters and ICG concentrations in each compartment. Moreover, recursive nature of the Kalman filter estimator potentially allows real time monitoring of time varying pharmacokinetic rates and concentration changes in different compartments. We tested our approach using the ICG concentration data acquired from four Fischer rats carrying adenocarcinoma tumor cells. Our study indicates that EKF model may provide additional parameters that may be useful for tumor differentiation.

We present a multi-dimensional TCSPC technique that combines multi-detector and multiplexing capability, and records fast and virtually unlimited sequences of time-of-flight distributions. The system consists of four fully parallel TCSPC channels. Each channel records simultaneously in up to eight detection channels. Up to four lasers and 32 source positions can be multiplexed. The total count rate is up to 4 x 10⁷ photons per second. Time-of-flight sequences can be recorded with a resolution of 50 to 100 ms per curve. The system is operated within a single personal computer.

Optical techniques based on photon migration are rapidly emerging as a promising alternative and/or augmentation of existing medical imaging modalities. For example, real time studies of hemodynamic changes in brain tissue are possible as a step towards optical functional brain imaging. Time-resolved implementations of these techniques allow for discrimination between scattering and absorption and for depth resolution. They require sub-nanosecond pulsed light sources with high repetition rate and sufficient power for deep enough tissue penetration. Picosecond diode lasers satisfy the clinical demands of economy, compact size, and reliability almost perfectly. Today multi-channel diode laser devices are commercially available and are widely used in diffuse optical imaging and spectroscopy, in particular in optical tomography and breast cancer detection. However, the output powers of these devices are just about sufficient for moderate tissue penetration depths. An improvement that does not compromise the advantages of the diode laser sources is amplification of the diode laser output by means of solid state tapered amplifiers. We present an amplified light source for use in NIR diffuse optical spectroscopy and imaging, providing pulse widths as short as 100 ps, adjustable repetition rates up to 80 MHz, and peak power levels as high as 7 Watts, corresponding to average power levels exceeding 100 mW. In combination with time-resolved photon counting electronics matching the high throughput demands in conjunction with the new source, state-of-the-art systems for diffuse optical imaging can be built. System design features and possible application examples are presented.

We report on an instrument for time-resolved spectroscopy (TRS) based on white-light generation in a highly non-linear crystal fiber. TRS in the visible and near-infrared region at picosecond-to-nanosecond time scales has attracted increased interest in recent years owing to the possibility of spectroscopic analysis of turbid media, such as biological tissues. A self-mode-locked Ti:Sapphire oscillator pumped by an Ar:ion laser provides pulses 50 - 100 fs long, at 85 MHz repetition rate. The light is focused into a crystal fiber, which consists of a core surrounded by a mesh of air-filled holes. White light is generated by a combination of several non-linear effects in the fiber. We optimize the spectrum for measurements in the region 600 - 1000 nm. For detection, we use an imaging spectrometer coupled to a 16-channel photomultiplier tube, enabling simultaneous detection in 16 wavelength bands. We use time-correlated single-photon counting to record the signal, with a temporal resolution of ~160 ps. To demonstrate the system, we have performed measurements of the diffuse time-resolved reflectance of tissue phantoms made of epoxy resin with added scattering and absorbing materials. The data was evaluated using a light propagation model based on diffusion theory, to extract the scattering and absorption coefficients of the medium. The results corresponded very well with previous measurements on the phantoms performed using other TRS instruments.

Until recently, optical mapping of electrical activity in the heart muscle using voltage-sensitive dyes has mainly been applied to subsurface imaging. Here we present a method for the three-dimensional (3D) reconstruction of electrical activity deep inside the myocardial wall. We propose an alternative approach to diffusive optical tomography, based on ideas from binocular vision. Detection and illumination occur on opposite sides of the preparation. Staining with absorptive voltage-sensitive dyes is assumed. Data acquisition follows a paraxial scanning procedure, which modifies coaxial scanning by the introduction of a vector offset between illumination and detection axes. Pairs of 2D images are obtained corresponding to offsets of opposite signs. Those image pairs created by parallax are used as an input for the reconstruction algorithm, whose output is a 3D optical image of intramural electrical excitation. We apply this method to the slab geometry. The procedure was tested for a variety of computer-generated sources including particles, lines, bubbles, and simulated electrophysiological patterns such as scroll waves. The limitations of the method and possible improvements are discussed.

In this paper, we report simultaneous reconstruction of absorption and scattering heterogeneities using a dual mesh scheme based on finite element method (FEM). Column normalization has been applied to the weight matrix obtained from FEM forward model to correct the depth dependent problem and to alleviate the crosstalk between the absorption coefficient and scattering coefficient. With this approach, phantom targets with both absorption and scattering heterogeneities can be reconstructed with good contrast and resolution. With this approach, the contrast between malignant breast cancers and benign lesions can be further improved compared with that obtained from the modified Born approximation, where the bulk reduced scattering coefficient has been used for reconstructing absorption heterogeneities.

The image resolution and contrast in Near-Infrared (NIR) tomographic image reconstruction is in part affected by the number of available boundary measurements. In the presented study, singular-value decomposition (SVD) of the Jacobian has been used to find the benefit of the total number of measurements that can be obtained in a two-dimensional (2D) and three-dimensional (3D) problem. Reconstructed images show an increase in improvement in the reconstructed images when the number of measurements are increased, with a central anomaly showing more improvement as compared to a more superficial one. It is also shown that given a 2D model of the domain, the increase in amount of useful data drops exponentially with an increase in total number of measurements. For 3D NIR tomography use of three fundamentally different data collection strategies are discussed and compared. It is shown that given a 3D NIR problem, using three planes of data gives more independent information compared to the single plane of data. Given a three planes of data collection fibers, it is shown that although more data can be collected in the out-of-plane data collection strategy as compared to the only in-plane case, the addition of new data does not increase image accuracy dramatically where as it increases the data collection and computation time.

With the advent of small animal imaging systems, it has become possible to non-invasively monitor the progression of diseases in living small animals and study the efficacy of drugs and treatment protocols. Magnetic resonance imaging (MRI) is an established imaging modality capable of obtaining high resolution anatomical images as well as studying cerebral blood volume (CBV), cerebral blood flow (CBF), and cerebral metabolic rate of oxygen (CMRO2). Optical tomography, on the other hand, is an emerging imaging modality, which, while much lower in spatial resolution and insensitive to CBF, can separate the effects of oxyhemoglobin, deoxyhemoglobin, and CBV with high temporal resolution. In this study we present our first results concerning coregistration of MRI and optical data. By applying both modalities to imaging of kidney tumors in mice that undergo VEGF treatment, we illustrate how these imaging modalities can supplement each other and cross validation can be performed.

To investigate the dependence of image features on the spatial distribution of tissue optical parameters, we have developed a parallel Monte Carlo code that is capable of simulating light propagation and distribution in heterogeneous tissue phantoms. With a parallel cluster of 16 processing elements running on a Linux operating system, this code can generate reflectance images with negligible statistical fluctuation by tracking 8x10⁸ photons in less than 30 minutes. Using this code, the effect of skin bulk optical parameters of μ _s, μ _a and g and their heterogeneous distributions on the reflectance image have been analyzed.

The characteristic lengths for linear and circular polarized light to depolarize have been studied by an analytical treatment of random walk of vector photons in turbid media. The depolarization lengths for linear and circular polarized light transmission through a slab are derived. A quantitative understanding of light depolarization behavior in turbid media comprised of Mie scatterers of arbitrary size and refractive index are presented.

It is well recognized that the spectral characteristics of light scattered from living tissue can provide valuable diagnostic information. In order to address the gap in the understanding of light scattering by complex cellular and tissue structures, we developed analytical and computational methods to characterize light scattering signals from irregular shapes. Recently, we investigated the total-scattering-cross-section (TSCS) spectra of complicated geometries based on the finite-difference-time-domain (FDTD) simulations. We found that the TSCS spectra of many inhomogeneous and nonspherical particles can be approximated with those of their best-fitting homogeneous ellipsoidal counterparts, and calculated using a simple formula provided by the equiphase-sphere approximation. Furthermore, we have characterized backscattering spectra of inhomogeneous particles with stochastic distribution of interior refractive index. We have investigated the backscattering signals of a wide range of inhomogeneous micro-particles based on the FDTD method and the Gaussian Random Field model. Our numerical results indicate that, contrary to the TSCS, the backscattering spectrum is sensitive to small structures within a particle, with scales down to tens of nanometers. We also note that the spectroscopic properties of the backscattering signals are directly linked to the coherence length of the interior refractive index distribution, which characterizes the texture of the inhomogeneous particle. The implication of this study is the possibility of using spectroscopic techniques to detect cellular morphological and textural changes within scales several-times smaller than the wavelength.

In this work, 2D and 3D distribution of exit angles of diffuse-reflected photons from a semi-infinite, homogeneous slab are obtained by Monte-Carlo simulations. During simulations, a single point source emits incident photons normally onto the slab. For the calculation of exit angles, we consider the exit angles from a hemisphere with a varying radius. Hemispheres are distributed on the upper side of the slab for reflectance geometry. In both 2D and 3D cases, only photons exiting the slab through bases of hemispheres are considered. As a photon intersects the hemisphere surface, exit angles of the photon with respect to a spherical coordinate system whose origin is on the center of base of the hemisphere are recorded. In the 2D case, projection of the direction of exit of a photon onto upper surface of the slab is tallied hence only a single exit angle (polar angle) with respect to a polar coordinate system is considered. In the 3D case however, two exit angles, namely polar and azimuth angles, are recorded. Our results show that, in both 2D and 3D distribution of exit angles of diffuse-reflected photons have similar patterns for all hemispheres. The distributions of exit angles are observed to aggregate on angles away from the source. This property is preserved independent of the hemisphere position with respect to the point source.

We have developed a model-based iterative image reconstruction scheme based on the equation of radiative transfer in the frequency domain for the applications in small animal optical tomographic imaging. To test the utility of such a code in small animal imaging we have furthermore developed a numerical phantom of a mouse. In simulation studies using this and other phantoms, we found that to make truly use of phase information in the reconstruction process modulation frequencies well above 100 MHz are necessary. Only at these higher frequencies the phase shifts introduced by the lesions of interest are large enough to be measured. For smaller frequencies no substantial improvements over steady-state systems are achieved in small geometries typical for small animal imaging.

The chest-wall layer underneath the breast tissue consists of muscle layer and induces distortion to measured near infrared diffused wave when the patient is imaged in the supine position. In this paper, we present results of using a simple two-layer model to correct the chest wall induced distortion. Four parameters of absorption and reduced scattering coefficients of both layers are used to describe the optical properties of the model. With the initially estimated absorption and reduced scattering coefficients, an iterative search method is used to find the best fitted parameters to minimize the difference between the measurements obtained at normal breast region and the model data. Then, a correction method is applied to correct the chest wall mismatch between the lesion site and reference site. With this correction scheme, phantom targets located on top of the chest-wall phantom layer can be reconstructed with good contrast and resolution. With the a priori chest wall depth information obtained from ultrasound at both normal and lesion regions, the contrast between malignant breast cancers and benign lesions can be further improved compared with that obtained from the modified Born approximation, where semi-infinite boundary is used.

While near-infrared tomography has advanced considerably over the past decade, key technological designs still limit what can be achieved, especially in terms of imaging acquisition speed. One of these fundamental limitations is the requirement that the source light be delivered sequentially or through frequency encoding of the time signal. Sequential delivery inherently limits the speed at which images can be acquired. Modulation frequency-dependent encoding of the sources solves the problem by allowing sources near the same location to be turned on simultaneously, thereby improving the speed for acquisition, but suffers from dynamic range problems. In this study, we demonstrate an alternative parallel source implementation approach which uses spectral wavelength encoding of the source. This new technique allows many sources to be input into the tissue at the same time, as long as the spectrally encoded signals can be decoded at the output. To test the implementation of this approach, 8 single-mode laser diodes of wavelengths distributed within a narrow range of 10 nm are used, and the lights are all input into tissue phantom simultaneously. On the detection side, a high-resolution spectrometer is used to spatially spread out the signals to facilitate parallel detection of the signal from each spectrally-encoded source. This robust approach allows rapid parallel sampling of all sources at all detection locations. The implementation of this technique in a NIR tomography application is examined, and the preliminary results of video-rate imaging at 30 Hz is presented.

We present a newly developed optical mammograph for concurrent diffuse optical and MR imaging of the compressed breast. A home-built MR coil allows optical measurements to be carried out along mediolateral as well as craniocaudal projections. Preliminary in-vivo experimental data are presented.

We study the scattering of diffused photon pair density wave (DPPDW) from a spherical and cylindrical inhomogeneity embedded in a homogeneous multiple scattering medium. In this study, DPPDW is composed of correlated polarized photon pairs at different temporal frequencies and parallel linear polarized states in a multiple scattering medium where two linear polarized photons are common-path propagating in the scattering medium. The heterodyne signal is then generated by a photomultiplier tube. By measuring the amplitude attenuation and the phase delay of heterodyne signal at different position of the scattering medium, the diffracted amplitude and phase wavefront of DPPDW are obtained precisely and simultaneously. In this experiment, the distorted phase wavefront and amplitude wavefront by a perfect spherical or cylindrical absorber in a multiple scattering medium is investigated. Because the feature of common-path propagation of polarized pair photons in a multiple scattering medium and the polarization gating and the spatial coherence gating working simultaneously, the sensitivity of the amplitude and phase detection of DPPDW are then enhanced significantly. The ability of detecting a smaller size of optical inhomogeneity in a multiple scattering medium is discussed and the experimental results consistent with the theoretical expectation are demonstrated.

Near-infrared spectroscopy (NIRS) has been used for functional brain imaging by employing properly designed source-detector matrices. We demonstrate that by embedding a NIRS source-detector matrix within an electroencephalography (EEG) standard multi-channel cap, we can perform functional brain mapping of hemodynamic response and neuronal response simultaneously. In this study, the P300 endogenous evoked response was generated in human subjects using an auditory odd-ball paradigm while concurrently monitoring the hemodynamic response both spatially and temporally with NIRS. The electrical measurements showed the localization of evoked potential P300, which appeared around 320 ms after the odd-ball stimulus. The NIRS measurements demonstrate a hemodynamic change in the fronto-temporal cortex a few seconds after the appearance of P300.

We present concurrent NIRS-fMRI measurements on a human subject during a finger tapping test. The optical data were collected with a frequency domain experimental apparatus (ISS, Inc., Champaign IL) comprising sixteen laser sources at 690 nm, sixteen laser sources at 830 nm and four photomultiplier tube detectors. The lasers were coupled to optical fibers that led the light onto the subject's head. A special optical helmet (fMRI-compatible) with a retractable and resilient set of optical fibers was devised to improve the coupling between the fibers and the scalp. The fMRI data were collected with a 3 Tesla Siemens Trio magnetic resonance scanner and a quadrature birdcage radiofrequency coil. The spatial and temporal comparison of the fMRI and NIRS signals associated with brain activation showed a very good agreement, confirming the role of NIRS as a reliable brain monitor for functional studies.

Functional Near Infrared Spectroscopy has been used to investigate changes in cerebral hemodynamics induced by hypercapnia challenges, such as carbon dioxide CO₂ inhalation and breath holding. The aim of this study was to investigate CO₂ pressure changes dependence of frequency spectrum of cerebral hemodynamic oscillations during breath holding task. Measurements of the relative changes in concentration of deoxy-hemoglobin ([Hb]) and oxyhemoglobin ([HbO₂]) are performed on nine healthy subjects during three breath holdings of 30 seconds (s.) interleaved with 90 s. of normal breathing. Power spectra are computed by continuous wavelet transform and averaged for normal and hold episodes. The percent change values between hold and normal episodes are given for frequency peaks at (0.035 Hz), where a 17% higher increase was observed for PC of [Hb] on the right side compared to left side,while this value was at 64.8% for [HbO₂] . Similarly, for a peak at 0.11 Hz these values were 54.5% and 9.5% for [Hb] and [HbO₂] PCs, respectively. The smallest changes were observed for breathing freq. range (around 0.2 Hz) where the values are -72% and 55.8% for [Hb] and [HbO₂], respectively.

Visually evoked hemodynamic responses and potentials were simultaneously measured using a 16-channel optical imaging instrument and a 60-channel electroencephalography instrument during normo-, hypo- and hypercapnia from three subjects. Flashing and pattern-reversed checkerboard stimuli were used. The study protocol included two counterbalanced measurements during both normo- and hypocapnia and normo- and hypercapnia. Hypocapnia was produced by controlled hyperventilation and hypercapnia by breathing carbon dioxide enriched air. Near-infrared imaging was also used to monitor the concentration changes of oxy- and deoxyhaemoglobin due to hypo- and hypercapnia. Hemodynamic responses and evoked potentials were successfully detected for each subject above the visual cortex. The latencies of the hemodynamic responses during hypocapnia were shorter whereas during hypercapnia they were longer when compared to the latencies during normocapnia. Hypocapnia tended to decrease the latencies of visually evoked potentials compared to those during normocapnia while hypercapnia did not show any consistent effect to the potentials. The developed measurement setup and the study protocol provide the opportunity to investigate the neurovascular coupling and the links between the baseline level of blood flow, electrical activity and hemodynamic responses in the human brain.

Fluorescence-enhance optical tomography is performed using (i) point illumination and point collection and (ii) area illumination and area collection geometrics in 3D. In both measurement techniques, an image-intensified charge-coupled (ICCD) imaging system is used in the frequency-domain to image near-infrared contrast agents. The experimental measurements are compared to diffusion model predictions in least squares form in the inverse problem. For image recovery for both area and point illumination geometries, an efficient gradient-based optimization technique based on the Penalty/modified barrier function (PMBF) method and the constrained truncated Newton with trust region (CONTN) method is developed. Targets in 3D were reconstructed from experimental data under two conditions of (i) perfect uptake (1:0, target to background ratio) and (ii) imperfect uptake (100:1, target to background ratio). Parameters of absorption cross section due to fluorophore and lifetimes are reconstructed. The present work demonstrates that 3D fluorescence enhanced optical tomography reconstructions can be successfully performed from both point/area illumination and collection experimental measurements of the time-dependent light propagation on clinically relevant tissue phantoms using a gain-modulated ICCD camera.

A discussion on recent works on diffusive inverse problems is presented with a special focus on three-dimensional imaging methods and their application to small animal imaging by fluorescence-enhanced Diffuse Optical Tomography. A numerical approach using the Finite Element Method for handling problems modelled by elliptic coupled partial differential equations is justified by the complexity of the geometry of the system but is known to be time- and memory-consuming. The resolution of the adjoint problem considerably speeds up the treatment and allows a full 3D resolution. Nevertheless, because of the ill-posedness of the problem, the reconstruction scheme is sensitive to a priori knowledge on the parameters to be reconstructed. In this study, a multiple step, self-regularized, reconstruction algorithm for the spatial distribution of the fluorescent regions is presented. We introduce the prior knowledge of the regions of interest via a segmentation of the results performed with a first rough reconstruction of the fluorescent regions. The results are then refined along iterations of the segmentation/reconstruction scheme.

A new approach for optical fluorescence tomographic imaging of targets in a turbid medium that uses the independent component analysis (ICA) from information theory is presented. Fluorescence signals from targets embedded in a turbid medium are measured on the boundary of the medium using a multi-source excitation and a multi-detector acquisition scheme. Differences between excitation and fluorescence wavelengths enable sensitive, minimal-background signal acquisition. ICA of the fluorescence signal on the medium boundary sorts out the embedded ob-jects, and their locations are obtained from Green’s function analysis based on any appropriate light propagation model. Fluorescence tomographic imaging experiments were carried out using Intralipid-10% suspension in water contained in a 50-mm thick rectangular transparent plastic cell as the turbid medium, and small glass spheres containing indocyanine green (ICG) solution as fluorescent targets. The near-infrared (NIR) fluorescence was excited using 785 nm light, and monitored over a narrow band around 830 nm. The transport mean free paths at 785 nm and 830 nm were 1.01 mm and 1.14 mm, respectively. The approach could image and determine the position of an ICG filled sphere of radius as small as 4 mm. It is applicable to small objects, different medium geometries, and amenable to near real time imaging applications.

We present a novel method for estimating the depth of a fluorescent lesion in tissue based on measurements of the fluorescence signal in different wavelength bands. The measured fluorescence spectrum following irradiation by excitation light at the surface is a function of several parameters, because the fluorescence light has to pass through tissue with characteristic scattering and absorption properties. Thus, the intrinsic fluorescence spectrum will be altered, in a way determined by the tissue optical properties, the depth of the fluorophore, and also by the geometry of the light irradiation and the detection system. By analyzing the ratio between the signals at two wavelengths we show that it is possible to estimate the depth of the lesion. We have performed Monte Carlo simulations and measurements on an Intralipid phantom in the wavelength range 850 - 1000 nm. By taking the ratio between the signals at the wavelengths 875 and 930 nm the depth of a fluorescing layer could be determined with 0.8 mm accuracy down at least a depth of 10 mm. Monte Carlo simulations were also performed for different tissue types with various composition. The results indicate that depth estimation of a lesion is possible with no assumptions made about the optical properties for a wide range of tissues.

We present a phased-array approach to diffuse optical imaging that is based on collecting continuous wave (CW) data using multiple sources and/or multiple detectors. The optical signals corresponding to individual source-detector pairs are combined by post processing after assigning individual amplitude and phase factors. Here, we demonstrate the enhancement in spatial resolution and the depth discrimination afforded by three-element linear phased arrays, either as stand alone units or in a 2-D crossed configuration. In this particular example, the (amplitude, phase) factors associated with the three array elements are (1,0), (2,π), (1,0), respectively. Experimental results on tissue like phantoms demonstrate the potential of the proposed source/detector arrays to enhance the performance of diffuse optical imaging.

In this paper we describe a novel integrated photonic system that can be applied for optical imaging of intracranial brain tumor (Glioblastoma) angiogenesis in small animal model. A non-invasive multi-modality approach based on near infrared spectroscopy (NIRS) technique namely: Steady State Diffuse Optical Spectroscopy (DOS) along with Magnetic Resonance Imaging (MRI) technique is applied for monitoring the concentration of oxyhemoglobin, deoxyhemoglobin and water within tumor region and for studying the vascular status of tumor and the physiological changes that occur during brain tumor angiogenesis.

Optical imaging and tomography in tissues can facilitate the quantitative study of several important chromophores and fluorophores in-vivo. Due to this fact, there has been great interest in developing imaging systems offering quantitative information on the location and concentration of chromophores and fluorescent probes. However, most imaging systems currently used in research make use of fiber technology for delivery and detection, which restricts the size of the photon collecting arrays leading to insufficient spatial sampling and field of view. To enable large data sets and full 360^o angular measurements, we developed a novel imaging system that enables 3D imaging of fluorescent signals in bodies of arbitrary shapes in a non-contact geometry in combination with a 3D surface reconstruction algorithm. The system consists of a rotating subject holder and a lens coupled Charge Coupled Device (CCD) camera in combination with a fiber coupled laser scanning device. An Argon ion laser is used as the source and different filters are used for the detection of various fluorophores or fluorescing proteins. With this new setup a large measurements dataset can be achieved while the use of inversion models give a high capacity for quantitative 3D reconstruction of fluorochrome distributions as well as high spatial resolution. The system is currently being tested in the observation of the distribution of Green Fluorescent Protein (GFP) expressing T-lymphocytes in order to study the function of the immune system in a murine model.

Hybrid NIR-MRI imaging has been used in a clinical breast imaging system to characterize breast tissue properties. The multi-spectral, frequency-domain tomography system operates inside a clinical scanner via long silica-glass optical fiber bundles and using a non-magnetic fiber-patient interface attached to a high resolution MR breast coil. Sixteen fiber bundles are positioned around the circumference of the female breast yielding 240 measurements of light transmission (amplitude and phase) at six optical wavelengths from 660-850nm through up to 12 cm of tissue. From optical measurements, we use a Newton-type algorithm to reconstruct images of tissue optical properties (absorption and scattering), and physiological tissue features such as oxy-hemoglobin [Hb-O₂], deoxy-hemoglobin concentrations [Hb-R], water concentration [water], scattering amplitude, and scattering power. We are exploring the synergistic benefits of a combined NIR-MRI data set, specifically the ways in which MRI (i.e. high spatial resolution) can be used to enhance NIR (i.e. high contrast resolution) image reconstruction. A priori knowledge can be applied to image reconstruction in the form of spatial and spectral constraints to improve spatial resolution, contrast, and quantitative accuracy of NIR images. In vivo results suggest that this combined system can accurately quantify contrast between the properties of tissues present in the breast (i.e. adipose and fibroglandular) regardless of their varied and complex spatial organization. For a group of healthy female volunteers, we observe greater contrast between the properties of adipose and glandular tissues when we use MR-guidance than when we do not, and values of total hemoglobin and water content are more consistent with what is physiologically expected.

In this paper, we propose a partially reflecting boundary method for diffusive optical imaging of shallow targets in turbid media. An appropriate range of the effective reflection coefficient R_eff has been suggested between 0.0 and 0.7, in which the extrapolated boundary condition can be used to simplify the forward model. For a shallow target of approximately 0.5 cm deep from the surface, experimental data acquired with a reflecting boundary of R_eff ≈ 0.6 lead to significant improvement in the image quality compared with that from an absorbing boundary of R_eff ≈ 0.

NIR handheld LED breast cancer imaging is a novel design that we are using widely in clinical study. During the test, we found the pressure change may affect the signal from the breast, especially to the blood flow inside of tissue. In this paper, we will show a new design of the probe combined with pressure sensors. And, we also did 15 cases of normal person’s breast test by using this design. We try to dig out how the pressure works and the relationships between the pressure effect to blood flow, also oxygen saturation in vivo. Some patients’ data may be presented here too.

In this report, clinical examples of using combined ultrasound and optical diffused wave technique to image tumor total hemoglobin concentration and tumor hypoxia are given. These examples demonstrate that the sensitivity and specificity of using tumor hemoglobin level as diagnostic index are much higher than that of tumor hypoxia.

We previously suggested that the two time constants quantified from the increase of tumor oxyhemoglobin concentration, ▵ [HbO₂], during hyperoxic gas intervention are associated with two blood flow/perfusion rates in well perfused and poorly perfused regions of tumors. In this study, our hypothesis is that when cancer therapy is applied to a tumor, changes in blood perfusion will occur and be detected by the NIRS. For experiments, systemic chemotherapy, cyclophosphamide (CTX), was applied to two groups of rats bearing syngeneic 13762NF mammary adenocarcinomas: one group received a single high dose i. p. (200 mg/kg CTX) and the other group continuous low doses (20 mg/kg CTX i. p. for 10 days). Time courses of changes in tumor ▵ [HbO₂] were measured at four different locations on the breast tumors non-invasively with an inhaled gas sequence of air-oxygen-air before and after CTX administration. Both rat body weight and tumor volume decreased after administration of high dose CTX, but continuous low doses showed decrease of tumor volume only. Baselines (without any therapy) intra- and inter-tumor heterogeneity of vascular oxygenation during oxygen inhalation were similar to our previous observations. After CTX treatment, significant changes in vascular hemodynamic response to oxygen inhalation were observed from both groups. By fitting the increase of ▵ [HbO₂] during oxygen inhalation, we have obtained changes of vascular structure ratio and also of perfusion rate ratio before and after chemotherapy. The preliminary results suggest that cyclophosphamide has greatest effect on the well perfused tumor vasculature. Overall, our study supports our earlier hypothesis, proving that the effects of chemotherapy in tumor may be monitored non-invasively by using NIRS to detect changes of hemodynamics induced with respiratory challenges.

A multi-spectral direct chromophore and scattering reconstruction for frequency domain NIR tomography has been implemented using constraints of the known molar spectra of the chromophores and a Mie theory approximation for scattering. This was tested in a tumor-simulating phantom containing an inclusion with higher hemoglobin, lower oxygenation and contrast in scatter. The recovered images were quantitatively accurate and showed substantial improvement over existing methods; and in addition, showed robust results tested for up to 5% noise in amplitude and phase measurements. When applied to a clinical subject with fibrocystic disease, the tumor was visible in hemoglobin and water, but no decrease in oxygenation was observed, making oxygen saturation, a potential diagnostic indicator.

We measured the changes of oxy-hemoglobin (Δ[HbO2]) and deoxy-hemoglobin concentration (Δ[Hb]) in rat breast 13762NF tumors with respect to oxygen or carbogen inhalation using near-infrared spectroscopy (NIRS). The changes in tumor blood flow can be estimated from the NIRS data provided with certain model assumptions. In the theoretical approach, we modified the Windkessel model so as to associate the mathematical model with such physiological parameters of tumor vasculature as total hemoglobin concentration ([HbT]), tumor blood flow (TBF), and tumor metabolic rate of oxygen (TMRO₂). The computational results show that hyperoxic gas administration to the rat tumors always gave rise to improvement of tumor Δ[HbO₂], while the same hyperoxic gas intervention could result in different responses in tumor [HbT], TBF, and TMRO₂. This preliminary study has demonstrated that NIRS, a noninvasive tool to monitor tumor oxygenation, may also be used to estimate tumor perfusion and oxygen consumption rate in response to therapeutic interventions, if a suitable mathematical model is provided.

Fluorescence optical tomography is an emerging tool for molecularly based medical imaging. In order to provide the required accuracy and resolution for imaging interior fluorescent yield and/or lifetime within the tissue, accurate experimental measurements as well as efficient and accurate numerical algorithms are needed. Herein, we present a new adaptive finite element approach to the inverse imaging problem that is able to significantly increase the resulting image resolution and accuracy, by (i) using finer meshes for the parameter estimation where the dye concentration varies significantly, (ii) using finer meshes for the fluence prediction where gradients are significant, while (iii) choosing coarse meshes in other locations. The nonlinear iterative optimization scheme is formulated in function spaces, rather than on a fixed grid. Each step is discretized separately, thus allowing for meshes that vary from one nonlinear step to the next. Furthermore, by employing adaptive schemes in the optimization, only the discretization level of the final mesh defines the achievable resolution, while the initial steps can be performed on coarse, cheap meshes. Using this technique, we can significantly reduce the total number of unknowns, which not only stabilizes the ill-posedness of the inverse problem, but also adapts the location and density of unknown parameters to achieve higher image resolution where it is needed. Specifically, we use an a posteriori error criterion to iteratively and adaptively refine meshes for both the forward and inverse problems based on derivatives of excitation and emission fluences as well as the sought parameter. We demonstrate this scheme on synthetically generated data similar to available experimental measurements.

The advent of optical molecular probes has taken optical imaging beyond approaches limited to intrinsic optical contrast mechanisms. Fluorophores are typically used as the source of contrast for optical molecular probes and the field of optical molecular imaging is concerned with measuring and quantifying their in vivo biodistribution and pharmacokinetics. Most optical molecular imaging systems are based on Continuous Wave (CW) approaches which enable rapid, full-body imaging of small animals and readily yield images of probe location, however quantification of probe concentration is challenging. Time Domain (TD) approaches, although more expensive and complicated than CW, provide more information to assist in determining the probe location and concentration. Moreover, the TD approach permits access to measuring the fluorophore lifetime which can be indicative of the probe’s environment. Existing TD approaches involve a point source and detector which are sequentially scanned over the sample and can take several minutes to acquire the data compared to the rapid imaging offered by CW. The system presented here employs a high power, near infrared, pulsed laser to provide area illumination and a temporally gated intensified CCD camera to permit area detection in order to enable rapid, full-body, TD optical molecular imaging of small animals in vivo. The system is described and preliminary in vitro and in vivo measurements are presented.

The laser “in vivo” autofluorescence diagnostics is now widely studied and applied in different areas of medicine, such as an oncology, dermatology, etc. Recently we have reported of created new professional multiwave laser diagnostic system (MLDS) for this purpose under the international scientific research and development project #1001 supported by the International Scientific and Technology Center. This presentation lights some results of application of the MLDS in a real clinical practice at Moscow Regional Research and Clinical Institute “MONIKI”, Department of Radiology. With the use of MLDS we investigated a skin and oral cavity cancer endogenous fluorescence before, during and after standard radiotherapy treatment. A statistical analysis showed that the best radiotherapy result was achieved for the patients with a small initial porfirines’ autofluorescence and a great initial flavines’ one from irradiated tumor tissues. It was shown that each radiotherapy procedure has an influence on a tissues’ autofluorescence intensity. The tendencies in porfirines’ fluorescence during a treatment course can be an additional prognostic factor for the prediction of the efficacy of a radiotherapy treatment. Moreover, it was estimated that a number of non-cancerous skin disease has a typical “cancer” initial autofluorescence, that makes it difficult to distinguish them one from another with the use of only the fluorescence diagnostics, but opens the way to investigate the non-cancerous tissues diseases with the help of tissues endogenous fluorescence phenomenon.

For efficacious transcutaneous monitoring of bone mineralization and matrix quality a spatially averaged measurement is needed, often over a large area. This precludes the use of confocal microscopy. We use picosecond pulsed laser excitation and Kerr-gated time-resolved data collection techniques to obtain marker bands of bone condition whilst rejecting interfering Raman scatter from skin, tendon and other overlying tissue. Alternatively, the methodology can be used to collect signals only from these overlying tissues. In all these experiments the 1 ps pulsed laser beam is focused to approximately 1 mm diameter. Raman light is then collected at specific times following the arrival of the pulse at time delays typically from 0 to 10 ps by opening an ultrafast optical shutter based on a Kerr cell that is driven by a second synchronized laser pulse. This permits specific probing of different layers of tissue. Individual delayed spectra are co-added and the resulting correction signal is subtracted from the ungated composite spectrum or from late-arriving time-resolved spectra. We have validated this methodology using tissue from the metacarpus and radius of several strains of laboratory mice. Overlying skin, flesh and tendon was removed from metacarpus and radius of one foreleg of a mouse and the tissue used as a control. The other foreleg served as the test specimen and was prepared by shaving the hair from the tissue, leaving the skin intact. Transcutaneous time-gated Raman spectra were measured on these specimens. With an 800 nm laser spatially resolved spectroscopy with depth penetration to greater than 1 mm was easily achieved. Normal and defective bone tissue were readily distinguished.

Fluorescence enhance optical tomography is an emerging imaging tool for investigating the molecular tissue environments in vivo. Owing to the scattering nature of near infrared radiation in tissue, iterative tomography approaches must employ the coupled diffusion equations for three-dimensional recovery of fluorescent properties from tissue boundary measurements. Unfortunately, the inverse problem suffers from computational inefficiency and ill-posed ness. Furthermore, the resolution attained in fluorescence tomography is limited by a priori fixed discretization of finite element/finite difference schemes used. These difficulties can be ameliorated by employing adaptive discretization strategies. To date, the efficacy of adaptive mesh refinement techniques has yet to be demonstrated in clinically relevant medical imaging situations. In this contribution we present a novel fluorescence tomography scheme which employs dual adaptive finite element meshes for three dimensional reconstructions of fluorescent targets beneath the simulated tissue surface. Image reconstructions for 1cm³ fluorescent target placed at the depth of 1 cm from the illumination surface are presented for target to background rations (TBRs) of 1:0 and 100:1 on the basis of dye concentration.

The ability to retrieve particle size information from back scattering reflectance with a small source-detector separation would significantly enhance the potential for development of non-invasive and minimally invasive diagnostic techniques. We present a technique for inverse determination of particle size distribution and volume fractions and validate it with polystyrene microspheres. Two of monotonic, third-degree polynomial equations were fitted from Mie theory to relate wavelength exponent 'n' and particle radii. These two equations allow us to inversely estimate the particle size from the measured 'n' value. A genetic algorithm was applied to optimize the particle size distribution and volume fraction. The experimental setup consisted of a tungsten light, CCD spectrometer with a bifurcated optical fiber for light delivery and detection. The measurement system was calibrated with a reflectance standard; different sizes and volume fractions of the suspensions were chosen for measurements. The wavelength dependence of reduced scattering coefficient was derived from the measured reflectance. Polystyrene microsphere suspensions with diameters 0.43 - 2.00 μm were characterized using the developed algorithm. The results show a good agreement between the particle size retrieved by our algorithm and manufacturer’s data, demonstrating a robust method for particle size determination using near infrared reflectance and small source-detector separation.

Determination of blood flow changes will be helpful for evaluation of tumor prognosis and therapy. Our study is to develop an in vitro hemodynamic phantom model, which allows us to show the feasibility of using near infrared spectroscopy (NIRS) to determine flow changes as a dynamic imaging modality to monitor tumor responses to therapy. In the hemodynamic phantom model, both single and multiple, transparent, plastic tubes were used to pass through a cylindral glass chamber. The chamber was filled with either an Intralipid solution or a soft gelatin phantom, while the tube or tubes were pumped with either an Intralipid-ink mixture or animal whole blood to simulate the tumor vasculature. The Intralipid solutions that were filled in the chamber and tubes had optical scattering and absorption properties similar to those of tumor tissues and tumor vasculature. A single-channel, broadband, NIRS system with a tungsten light source and a CCD-array spectrometer was used to quantify the changes in optical density (OD) of the intralipid-ink mixture with variations in flow rate and concentration. A single-exponential curve fit has been used to determine the time constant (τ) from the change in OD to estimate the flow rate. The obtained preliminary results show a strong correlation between changing rates of concentration and flow; a multivariable dynamic mathematical model may be also established to relate changes of Hb, HbO and blood volume with blood flow.

The Mueller matrix describes all the polarizing properties of a sample, and therefore the optical differences between non-cancerous and pre-cancerous tissue should be present within the matrix elements. We present in this paper that a high speedpolarimetry system generates 16 full Mueller matrices to characterize tissues. Feature extraction is done on the Mueller matrix elements resulting in the depolarizance and retardance images by polar decomposition to detect and classify of early oral cancers and pre-cancerous changes in epithelium, such as dysplasia. These images are compared with orthogonal polarization image and analyzed in an attempt to determine the important factors for the identification of cancerous lesions from their benign counterparts. Our results indicate that polarimetry has potential as a method for the in vivo early detection and diagnosis of oral premalignancy and malignancy.

We used a scanning laser pulse mammograph to record craniocaudal and mediolateral projection optical mammograms of 154 patients suspected to have breast cancer. Optical mammograms were analyzed by comparing them with x-ray and MR mammograms, including results of histopathology. Out of 102 histologically confirmed carcinomas, 92 carcinomas were visible in at least one of the two projection mammograms. On average optical mammograms based on photon counts in a late time window exhibited the carcinomas with highest contrast compared to mammograms displaying absorption coefficients or hemoglobin concentration. Optical properties of carcinomas visible in optical mammograms were determined employing the model of diffraction of photon density waves by a spherical inhomogeneity, located in an otherwise homogeneous tissue slab. On average, tumor absorption coefficients exceeded those of surrounding healthy breast tissue by a factor of about 2.5 at the shortest wavelength used (670 nm), whereas tumor reduced scattering coefficients were larger by about 20% at this wavelength. Total hemoglobin concentration was observed to be systematically larger in tumors compared to healthy breast tissue. In contrast, blood oxygen saturation was found to be a poor discriminator for tumors and healthy breast tissue.

We have used near-infrared spectroscopy (NIRS) to study hemodynamic auditory evoked responses on 7 full-term neonates. Measurements were done simultaneously above both auditory cortices to study the distribution of speech and music processing between hemispheres using a 16-channel frequency-domain instrument. The stimulation consisted of 5-second samples of music and speech with a 25-second silent interval. In response to stimulation, a significant increase in the concentration of oxygenated hemoglobin ([HbO₂]) was detected in 6 out of 7 subjects. The strongest responses in [HbO₂] were seen near the measurement location above the ear on both hemispheres. The mean latency of the maximum responses was 9.42±1.51 s. On the left hemisphere (LH), the maximum amplitude of the average [HbO₂] response to the music stimuli was 0.76± 0.38 μ M (mean±std.) and to the speech stimuli 1.00± 0.45 μ± μM. On the right hemisphere (RH), the maximum amplitude of the average [HbO₂] response was 1.29± 0.85 μM to the music stimuli and 1.23± 0.93 μM to the speech stimuli. The results indicate that auditory information is processed on both auditory cortices, but LH is more concentrated to process speech than music information. No significant differences in the locations and the latencies of the maximum responses relative to the stimulus type were found.

Motivation: Early diagnosis of Alzheimer's disease (AD) is crucial because symptoms respond best to available treatments in the initial stages of the disease. Recent studies have shown that marked changes in brain oxygenation during mental and physical tasks can be used for noninvasive functional brain imaging to detect Alzheimer’s disease. The goal of our study is to explore the possibility of using near infrared spectroscopy (NIRS) and mapping (NIRM) as a diagnostic tool for AD before the onset of significant morphological changes in the brain. Methods: A 16-channel NIRS brain imager was used to noninvasively measure spatial and temporal changes in cerebral hemodynamics induced during verbal fluency task and physical activity. The experiments involved healthy subjects (n = 10) in the age range of 25±5 years. The NIRS signals were taken from the subjects' prefrontal cortex during the activities. Results and Conclusion: Trends of oxygenated and deoxygenated hemoglobin in the prefrontal cortex of the brain were observed. During the mental stimulation, the subjects showed significant increase in oxygenated hemoglobin [HbO₂] with a simultaneous decrease in deoxygenated hemoglobin [Hb]. However, physical exercise caused a rise in levels of HbO₂ with small variations in Hb. This study basically demonstrates that NIRM taken from the prefrontal cortex of the human brain is sensitive to both mental and physical tasks and holds potential to serve as a diagnostic means for early detection of Alzheimer's disease.

Our group has recently established that joints affected by Rheumatoid Arthritis (RA) can be distinguished from healthy joints through measurements of the scattering coefficient. We showed that a high scattering coefficient in the center of the joint is indicative of a joint with RA. While these results were encouraging, data to date still suffers from low sensitivity and specificity. Possibly higher specificities and sensitivities can be achieved if dynamic measurements of hemodynamic and metabolic processes in the synovium are considered. Using our dual-wavelength imaging system together with previously implemented model-based iterative image reconstruction schemes, we have performed initial dynamic imaging studies involving healthy human volunteers and patients affected by RA. These case studies seem to confirm our hypothesis that differences in the vascular reactivity exist between affected and unaffected joints.

Migraine is a complex chronic neurovascular disorder in which the interictal changes in neuronal excitability and vascular reactivity in the cerebral cortex were detected. The extent and direction of the changes in cerebral blood flow that affect cerebral hemodynamics during attacks, however, are still a matter of debate. This may have been due to the logistic and technical problems posed by the different techniques to determine cerebral blood flow during migraine attacks and the different definitions of patient populations. In this study, we have investigated hypercapnia challenges by breath holding task on subjects with and without migraine by using functional near infrared spectroscopy (fNIRS). Measurements of the relative changes in concentration of deoxy-hemoglobin [Hb] and oxy-hemoglobin [HbO₂] are performed on four healthy subjects during three breath holdings of 30 seconds (s.) interleaved with 90 s. of normal breathing. We have observed [Hb]increase during breath holding interval in subject without migraine whereas in subject with migraine [Hb] decreases during breath holding interval. The result of our study suggest that hypercapnia effect on cerebral hemodynamic of subject with migraine and without migraine could be due to different vascular reactivity to P_CO2 (carbon dioxide partial pressure) in arteries.

Recently, conventional near infrared spectroscopy (NIRS) for oxymetry has been extended with an indocyanine green (ICG) dye dilution method which allows the estimation of cerebral blood flow (CBF) and cerebral blood volume (CBV). The signal obtained through the skull is substantially influenced by extracerebral tissue. In order to quantify and eliminate extracerebral contamination of the optical density signal we have applied two approaches. Firstly, we used spatially resolved spectroscopy (SRS) with a two receiver arrangement, with separations between emitter and two receivers in distances of d1=4.0cm and d2=6.5cm. The magnitude of the determined extracerebral contamination was verified with NIRS measurements in patients after brain herniation. Intracerebral circulatory arrest was confirmed by transcerebral Doppler examination. Secondly, Monte Carlo simulation was used to simulate the light propagation through the head to quantify the extracerebral contamination of the optical density signal of NIRS. The anatomical structure is determined from 3D-magnetic resonance imaging (MRI) using a voxel resolution of 0.8 x 0.8 x 0 .8 mm³ for three different pairs of T1/T2 values. We segment the MRI data to obtain a material matrix describing the composition of skin, skull, cerebral spinal fluid (CSF), grey and white matter. Each voxel in this material matrix characterizes the light absorption and dispersion coefficient of the identified material. This material matrix is applied in the Monte Carlo simulation. With SRS an extracerebral contamination of 65% of the optical density signal is extracted, while the Monte Carlo simulation results show that the extracerebral contamination decreases from 70% to 30% with increasing emitter-receiver distance. Differences between the NIRS ICG dye dilution technique and conventional NIRS oxymetry concerning the extracerebral contamination are discussed.

The absolute diffuse optical tomography (DOT) has been rather difficult to achieve due to the problems arising upon the robustness of the algorithm to uncertainties in measuring conditions. Alternatively, the differential imaging scheme was applied to reconstruct a difference image between a target and a baseline reference from the difference data. Nevertheless, the absolute imaging scheme is desirable for unavailability of the reference in many situations. The absolute imaging usually uses intensity-independent data-type, which has been popularly the mean time of flight (TOF) in time-resolved (TR) detection, to avoid absolute instrument scaling. A problem with the mean TOF is its is insufficient sensitivity to deep absorption change to cope with the measuring noises, such as uncertainty of the optode positions. Therefore seeking for more robust data-type has been a key task in the community. We have previously developed an image reconstruction algorithm for TR-DOT, based on the modified generalized pulse spectrum technique (GPST), where the ratio between the Laplace-transformed TR re-emissions at two real-domain frequencies is used as the data-type. It is computationally the same efficient as the mean TOF but offers a potentiality to enhance noise-robustness by optimizing the working frequencies. We demonstrate here that the robustness of this data-type to optode position uncertainty can be substantially increased by enlarge the difference between the two working frequencies. We optimize the working frequencies within the range of physical sense and numerically validate the method for brain-simulating two-layer geometry using the TR reflected light.

Diffuse optical tomography (DOT) in the near infrared involves reconstruction of spatially varying optical properties of turbid medium from boundary measurements based on a forward model of photon propagation. Due to highly non-linear nature of the DOT, high quality image reconstruction is a computationally demanding problem that requires repeated solutions of both the forward and the inverse problems. Therefore, it is highly desirable to develop methods and algorithms that are computationally efficient. In this paper, we propose a domain decomposition approach to address the computational complexity of the DOT problem. We propose a two-level multiplicative overlapping domain decomposition method for the forward problem and a two-level space decomposition method for the inverse problem. We showed the convergence of the inverse solver and derived the computational complexity of each method. We demonstrate the performance of the proposed approach in numerical simulations.

We have performed a detailed analysis for the problem of photon migration through a scattering slab containing a single absorptive inclusion whose absorption coefficient is characterized by a spatially varying Gaussian distribution law. The analysis has been performed within the framework of the first-order perturbation approach to the diffusion theory for a slab geometry and a coaxial measurements arrangement. An analytical expression has been derived that account for the change in the time-resolved transmittance in presence of the absorptive Gaussian inclusion. We present experimental results of time-resolved measurements that have been performed on absorptive phantoms with the aim to investigate and to validate the ability of this model to predict the optical properties of absorptive inhomogeneities hidden inside a scattering medium.

A wireless noninvasive optical oxygen monitor based on near infrared spectroscopy is presented. Multi-objects (up to 20) could be monitored wirelessly by the center control system. The monitor is made of two parts. Part one is for the measurement of relative changes of oxyhemoglobin, deoxyhemoglobin and blood volume. The data from probe is transferred by GPRS Data Module. The system includes hardware for data receiving, data pre-processing, A/D converting, communication with a portable computer, and software for real time data processing. Experimental results are demonstrated.

Near infrared (NIR) diffused optical tomography (DOT) is emerging as a potential tool of non-invasively diagnosing woman breast cancers, neonatal brain hypoxia, and other human organ diseases. The intensive and worldwide investigations in theory and experiment have revealed the possibility of NIR DOT in providing both anatomical and functional information of biological tissue simultaneously, which is important for distinguishing between healthy and diseased tissues, such as benign and malignant tumors. In this paper, our recent DOT experiments on human lower legs and forearms are presented using our time-resolved measuring system and image reconstruction algorithm based on the modified generalized pulse spectrum technique. It was found that the image quality in the experiments, including both the spatial resolution and the quantitativeness of the targets, was rather poor, and the interior blood vessels undisclosed in the absorption images. To clarify this issue, the influences of target contrast and size on the image reconstruction were investigated with simulated data. We have accordingly obtained the following observations: the quantitativeness of the reconstructed optical properties was prone to be spoiled by the small size ratio and high contrast of the interior targets (such as blood vessels) to the background, and the incompleteness of information embedded in the featured data-types, in addition to the experimental noise, evidently answers for the degradation of the spatial resolution and quantitativeness. It was shown in a further simulative investigation that the image quality could be substantially improved by making full use of the time-resolved data.

We present data collected with near-infrared spectroscopy on the human forearm (brachioradialis muscle) to characterize the hemodynamic response to venous occlusion in muscle. Venous occlusion was achieved in the upper arm by inflating a pneumatic cuff to a pressure of 60 mmHg. We performed absolute measurements of concentration and oxygen saturation of hemoglobin on six healthy adult human subjects. On all six subjects, we consistently found that during a 40-s venous occlusion the hemoglobin concentration increases (by 5.6±2.3 μM), while the oxygen saturation of hemoglobin decreases (by 2.1±0.7%). This accumulation and desaturation of blood in the forearm in response to the upper arm occlusion can be described with an electrical model in which the charge stored by a capacitor represents the local blood volume, and the electrical current represents blood flow.

Several phantom and in vivo small animal imaging studies have been performed to detect the re-emitted fluorescence signal arising from micro to pico molar concentrations of fluorophore by employing band-pass and band-rejection filters. However, elimination of the back-reflected excitation light still remains a major challenge for further reducing the noise floor in fluorescence imaging. Furthermore, despite the well-known deterioration of interference filter performance as the angle of incidence deviates from zero degrees, most studies do not employ collimated light optical design required for efficient excitation light rejection using interference filters. In this study, we measured quantities in frequency domain data for the combination of three-cavity interference and holographic super notch filters. To assess excitation leakage, the “out-of-band (S (λ_x ) )” to “in-band (S (λ_m ) - S (λ_x ) )” signal ratio, AC amplitude (I_AC ), and phase delay (δ-δ*) measured from a gain modulated, intensified CCD imaging system with and without collimating optics was evaluated. The addition of collimating optics resulted in a reduction of 82% to 91% of the out-of-band to in-band ratio for the phantom studies and an increase of 1.4 to 3.7 times of the target-to-background ratio (T:B) for small animal studies.

Fluorescence enhanced optical tomography uses modulated NIR light and an exogenous fluorescent contrast agent in order to assess its spatial distribution deep within tissues. However, a spatial distribution of endogenous optical properties of absorption and scattering coefficients arises due to normal, structured anatomical background which can be expected to vary from patient to patient. This structured background can be a source of randomness (or “noise”) in the task of detecting a fluorescent target. In addition, there may be non-uniform distribution of exogenous fluorescent agent. Our objective is to develop the tools for OAIQ in order to assess the performance of image reconstruction algorithms in the presence of anatomical backgrounds owing to both endogenous and exogenous optical properties. We consider the lumpy-object model developed by Rolland and Barrett to simulate the normal background anatomy as a representation of the non-specific distribution of the fluorescent agent as well as the natural heterogeneity of the endogenous tissue optical properties. Reconstructed images show the insensitivity to endogenous optical property lumps in fluorescence enhanced optical imaging. The reconstructed images are more sensitive to uneven distribution of exogenous fluorophore in normal tissues, but can nonetheless be achieved.

The use of near-infrared (NIR) spectroscopy to interrogate deeper tissue volume has shown enormous potential for molecular-based non-invasive imaging when coupled with appropriate excitable dyes. As most of the breast cancers are hormone dependent hence determination of the hormonal receptor status gains paramount importance when deciding the treatment regime for the patient. Since proliferations of the breast cancer cells are often driven by estrogen, we focus on to developing a technique to detect estrogen receptor status. As a first step, the objective of this work was to synthesize and characterize one such novel NIR fluorescent (NIRF) conjugate, which could potentially be used to detect estrogen receptors. The conjugate was synthesized by ester formation between 17-b estradiol and a cyanine dye namely: bis-1, 1-(4-sulfobutyl) indotricarbocyanine-5-carboxylic acid, sodium salt. The cyanine dye is a hydrophilic derivative of indocyanine green (ICG). The ester formed was found to have an extra binding ability with the receptor cites as compared to ICG, which was established by the partition coefficient studies. This cyanine dye has a partition coefficient less than 0.005 as compared to that of ICG (>200)[1]. In addition the ester showed enhanced fluorescent quantum yield than ICG. The replacement of the sodium ion in the ester by a larger glucosammonium ion was found to enhance the hydrophilicity and reduce the toxic effect on the cell lines. The excitation and emission peaks for the conjugate were recorded in the NIR region as 750nm and 788nm respectively. The ester developed was tested on the breast cancer cell lines MCF-7 and found non-toxic. The tagging characteristics were pivotal determinants underlying the ability of the fluorescent conjugate in binding the estrogen receptor of the breast cancer cells. This technique offers the potential of non-invasive detection of hormone receptor status in vivo and may help in decreasing the load of unnecessary biopsies. Here, we have reported the progress made in the development of a novel NIR external contrast agent and the work is in progress to use this conjugate for the molecular-based, diagnostic imaging of breast cancer.

A new imaging approach for three-dimensional localization and characterization of absorptive, scattering or fluorescent objects in a turbid medium is presented and demonstrated using simulated and experimental data. This approach uses a multi-source and multi-detector signal acquisition scheme and independent component analysis (ICA) from information theory for target localization and characterization. Independent component analysis of the perturbation in the spatial intensity distribution or the fluorescent signal measured on the medium boundary sorts out the embedded objects. The location and optical characteristics (size, shape and optical property) of the embedded objects are obtained from a Green's function analysis based on an appropriate model for light propagation in the background medium and back-projections of the retrieved independent components.

No abstract provided

This paper points to the optimization of recording of signals from the prefrontal cortex (PFC) of an age-matched group of students studied annually over six years and challenged by various interrogations with resulting localized PFC signals. Here we discuss the evolution of a successful technology and cite exemplary results.

Near infrared spectroscopy (NIRS) has the ability to record, at high temporal resolution, hemodynamic changes within the brain during functional activity. Although alone, NIRS has a poorer spatial resolution compared to other imaging methods such as functional MRI (fMRI), multi-modality approaches, which attempt to fuse the spatial resolution of MRI with the hemoglobin oxygenation information and temporal resolution of NIRS, show promise to yielding better insight into the hemodynamic and metabolic response of the functional brain in future research. However, paramount to the development of these multi-modality approaches, proper control experiments to validate the correlation between NIRS and fMRI methods must be preformed. In this experiment, we have examined the spatial and temporal relationship between the NIRS measure of deoxy-hemoglobin and the fMRI blood oxygen level dependent (BOLD) signal. Here, we have modeled the propagation of light through realistic, tissue segmented, head models for each of five subjects. Using these sensitivity profiles, we predicted the measurement of deoxy-hemoglobin for each individual NIRS source-detector pair from the projection of the volume-wise fMRI BOLD changes, thus allowing a quantitative spatial and temporal comparison between NIRS and fMRI. We report a linear correlation of R = 0.73 (p < 2x10 ^-8) between the spatial profiles between the NIRS measure of deoxy-hemoglobin and BOLD signal. We also report a temporal correlation of R=0.88 (p<9x10 ^-18) between the region-of-interest averaged responses using the projected BOLD response.