We report non-invasive, bilateral measurements of cerebral oxygenation performed with near-infrared spectroscopy (NIRS) on ten patients undergoing right unilateral electro-convulsive therapy (ECT). Right unilateral ECT consists of delivering an electrical current through the right brain hemisphere to induce a seizure, which is associated with significant changes in systemic and regional physiological parameters. In this work, we have examined the regional cerebral oxygenation (StO₂) measured with NIRS on the right and left sides of the frontal brain region, and the systemic arterial oxygenation (SaO₂) measured with pulse oximetry. On the ten patients examined, we have found that the decrease in the cerebral oxygenation on the side ipsilateral to the ECT electrical discharge (ΔStO₂^(ipsi)) is consistently stronger than the decrease on the contralateral side (ΔStO₂^(contra)). The average and standard deviation for the ipsilateral and contralateral oxygenation changes across the ten patients are ΔStO₂^(ipsi) = -22 ± 10% and ΔStO₂^(contra) = -6 ± 10%, respectively. By contrast, we observed two distinct behaviors in the arterial saturation; in five patients, SaO₂ did not significantly change during the ECT procedure, and in three patients, SaO₂ decreased by -16± 6%, an intermediate value with respect to the changes observed in StO₂^(ipsi) and StO₂^(contra) (we do not have the SaO₂ recording in the remaining two patients for technical reasons). These results indicate that NIRS monitoring of the cerebral oxygenation during ECT has the potential of being a valuable addition to the standard monitoring of arterial saturation with pulse oximetry.

Optical tomography is being developed at UCL as a tool for understanding the mechanisms of haemorrhagic and hypoxic-ischaemic brain injury and assessing the effectiveness of novel neural rescue therapies in the newborn infant. Our 32-channel time-resolved optical imaging system measures photon flight times between multiple pairs of points on the surface of the head, and images sensitive to local variation in tissue absorption and scattering properties are reconstructed using non-linear algorithms. Several studies have been performed on premature infants using custom-built helmets, which hold up to 32 sources and detectors in contact with the head. Simulations have revealed that combining data with reference measurements acquired on a homogeneous object using the same fibre locations can significantly reduce errors in reconstructions due to uncertainty in the location of the sources and detectors. To provide the reference data, a homogeneous phantom based on a balloon filled with a scattering fluid of precisely known optical properties was made and inserted into the helmet immediately following each infant scan. In this paper, we evaluate the effectiveness of this approach by acquiring data on a realistically head-shaped phantom containing a small perturbation, and reconstructing it using the homogeneous head-shaped phantom and the fluid-filled balloon.

There has been considerable discussion concerning the effects of the cerebrospinal fluid on measurements of blood-related parameters in the human brain, and if diffusion-theory-based image reconstruction algorithms can accurately account for the light propagation in the head. All of these studies have been performed either with synthetic data generate from numerical models or from phantom studies. We present here the first comparative study that involves clinical data from optical tomographic measurements. Data obtained from the human forehead during a Valsalva maneuver were input to two different model-based iterative image reconstruction algorithms recently developed in our laboratories. One code is based on the equation of radiative transfer, while the other algorithm uses a diffusion model to describe the light propagation in the head. Both codes use finite-element formulations of the respective theories and were used to obtain three-dimensional volumetric images of oxy, dexoy and total hemoglobin. The reconstructed overall spatial heterogeneity in changes of these parameters is similar using both algorithms. The two codes differ mostly in the amplitude of the observed changes. In general the transport based codes reconstructs changes 10-40% stronger than the diffusion code.

Monitoring of intra-mitochondrial NADH redox state is a common in-vivo technique in experimental animals and is rare in clinical studies. The combination of NADH monitoring with the Laser Doppler flowmetry for cerebral blood flow monitoring was described in various publications. Until now, very small effort was made to monitor NADH and CBF inside the cortex of experimental animals. The significance of this monitoring is in its application to experimental models of Parkinson’s disease or to clinical monitoring situations in the intensive care unit, when ICP is monitored. Here we compared the responses of the gerbil or rat brain to oxygen deficiency, monitored on the brain surface and in different depths. After the animals were anesthetized, the two common carotid arteries (gerbil) were isolated and prepared for following occlusion. The brain was exposed and two optical probes were located on its surface. Ischemia was induced by occluding the two carotid arteries, and anoxia was preformed by inhalation of pure N₂. After recovery, one of the probes was inserted into the cortex (0.5-3mm) and a second ischemia or anoxia was preformed. The results showed that: 1. It is possible to monitor both CBF and NADH on the brain surface simultaneously with intracortical location. 2. The responses of the brain to ischemia or anoxia was smaller inside the cortex comparing to brain surface. 3. Negative correlation was found between CBF and NADH in both locations and models. In conclusion, this new model of simultaneously monitoring of CBF and NADH in different cerebral locations can shed light on various pathophysiological situations.

The emerging sub-field of dynamic medical optical tomography shows great potential for conferring significantly enhanced early diagnosis and treatment monitoring capabilities upon researchers and clinicians. In previous reports we have showed that adoption of elementary time-series analysis techniques can bring about large large improvements in localization and contrast in optical tomographic images. Here we build upon the earlier work, and show that well-known techniques for extraction and localization of signals embedded in a noisy background, and for deconvolution of signal mixtures, also can be successfully applied to the problem of interpreting dynamic optical tomography data sets. A general linear model computation is used for the signal extraction/localization problem, while the deconvolution problem is addressed by means of a blind source separation technique extensively reported.

In this paper, a 3D dual-mesh imaging reconstruction method is demonstrated, which can reconstruct absorption and scattering coefficients simultaneously. In the dual-mesh scheme, the total number of voxels with unknown absorption and scattering perturbations are maintained on the same order of total measurements by using a fine grid for target region and a coarse grid for background region. Certain row/column normalization has been applied to alleviate the crosstalk between the absorption coefficient and scattering coefficient, and to minimize the depth dependent problem. Experimental results of targets with different absorption and scattering contrasts have shown that accurate reconstruction of both absorption and scattering coefficients can be achieved.

In large 3-D finite element optical tomography problems, computation times for forward and adjoint solutions and for calculation of sensitivities can become prohibitive. Parallelization of computer codes can be used to obtain speedups approaching the number of processors employed, but parallel codes and computer systems can be difficult and expensive to develop and maintain. We show that by employing highly vectorized code that takes advantage of pipelining capabilities in the microprocessor we achieve dramatic speedups for forward and adjoint sensitivity calculations on a single processor microcomputer, and that these speedups actually increase as the problem size increases. Our vectorized implementations involve replication of large amounts of data and are thus memory intensive, however we effectively remove memory constraints by using domain decomposition to control the use of virtual memory. We show that global matrix assembly for a large (98,304 element) mesh is speeded up by a factor of 6.5 and adjoint sensitivity calculations of emission fluence with respect to fluorescence absorption are speeded up by a factor of 502 on a single-processor 2.2 GHz Pentium IV.

Noninvasive diagnosis of solid benign and malignant tumors presents a unique challenge to all imaging modalidies. Ultrasound is routinely used in conjunction with mammography to deffierenciate simple cysts from solid lesions. However, the overlapping appearances of benign and malignant lesions make ultrasound less useful in differenciating solid lesions, which results in a large number of normal biopsies. Optical tomography using near infrared diffused light has a great potentional for imaging functional parameters of tumor total hemoglogin concentration, oxygeon saturation, and metabolism, etc, and these parameters can differenciate benign from malignant lesions. However, optical tomography alone suffers from low spatial resolution and target localization uncertainty due to the intensive light scattering inside the tissue. Our aims were to combine diffused light imaging with conventional ultrasound for detection and diagnosis of solid lesions. Initial findings of palpable and non-palpable solid breast lesions have shown that early stage invasive cancers have much higher total hemoglobin concentration due to angiogenesis than benign lesions. Furthermore, the spatial distribution of the total hemoglobin concentration is well localized in small invasive cancer cases, and is quite diffused in benign lesions.

We have introduced working technology that provides for time-series imaging of the hemoglobin signal in large tissue structures. In this study we have explored our ability to detect aberrant time-frequency responses of breast vasculature for subjects with Stage II breast cancer at rest and in response to simple provocations. The hypothesis being explored is that time-series imaging will be sensitive to the known structural and functional malformations of the tumor vasculature. Mammographic studies were conducted using an adjustable hemisheric measuring head containing 21 source and 21 detector locations (441 source-detector pairs). Simultaneous dual-wavelength studies were performed at 760 and 830 nm at a framing rate of ~2.7 Hz. Optical measures were performed on women lying prone with the breast hanging in a pendant position. Two class of measures were performed: (1) 20- minute baseline measure wherein the subject was at rest; (2) provocation studies wherein the subject was asked to perform some simple breathing maneuvers. Collected data were analyzed to identify the time-frequency structure and central tendencies of the detector responses and those of the image time series. Imaging data were generated using the Normalized Difference Method (Pei et al., Appl. Opt. 40, 5755-5769, 2001). Results obtained clearly document three classes of anomalies when compared to the normal contralateral breast. 1) Breast tumors exhibit altered oxygen supply/demand imbalance in response to an oxidative challenge (breath hold). 2) The vasomotor response of the tumor vasculature is mainly depressed and exhibits an altered modulation. 3) The affected area of the breast wherein the altered vasomotor signature is seen extends well beyond the limits of the tumor itself.

Near Infrared Spectroscopy (NIRS) is a powerful method for non-invasive mapping of cerebral functional activation. We have developed an NIRS instrument that is portable, inexpensive and lightweight consisting of an array of light emitting diodes (LEDs) and photodiodes (PDs) mounted on a flexible printed circuit board (PCB). The flexibility and portability of the instrument makes it easy to apply to subjects ranging from premature babies in intensive care to adults. The flexible PCB array consists of 48 LEDs operating at two different wavelengths (780nm & 880nm) together with 14 photodiodes. Transimpedence amplifiers for each of the PDs are located on the flexible pad to minimise noise pick up. The LEDs are pulsed (10ms) at a peak optical power of 20mW, while a sample and hold circuit monitors the voltages at all of the PDs. The array is also encapsulated in black silicon rubber, except for the regions directly above the LED's and PD's, which have a clear silicone rubber cover. The sensor array is attached to the electronics by two one metre long flexible ribbon cables. The monitoring circuit provides medical grade electrical isolation between the patient and computer. Studies have been conducted on phantoms to test the penetration depth of the sensor array for two different separations of LEDs and PDs (11mm and 27mm). The maximum depths that can be probed are 5mm and 11mm respectively. This makes it suitable for studying cortical activation in babies.

Diffuse near-infrared tomography of tissue can provide intrinsically useful information about total hemoglobin, oxygen saturation, water and cytochromes within tissue, yet to extract this information spectral data is required at many wavelengths. In this study, we examine a new approach to using multispectral continuous wave measurements through tissue along with second-derivative data analysis methods to estimate the pathlength in tissues over multiple tomographic paths. The goal of this work has been to demonstrate that the optical differential pathlength that is estimated by spectroscopy methods is directly related to the optical pathlength as measured by frequency-domain signals. This direct relation then allows the use of tomographic algorithms which have been developed for frequency-domain optical tomography to be applied to multispectral continuous wave data. The theoretical development is presented here, along with numerical validation in homogeneous and heterogeneous tissue-simulating regions. These results indicate that the approach outlined is valid and provides the theoretical basis for developing multispectral near-infrared tomography of tissues.

Noninvasive examination of the hemodynamics of brain tissue is of general interest in many areas of medicine and physiology. To date, optical brain studies generate topographic maps, where signals obtained form within the brain are projected onto two-dimensional surface maps. Recently, our group has presented the first three-dimensional, volumetric reconstruction of hemodynamic changes during a Valsalva maneuver in the human forehead. To further validate our three-dimensional diffusion optical tomographic reconstruction algorithm we have turned to experimental studies involving small animals. Here we report on hypercapnia studies performed with 3-month old Sprague-Dawley rats. After anesthetizing the animal a tracheotomy was performed and the rat was artificially respirated. The head shaven and secured in a stereotaxic frame an optical probe was positioned between the bregma and lambda skull landmarks. A baseline measurement was recorded and then the inspired gas content was altered. The experimental studies verified the ability of our code to three-dimensionally visualize a global hemodynamic phenomenon in the rat head in response to perturbations in the inspired CO₂ concentrations. Specifically, we incrementally increased the concentration of inspiratory CO₂ (hypercapnia) and visualized the resulting hemodynamic change. We observed a global increase in blood volume and oxygenation, which was consistent with the known physiologic response to hypercapnia. A second set of experiments were designed to determine the sensitivity of the reconstruction to inaccurate probe positioning versus assumed model optrode-position mismatch. We determined that shifts on the order of 1/10 the maximum optrode separation significantly influence the reconstruction and may falsely produce lateralized effects.

The possibility of obtaining images of human organs by optical techniques for diagnostic purposes is still under intense investigation. Useful information for devising the most efficient data acquisition procedures and reconstruction algorithms can be obtained by imaging well-controlled small inhomogeneities using different reconstruction procedures. For this purpose we performed time-resolved imaging experiments by embedding in a turbid slab small inclusions with different absorption coefficients. The ability of the imaging system to detect these hidden objects strongly depends on its optical properties and on the method that is adopted to reconstruct images. To this regard, in order to construct images different procedures can be adopted. The efficiency of these different procedures has been valuated by investigating the contrast, the noise and the effective signal-to-noise ratio of constructed images. Furthermore, this data analysis has been used to study the role of the inclusion’s object contrast on the efficiency of the adopted reconstruction algorithm.

We used a four-channel intensity-modulated near-infrared spectroscopy device to study the hemodynamic responses due to brain activation in adults and neonates. The stimuli included finger tapping, tickling of the heel, and auditory stimuli. The subjects included two adults and ten neonates of age between 0.5 and 4 days. A block paradigm was used in the studies, and responses were successfully obtained from both subject groups.

The spatio-temporal characteristics of changes in cerebral blood volume associated with neuronal activity were investigated in the hindlimb somatosensory cortex of α-chloralose/urethan anesthetized rats (n=10) with optical imaging at 570nm through a thinned skull. Activation of cortex was carried out by electrical stimulation of the contralateral sciatic nerve with 5Hz, 0.3V pulses (0.5ms) for duration of 2s. The stimulation evoked a monophasic optical reflectance decrease at cortical parenchyma and arteries sites rapidly after the onset of stimulation, whereas no similar response was observed at vein compartments. The optical signal changes reached 10% of the peak response 0.70±0.32s after stimulation onset and no significant time lag in this 10% start latency time was observed between the response at cortical parenchyma and arteries compartments. The evoked optical reflectance decrease reached the peak (0.25%±0.047%) 2.66±0.61s after the stimulus onset at parenchyma site, 0.40±0.20s earlier (P<0.05) than that at arteries site (0.50±0.068% 3.06±0.70s). Variable location within the cortical parenchyma and arteries compartment themselves didn’t affect the temporal characteristics of the evoked signal significantly. These results suggest that the sciatic nerve stimulation evokes a local blood volume increase at both capillaries (cortical parenchyma) and arterioles rapidly after the stimulus onset but the evoked blood volume increase in capillaries could not be entirely accounted for by the dilation of arterioles.

We are using rodent animal models to study and compare contrast mechanisms for detection of breast cancer. These measurements are performed with the animals immersed in a matching scattering medium. The matching scattering medium or liquid tissue phantom comprises a mixture of Ropaque (hollow acrylic/styrene microspheres) and ink. We have previously applied matched imaging to imaging in humans. Surrounding the imaged region with a matched tissue phantom compensates for variations in tissue thickness and geometry, provides more uniform illumination, and allows better use of the dynamic range of the imaging system. If the match is good, the boundaries of the imaged region should almost vanish, enhancing the contrast from internal structure as compared to contrast from the boundaries and surface topography. For our measurements in animals, the immersion plays two additional roles. First, we can readily study tumors through tissue thickness similar to that of a human breast. Although the heterogeneity of the breast is lost, this is a practical method to study the detection of small tumors and monitor changes as they grow. Second, the immersion enhances our ability to quantify the contrast mechanisms for peripheral tumors on the animal because the boundary effects on photon migration are eliminated. We are currently developing two systems for these measurements. One is a continuous-wave (CW) system based on near-infrared LED illumination and a CCD (charge-coupled device) camera. The second system, a frequency domain system, can help quantify the changes observed with the CW system.

Two experimental setups for time-resolved diffuse optical tomography (DOT) are described. The first involves a titanium-sapphire laser and a streak camera in a synchroscan mode. A multiple-arm light guide allows to measure simultaneously the temporal profiles of photons re-emitted at different boundary sites of the objects studied. The second one uses picosecond laser diodes and a multiple-anode micro-channel plate photomultiplier tube followed by parallel time-correlated single photon counting channels. The two instruments are described and their main specifications compared: instrument response and time resolution, sweep time, temporal stability, sensitivity and its inter-channel variation, cross-talk between adjacent channels. We will also discuss detector-specific data analysis prior to image reconstruction. Both instruments have been tested with phantoms simulating tissues and the absorption and reduced scattering images obtained have been compared.

Measurement of concentration, oxygenation, and flow characteristics of blood cells can reveal information about tissue metabolism and functional heterogeneity. An improved multifunctional hybrid system has been built on the basis of our previous hybrid instrument that combines two near-infrared diffuse optical techniques to simultaneously monitor the changes of blood flow, total hemoglobin concentration (THC) and blood oxygen saturation (StO2). Diffuse correlation spectroscopy (DCS) monitors blood flow (BF) by measuring the optical phase shifts caused by moving blood cells, while diffuse photon density wave spectroscopy (DPDW) measures tissue absorption and scattering. Higher spatial resolution, higher data acquisition rate and higher dynamic range of the improved system allow us to monitor rapid hemodynamic changes in rat brain and human muscles. We have designed two probes with different source-detector pairs and different separations for the two types of experiments. A unique non-contact probe mounted on the back of a camera, which allows continuous measurements without altering the blood flow, was employed to in vivo monitor the metabolic responses in rat brain during KCl induced cortical spreading depression (CSD). A contact probe was used to measure changes of blood flow and oxygenation in human muscle during and after cuff occlusion or exercise, where the non-contact probe is not appropriate for monitoring the moving target. The experimental results indicate that our multifunctional hybrid system is capable of in vivo and non-invasive monitoring of the hemodynamic changes in different tissues (smaller tissues in rat brain, larger tissues in human muscle) under different conditions (static versus moving). The time series images of flow during CSD obtained by our technique revealed spatial and temporal hemodynamic changes in rat brain. Two to three fold longer recovery times of flow and oxygenation after cuff occlusion or exercise from calf flexors in a patient with peripheral vascular disease (PVD) were found.

We have developed a near infrared optical tomography system features fast optical switching, three-wavelength excitations, and avalanche photodiode (APD) detectors with a high dynamic range. Pigtailed laser diodes at 660, 780, and 830 nm are used as light sources and their outputs are distributed sequentially to one of nine source fibers. The crosstalk between source channels is around 65 dB, equivalent to 130 dB in opto-electrical signals. 10 Silicon APD’s detect diffusive photon density waves simultaneously. The dynamic range of an APD is several orders higher than that of a photomultiplier tube (PMT), which eliminates the need of multi-step system gain control. However, the internal gain of the APD we are using is about 3 orders lower than an ordinary PMT. Efforts have been made to suppress the feed through interferences from the transmission part to the reception part so as to reduce the errors in amplitude and phase measurements.

We present a combination of (1) single-wavelength optical mammograms representing either the optical absorbance (N-images) or its spatial second-derivative (N’’-images), and (2) oxygenation-index images based on optical data at four wavelengths (690, 750, 788, and 856 nm). The enhanced contrast of the second-derivative images with respect to the absorbance images is complemented by the functional information provided by the oxygenation-index images. In a number of clinical cases, we have found that breast cancer is often associated with relatively low oxygenation indices, while benign breast lesions correspond to areas with relatively high oxygenation indices. However, the relative nature of the oxygenation-index presented in this work does not allow for a comparison of oxygenation values in different images, for which absolute oxygenation measurements would be needed.

Near Infrared (NIR) tomography has the potential for characterization of different tissue types based upon cellular and vascular alternations that are optically apparent. This is especially useful for characterizing cancerous regions within normal tissue due to the high available contrast. Reconstructed images from NIR light propagation measurements through the female breast hold promise of providing clinically useful information about the pathophysiologic change of the tissue. We have developed a fast three-dimensional finite element model and image reconstruction algorithm, NIRFAST (Near Infrared Frequency-Domain Absorption and Scatter Tomography) and have previously tested the results extensively against simulated and phantom data. The results have shown that the reconstructed images have good accuracy in recovering optical changes within the medium under investigation, and that with appropriate constraints, the calculated quantitative values agree well with the true values. Based on these algorithms, we present true three-dimensional images of the breast, from patient data. These images, which are reconstructed using NIR measurements over a range of wavelengths, provide additional information regarding the blood content and oxygen saturation distribution within the breast.

Spectroscopic and time-sliced two-dimensional (2-D) transillumination imaging methods were used to investigate ex vivo tumor and normal tissues of human breast and parotid gland. The experimental arrangement for time-sliced optical imaging uses 120-fs, 1 kHz repetition-rate, 800-nm light pulses from a Ti:sapphire laser system for sample illumination and a charge coupled device (CCD) camera coupled to a gated image intensifier for recording 2-D images. The spectroscopic imaging arrangement uses 1210-1325 nm tunable output of a Cr: forsterite laser for sample illumination, a Fourier space gate to discriminate against multiple-scattered light, and a near-infrared (NIR) area camera to record 2-D images. Images recorded with earlier temporal slices of transmitted light highlighted tumors, while those recorded with later slices accentuated normal tissues. When light was tuned closer to the 1203 nm absorption resonance of adipose tissues, a marked enhancement in contrast between the images of adipose and fibrous tissues was observed. A similar wavelength-dependent difference between normal and cancerous tissues was observed. These results correlate well with pathology and nuclear magnetic resonance based analyses of the samples. This work demonstrates the advantages of time-resolved spectroscopic approach for imaging tumors in body organs.

An instrument for time-resolved optical mammography was constructed and is regularly used in an ongoing clinical trial. The system operates with four pulsed diode lasers emitting at 683, 785, 913, and 975 nm. The breast is slightly compressed between two Plexiglas plates and scanned along the X and Y directions. Light transmitted through the breast is collected by a 5 mm diameter optical bundle and coupled into two compact photomultipliers with enhanced sensitivity in the visible and near infrared spectral region, respectively. A multichannel PC board for time-correlated single photon counting is used for signal processing. Continuous acquisition and on line display of the processed images permit optimal measurement efficiency and control of the operation. A typical optical examination complies measurements obtained on both breasts in craniocaudal and oblique projections, with a typical measurement time of 5 minutes for each scan. In this paper we demonstrate the feasibility of in vivo time-resolved optical mammograms acquired at wavelengths longer than 900 nm in the lipid and water absorption region. The results of a clinical trial performed on 101 patients are shown, demonstrating capability to discriminate between benign and malignant breast lesions on the basis of different spectral behaviors. Furthermore, a non-linear perturbation model was successfully applied to the in vivo characterization of breast lesions. Preliminary results on tissue constituent determination (in term of blood volume, blood oxygen saturation, lipids and water content) were presented.

We present results of ongoing research in 3-D fluorescence tomography on large clinically-relevant tissue-mimicking domains. Finite element predictions of excitation and emission phase shift and amplitude attenuation are compared to experimental data from both column-shaped and breast-shaped tissue mimicking phantoms containing embedded fluorophore; system noise and measurement noise are characterized and utilized in image reconstruction using the Bayesian APPRIZE algorithm.

Optical fluorescence tomography recovers the spatial distribution of light emitting fluorophores inside a highly scattering medium. The quantification of a non-uniform quantum yield and fluorophore absorption distribution is of major interest in molecular imaging of biological tissue. We have developed a fluorescence image reconstruction code that is based on the particle transport equation. Since the algorithm does not rely on the diffusion approximation it promises to yield more accurate results in highly absorbing media or media with small geometries. We show that the code can be employed in a two-stage reconstruction process to obtain images of the fluorophore absorption and of the quantum yield.

In this paper we describe a Monte Carlo simulation for time resolved fluorescence. In the past information on steady state measurements have been reported. However we feel that a lot more information and insight could be gained by the use of time resolved fluorescence spectroscopy. We have developed a Monte Carlo simulation to study the fluorescence signal generated by fluorophores distributed in a scattering medium. The simulation uses a semi-infinite medium with a thickness of 1cm. We have used the simulation to study the effect of the change in optical properties of the medium on the TPSF (temporal point spread function) generated. We have also investigated the effect of the increased radial separation of the detector on the TPSF. We have observed a shift in the Tmax (time at which the peak intensity is reached) in accordance with diffusion theory. We wanted to validate our simulation by seeing how well we could derive the optical properties of the medium from the TPSF produced from simulation. We fitted the TPSF to an adjusted form of the diffusion theory to find scattering coefficient, μ_s, and we have used an analytical model of time resolved fluorescence to extract the absorption coefficient, μ_a. The results obtained were better than previously reported.

A reconstruction technique that allows for real-time recovery of a time series of images is presented. The reconstruction technique is intended for use in a dynamic DC imaging system, and uses a model-based normalized-preconditioned-transformed system equation that considers light diffusion and is suitable for structures having arbitrary geometry and composition. The algorithm was tested on synthetic 2D media containing dynamic optical contrast features, with the presence of added noise and other uncertainties that commonly are present in experimental data. Post-reconstruction analyses performed on the reconstructed image time series show that the algorithm is capable of correctly locating the inclusions (total of eight) with excellent spatiotemporal accuracy, even in the presence of 32% noise and other uncertainties. In addition, a semi-analytic graphical tool we have identified, that can guide design of suitable measurement geometries and selection of appropriate refinement of the imaging operator, is described.

Recently, there has been a great deal of interest in further improving the in-vivo characterization of tumors by interrogating tissues with multiple simultaneous image based measurement techniques that provide complementary information about the functional state of the tissue under investigation. Besides offering the possibility of complementary information, multi-modality techniques could also be used to improve the measurements done by either modality as well as providing cross validation measurements. Magnetic resonance imaging (MRI) and near-infrared diffuse optical tomography (DOT) are two techniques that, when combined, provide complementary structural and functional information. MRI can be used to obtain detailed structural and metabolic information regarding tumors. On the other hand, DOT can be used to determine local quantitative information regarding tumor composition and metabolism. Thus, a combined system consisting of MRI and DOT has the potential to enhance understanding of the complex biological processes in tumors. In this work, we describe the design of the hybrid DOT-MRI system.

Optical imaging can be used for localizing the oxygenation changes in the cortex in response to certain stimuli such as visual or sensory tasks, with the advantages of flexibility to subject's movement as well as being cheap and very fast. Up to now, data from optical imager is simply presented as a two-dimensional (2-D) topographic map rather than being tomographically reconstructed onto the cerebral cortex, based on the assumptions that the optical properties beneath each optode pair are homogeneous and the modified Beer-Lambert law can be used. Due to the high heterogeneity of optical properties in the brain, the above assumptions are evidently invalid, leading to both low spatial resolution and inaccuracy in the assessment of hemodynamic changes. To solve the problem, we propose a nonlinear image reconstruction algorithm for a two-layered slab geometry using time-resolved reflected light and demonstrate its advantages in quantifying simulated changes in hemoglobin concentrations. The algorithm is based on the previously developed generalized pulse spectrum technique, and implemented within a semi three-dimensional (3D) framework, where the changes of optical properties assumed invariable in depth, to conform to the topographic visualization and to reduce computational load. We also investigate the robustness of the algorithm to the uncertainties in the cortical structure and optical properties.

We describe a method to increase the convergence speed for simultaneous reconstruction of absorption and scattering images in Diffuse Optical Tomography (DOT). We used the diffusion approximation of the radiative tranfer equation and the Finite Element Method (FEM) to solve the forward problem. The absorption and reduced scattering images were reconstructed by inverting the distribution of the moments of the time-dependent detected light flux. The inverse problem is solved with an optimization algorithm such as ART or Conjugate Gradient. This ill-posed inverse problem can be simplified by using a priori knowledge of the objects studied. In this paper, we consider that DOT is a functional imaging technique that can be complemented by an anatomical imaging technique such as Magnetic Resonance Imaging (MRI). The algorithm is built as follows: the first step consists in to dividing the observed zone into regions, thanks to another imaging technique such as MRI. In practice, this allows to adapt the mesh to the internal geometry. Then, supposing that each region is homogeneous absorption- and scattering-wise, a few parameters are reconstructed with an optimization technique. With a few iterative steps, well-averaged parameters can be obtained, which could be used to initialize the first stage of a global process. This process could reconstruct smaller inhomogeneities. We compare this method with direct global reconstruction, beginning with homogeneous parameters.

An exact solution of the time dependent diffusion equation for a layered diffusive medium has been obtained by using a method based on the decomposition of the fluence rate in a series of eigenfunctions. Making use of this exact analytical solution inversion procedures have been implemented to retrieve the optical properties of the medium. Time resolved measurements of diffuse reflectance have been generated with Monte Carlo simulations. Inversion procedures have been carried out both for a single time resolved measurement and for multi-distance time resolved measurements. The characteristics of both procedures have been investigated. For the case of a two-layered medium the fitting procedure carried out on a single measurement showed that it is impossible to retrieve correctly all the optical parameters of the layered medium. The inversion procedure correctly retrieves the absorption coefficient of the second layer and the reduced scattering coefficient of the first layer. A significant improvement is obtained by using multi-distance measurements. A particular attention has been devoted to determine the optical properties of muscle with the subcutaneous fat layer.

A linear forward model of light propagation in turbid media for three dimensional (3D) optical tomography is formed based on the analytical cumulant solution of the Boltzmann radiative transfer equation., for time resolved, CW, and frequency-domain measurements in parallel slab geometry. This forward model is more accurate than the forward model based on the diffusive approximation. An inverse algorithm is developed, based on a fast 3D hybrid-dual-Fourier tomographic approach using multiple detectors and multiple sources in parallel geometry, which incorporates the cumulant radiative transfer model. The inverse algorithm runs fast that can produce a 3D image of a turbid medium with more than 20,000 voxels in 1-2 minutes using a PC computer. A 3D image is presented using simulated data.

In this work, we analyze the image reconstruction problem within the first order Rytov approximation and optimize the image reconstruction by looking at the singular value spectrum of the "weight matrix". We show that the optimal data is from the tail of the temporal intensity autocorrelation curve by investigating the condition number. Furthermore, by doing a standard L-curve analysis, we find the optimal regularization constant (Tikhonov regularization). Images from simulated data are shown to illustrate the advantage of the optimization process.

This investigation explores the effect of index of refraction, as an optical property, on light transport through optically turbid media. We describe a model of light propagation that incorporates the influence of refractive index mismatch at boundaries within a domain. We measure light transmission through turbid cylindrical phantoms with different distributions of refractive index. These distributions approximate the heterogeneous, layered nature of biological tissue. Finite element method model calculations of diffuse transmittance through these phantoms show good agreement with the trends observed experimentally. We see that phase measurements of light that propagates through approximately 90 (mm) of scatter-dominated media may vary by 10 degrees depending upon the values of refractive index of the medium. Amplitude measurements are not as sensitive to this parameter as phase. Model calculations of diffuse reflectance from a semi-infinite slab geometry with different layers also shows good agreement with Monte Carlo simulations. We conclude that incorporating refractive index into light transport models may be worthwhile. Applying such a model in tomographic image reconstruction may improve the estimation of optical properties of biological tissues.

We present the method which can calculate the spectral reflectance from physical parameters corresponding to the pathological features, e.g. average size of cell nuclei and standard deviation of cell nuclear size distribution, in consideration of multiple scattering in biological tissue. In this paper, the method combined the Monte Carlo method which simulates multiple scattering effects and the Mie theory which provides phase function (angular properties of light scattering) and scattering coefficient was employed. In order to investigate the validity of this method, the calculated spectra by the method and Monte Carlo method with Henyey-Greenstein phase functions were compared with measurement spectra derived from the tissue phantom whose size distribution has double peaks. From the results, it is shown that the method can better predict the spectral reflectance of tissue phantom rather than Monte Carlo method with Henyey-Greenstein phase function.

Diseased tissue may be specifically marked by an exogenous fluorescent marker and then, following laser activation of the marker, optically and non-invasively detected through fluorescence imaging. Interaction of a fluorophore, conjugated to an appropriate antibody, with the antigen expressed by the diseased tissue, can indicate the presence of a specific disease. Using an optical detection system and a reconstruction algorithm, we were able to determine the fluorophore’s position in the tissue. We present 3D reconstructions of the location of a fluorescent marker, FITC, in the tongues of mice. One group of BALB/c mice was injected with squamous cell carcinoma (SqCC) cell line to the tongue, while another group served as the control. After tumor development, the mice’s tongues were injected with FITC conjugated to anti-CD3 and anti-CD 19 antibodies. An Argon laser excited the marker at 488 nm while a high precision fluorescent camera collected the emitted fluorescence. Measurements were performed with the fluorescent marker embedded at various simulated depths. The simulation was performed using agarose-based gel slabs applied to the tongue as tissue-like phantoms. A biopsy was taken from every mouse after the procedure and the excised tissue was histologically evaluated. We reconstruct the fluorescent marker’s location in 3D using an algorithm based on the random walk theory.

We have developed a novel imaging system for determining the localization of tumors labeled by fluorescent contrast agents and embedded several centimeters inside the highly scattered medium. This frequency-domain system utilizes the phased cancellation configuration with a goniometric probe. The instrumentation performance on the phantom test can detect 3 mm diameter sphere filled with 1 nM fluorescent dye, Indocyanine Green (ICG), and 3 cm deep inside the scattering medium with similar optical properties as human breast tissue within a 1 mm localization confidence. Mouse tumor model immersed in appropriate scattering/absorbing medium is used for animal test. Intra-tumor injection of ICG demonstrates the localization of the tumor (5 mm in diameter) submerged 3 cm deep inside the highly scattered medium with 2 mm position error. Results with NIR804-D-Glucosamide on the AR42J tumor bearing nude mouse are also presented with 3 mm localization error. The accuracy of the localization suggest that this system would be helpful to guide the clinical fine-needle biopsy for early breast cancer detection.

Current state of the knowledge in the field of optical diagnostics indicates that application of both steady state and time-resolved fluorescence spectroscopy permits the discrimination among normal and pathological tissues. Most data analysis algorithms, however, separately address either the spectral profile of the fluorescence emission or the time domain features of fluorescence decay of tissue sample. We present a novel system identification technique that allows for intrinsic Fluorescent Impulse Response Kernel (FIRK) estimation in combined time and wavelength spaces. The fluorescence experiment is described as a discrete two-dimensional single input - single output system. Modeling of FIRK is based on the characteristics of spectro-temporal fluorescence waveforms. Taking advantage of the exponentially decaying temporal behavior of the fluorescent output, we use a linear combination of Laguerre functions to model the fluorescence response as a function of time. To address the large variability of spectral profiles from different biological samples, we use the discrete Fourier series expansion to model the variation of fluorescence intensity across wavelength. The validation of the proposed approach is performed on simulated fluorescence data and fluorescence data measured from fluorescence lifetime standards. The algorithm is applicable for an effective recovery of intrinsic FIRK from any excitation light probe spectro-temporal waveforms. The retrieved FIRK can be further used for modeling and/or classification of pathological and physiological conditions of tissues.

Near-infrared imaging can be used in humans to characterize changes in breast tumor tissue by imaging total hemoglobin and water concentrations as well as oxygen saturation. In order to improve our understanding of these changes, we need to carefully quantify the range of variation possible in normal tissues for these parameters. In this study, the effect of the subject’s menstrual cycle was examined by imaging their breast at the follicular (7-14 days of the cycle) and secretory phases (21-28 days of the cycle), using our NIR tomographic system. In this system, a three layer patient interface is used to measure 3 planes along the breast from chest wall towards the nipple at 1cm increments. Seven volunteers in their 40s were observed for 2 menstrual cycles and all of these volunteers recently had normal mammograms (ACR 1) with heterogeneously dense breast composition. The results show that average total hemoglobin in the breast increased in many subjects between 0 to 15% from the follicular phase to secretory phase. Oxygen saturation and water concentration changes between these 2 parts of the cycle were between −6.5% to 12% for saturation and between −33% to 28% for water concentration. While the data averaged between subjects showed no significant change existed between phases, it was clear that individual subjects did exhibit changes in composition which were consistent from cycle to cycle. Understanding what leads to this heterogeneity between subjects will be an important factor in utilizing these measurements in clinical practice.

We have introduced a three-dimensional (3-D) ultrasound imaging system for use in a combined ultrasound and near-infrared (NIR) imager for breast imaging. Compared with commercial ultrasound scanners which can only provide a single 2-D cross-section image at one scan, this 3-D system can provide 3-D co-registered ultrasound images for constraining regions of interest for 3-D NIR imaging reconstruction at one scan. The 3-D high-resolution ultrasound system consists of a state-of-the-art 1.75D ultrasound array of 1280 elements manufactured by Tetrad Inc. We have developed corresponding circuitry to drive the 1.75D array. The circuitry consists of 1280 parallel transmission channels, a 1280-to-40 multiplexer and 40 parallel receiving/data acquisition channels. A Texas Instrument DSP board has been embedded to speed up on-board data processing. We have tested the 3-D ultrasound imaging system with a 3-D ultrasound calibration phantom.

We report the development of a general, quantitative, and localized visible light clinical tissue oximeter, sensitive to both hypoxemia and ischemia. Monitor design and operation were optimized over four instrument generations. A range of clinical probes were developed, including non-contact wands, invasive catheters, and penetrating needles with injection ports. Real-time data were collected (a) from probes, standards, and reference solutions to optimize each component, (b) from ex vivo hemoglobin solutions co-analyzed for StO₂% and pO₂ during deoxygenation, and (c) from normoxic human subject skin and mucosal tissue surfaces. Results show that (a) differential spectroscopy allows extraction of features with minimization of the effects of scattering, (b) in vitro oximetry produces a hemoglobin saturation binding curve of expected sigmoid shape and values, and (c) that monitoring human tissues allows real-time tissue spectroscopic features to be monitored. Unlike with near-infrared (NIRS) or pulse oximetry (SpO₂%) methods, we found non-pulsatile, diffusion-based tissue oximetry (StO₂%) to work most reliably for non-contact reflectance monitoring and for invasive catheter- or needle-based monitoring, using blue to orange light (475-600 nm). Measured values were insensitive to motion artifact. Down time was non-existent. We conclude that the T-Stat oximeter design is suitable for the collection of spectroscopic data from human subjects, and that the oximeter may have application in the monitoring of regional hemoglobin oxygen saturation in the capillary tissue spaces of human subjects.

Photonify’s tissue spectrometer uses Near-Infrared Spectroscopy for real-time, noninvasive measurement of hemoglobin concentration and oxygen saturation [SO₂] of biological tissues. The technology was validated by a series of ex vivo and animal studies. In the ex vivo experiment, a close loop blood circulation system was built, precisely controlling the oxygen saturation and the hemoglobin concentration of a liquid phantom. Photonify’s tissue spectrometer was placed on the surface of the liquid phantom for real time measurement and compared with a gas analyzer, considered the gold standard to measure oxygen saturation and hemoglobin concentration. In the animal experiment, the right hind limb of each dog accepted onto the study was surgically removed. The limb was kept viable by connecting the femoral vein and artery to a blood-primed extracorporeal circuit. Different concentrations of hemoglobin were obtained by adding designated amount of saline solution into the perfusion circuit. Photonify’s tissue spectrometers measured oxygen saturation and hemoglobin concentration at various locations on the limb and compared with gas analyzer results. The test results demonstrated that Photonify’s tissue spectrometers were able to detect the relative changes in tissue oxygen saturation and hemoglobin concentration with a high linear correlation compared to the gas analyzer

A study using pregnant sheep was designed to simulate fetal hypoxia in order to investigate the ability of near-infrared spectroscopy (NIRS) to detect and quantify fetal hypoxia in utero. The near-infrared spectroscopic probe consisted of two detectors and six source positions. It was placed on the maternal ewe abdomen above the fetal head. The light sources were modulated at 70 MHz and frequency-encoded so that simultaneous measurements at 675, 786, 830 nm for each source position were possible. After the baseline measurements, fetal hypoxia was induced by blocking the aorta of pregnant ewe and thus compromising the blood supply to the uterus. Blood gas samples were concurrently drawn from the fetal brachial artery and jugular veins. Analysis of the diffuse optical data used a two-layer model to separate the maternal layer from the fetal head. The analysis also employed a priori spectral information about tissue chromophores. This approach provided good quantification of blood oxygenation changes, which correlated well with the blood gas analyses. By contrast the homogeneous model underestimated oxygenation changes during hypoxia.

In a previous report, we showed that there was distinct heterogeneity in tumor oxygenation dynamics even within the same tumor monitored by F MR pO₂ mapping and near-infrared spectroscopy (NIRS). In this study, we applied a three-channel NIRS system to reveal dynamic heterogeneity of tumor vascular oxygenation during respiratory challenges. When the implanted tumors reached ~3 cm, three photo detectors were attached to the surface of the tumor in three different positions to measure changes in oxygenated hemoglobin concentration ([HbO₂]), while the inhaled gas was alternated between air and carbogen (95% O₂ + 5% CO₂). Significant changes in tumor oxygenation were observed at all three detectors accompanying the respiratory challenge, and these changes could be modeled with two exponential components with fast and slow time constants. Time constant, amplitude, vascular coefficient, and the ratio of perfusion rate were obtained from fitting curve of Δ[HbO₂]. These parameters revealed that there were indeed different responses of tumor vascular oxygenation during carbogen inhalation at the three different locations. These results clearly reveal that tumors are highly heterogeneous during vascular oxygenation in response to carbogen inhalation. This study demonstrates that the NIR technology can provide an efficient, real-time, non-invasive way for monitoring tumor physiology, and it may have prognostic value, providing insight into tumor vascular development and angiogenesis.

It has been observed for a long time that the compressive pressure applied on the tissue significantly effects the measurement of Near-Infrared Spectroscopy. The clinical study implies a certain level of correlation between the tissue type and its response to pressure. However, this correlation is not quite understood nor quantified yet, especially in the diffuse optical imaging arena. This paper discusses the development of NIRS tissue spectrometer with real time pressure monitoring and a series of pressure regulated experiments we have conducted on human subjects. By analyzing the pattern of the 2D mapping of tissue hemoglobin concentration and oxygen saturation in response to the temporal profile of the compression force, we are able to explore major factors that influence the pressure regulated measurement

Near Infrared diffuse optical tomography was used to measure the molar absorption spectra for de-oxyhemoglobin, oxyhemoglobin and water through phantom experiments. The measured spectra compensate for any systemic errors and errors arising from model-based estimation of concentrations of these chromophores, and show reasonable agreement with existing literature values. The spectra were validated by phantom experiments to obtain the oxygen dissociation curve. Variation of the partial pressure of oxygen in solution yielded results consistent with theory of mean error of 7.6% in estimation of oxygen saturation, and accurate within 5% in estimation of hemoglobin.

Three oxygen-sensitive parameters (arterial hemoglobin oxygen saturation SaO₂, tumor vascular oxygenated hemoglobin concentration [HbO₂], and tumor oxygen tension pO₂) were measured simultaneously by three different optical techniques (pulse oximeter, near infrared spectroscopy, and FOXY) to evaluate dynamic responses of breast tumors to carbogen (5% CO₂ and 95% O₂) intervention. All three parameters displayed similar trends in dynamic response to carbogen challenge, but with different response times. These response times were quantified by the time constants of the exponential fitting curves, revealing the immediate and the fastest response from the arterial SaO₂, followed by changes in global tumor vascular [HbO₂], and delayed responses for pO₂. The consistency of the three oxygen-sensitive parameters demonstrated the ability of NIRS to monitor therapeutic interventions for rat breast tumors in-vivo in real time.

A portable instrument for tissue oximetry based on time-resolved reflectance spectroscopy was developed. The output pulses of 2 laser diodes (683 and 785 nm, 80 MHz pulse repetition rate, 1 mW average power, 100 ps FWHM) are delayed and coupled into a multimode graded-index fiber (50/125 μm and injected into the tissue. The reflectance photons are collected by 8 independent 1 mm fibers and detected by a 16-anode photomultiplier. A time-correlated single photon counting PC board is used for the parallel acquisition of the curves. Simultaneous estimate of the transport scattering and absorption coefficients is achieved by best fitting of time-resolved reflectance curves with a standard model of Diffusion Theory. The performances of the system were tested on phantoms in terms of stability, reproducibility among channels, and accuracy in the determination of the optical properties. Preliminary in vivo measurements were performed on healthy volunteers to monitor spatial changes in calf (medical and lateral gastrocnemius) oxygen hemoglobin saturation and blood volume during dynamic plantar flexion exercise.

We report results of a feasibility clinical evaluation of the near-infrared Photonify Tissue Spectrometer in detection of peripheral vascular disease (PVD). In the evaluation, using a blood pressured cuff, we measure changes in oxygen saturation (S_tO₂) during a temporary occlusion of blood inflow. We use the post-occlusive S_tO₂ recovery rate and the time needed for 80% recovery as parameters to differentiate healthy subjects from subjects with PVD, and obtained high diagnostic accuracy in 19 subjects enrolled in the clinical evaluation. This may suggest that the tissue spectrometer could be useful in diagnosis of PVD.

To date, the applications of near infrared (NIR) diffusion optical tomography (DOT) are mostly focused on the potential of imaging woman breast, human head hemodynamics and neonatal head. For the neonates, who are suffered from ischaemia or hemorrhages in brain, bedside monitoring of the cerebral perfusion situation, e.g., the blood oxygen saturation and blood volume, is necessary for avoiding permanent injure. NIR DOT is on the promising tools because it is noninvasive, smaller in size, and moveable. Prior to achieving the ultimate goal of imaging infant brain and woman breast using DOT, in this paper, the developed methodologies are justified by imaging in vivo human forearms. The absolute absorption- and scattering-coefficient images revealed the inner structure of the forearm and the bones were clearly distinguished from the muscle. The differential images showed the changes in oxy-hemoglobin, deoxy-hemoglobin and blood volume during the hand-gripping exercises, which are consistent with the physiological process reported on literatures.

A hybrid instrument combining near infrared and diffuse correlation spectroscopies was used to measure muscle oxygenation and blood flow dynamics during cuff occlusion and ischemia. Measurements were done on six healthy subjects on their arm and leg flexor muscles. Hemodynamic response was characterized for blood oxygen saturation, total hemoglobin concenration and relative blood flow speed. The characterization allowed us to define the normal response range as well as showing the feasibility of using a hybrid instrument for dynamic measurements.

Background: Repetitive Strain Injury (RSI) is a major problem in nowadays health care and creates high financial costs and personal distress. Average prevalence rates in the Netherlands vary from 20-40% of the working population. Insight into the patho-physiological mechanism of RSI is important in order to establish adequate treatment and prevention programs. Objective: The aim of this study was to gain insight in muscle oxygen consumption (mVO₂), blood flow (BF), and reoxygenation (ReOx) in the forearm of computer workers with stage III Repetitive Strain Injury (RSI). Method: We have used continuous wave infrared spectroscopy (NIRS) to measure these variables. Measurements were conducted on the extensor and flexor muscle in both arms as well in RSI-patients (n=10) as in control subjects (n=21). A protocol of increased isometric repetitive contraction in a handgrip ergonometer was used with increasing levels of strength. Results: mVO₂ in the extensor muscle in RSI-subjects (dominant side) was increased compared to control subjects and compared to the non-dominant side (p<0.05). ReOx was not increased in RSI (dominant side-extensor muscle). However, there was a tendency towards statistical significance (p=0.065). BF in rest was equal in both groups, however after exercise it tended to be increased. Half-time recovery (T _½) was measured during only one part of the protocol and it was significantly increased (p<0.05). Conclusion: mVO₂ in RSI is impaired. BF and ReOx did not show difference between both groups. Future research should aim at a microvascular dysfunction in RSI.

Optically Tomography within highly scattering material has focused on Coherence Domain and Time Domain methods: both detecting the shorting path photons over the dominant randomly scattered background light. Angular Domain Imaging instead uses collimators, small acceptance angle filters, to observe only those photons closely aligned to a laser light source. A linear collimating array was fabricated using silicon surface micromachining consisting of 51 μm wide by 10 mm long etched channels with 102 μm spacing very high aspect ration (200:1) 20 mm wide array. With careful array alignment to a laser source, restricted to a linear beam, the unscattered laser light passes directly through the channels to a CCD detector, and the channel walls absorb the scattered light at angles >0.29 degrees. With a computer controlled Z axis objects within a 1 cm thick scattering material were scanned quickly. High contrast 150 μm lines/spaces at the medium front were observed at scattered to ballistic photon ratios >5×10⁵:1 with a 10 mm beam. Narrowing the beam to 130 μm width produces detectable images >3×10⁸:1. Objects closer to the detector were more visible, and mid point objects were detectable >10⁹:1. Smaller channels and longer arrays should enhance detection by factors of >100.

The image quality of optical tomography depends heavily on the amount of independent information and the signal to noise ratio. Time-resolved photon migration imaging (TPMI) measures a temporal point spread function (TPSF) of light diffusing through a turbid medium. The temporal profile can generally provide more useful information for accurate image reconstruction than continuous wave (CW) and frequency domain (FD) measurements, providing that adequate signal to noise ratio has been achieved. Currently, measurement of TPSF is most commonly realized with the time-correlated single photon counting (TCSPC) technique. Nonetheless, the principle of single photon counting requires that no more than one photon is detected in each cycle. This usually results in a too long data acquisition time, or a rather low signal to noise ratio in a short time. We propose a more efficient coded excitation and correlation detection mechanism that borrows ideas from spread spectrum communications. A train of excitation pulses modulated by a pseudo-random bit sequence is used as the light source, while a cross-correlation scheme is used to retrieve the impulse response. We have simulated a TPMI system and evaluated its performance. It has been demonstrated that the data acquisition time can be reduced by several orders, in comparison with a TCSPC system.

Optical imaging through turbid media has gathered a great interest as a non-invasive method to localize tumors inside the body. A step forward in this direction could be provided by a parallel detector that allows one to acquire a huge amount of data in a short time. This could result in a more stable reconstruction of the optical properties of the sample and in a more reliable diagnosis. We used a system based on a fast gated intensified CCD camera capable of a gate width <200 ps to acquire time sliced images of the light transmitted though a phantom simulating the optical properties of the human breast. Light pulses emitted by a laser diode (633 nm) were delivered to the central position of the phantom, while the output plane was imaged by the ICCD (fan geometry). Experimental data were analyzed using an analytical expression for homogeneous medium and a perturbative model, both derived from the Random Walk Theory (RWT). By using the perturbative approach we evaluated the depth (along Z axis) of scattering inclusions and their scattering coefficient.

In medical instrumentation, optical topography (OT) refers to the use of near-infrared spectroscopy for measuring brain function in systems. Arrays of optical fibers are attached to the scalps of subjects; infrared light is passed through the fibers, and changes in the reflections depict blood-volume changes in the cortex. In this study, the spatial resolution and locational accuracy of topographical images obtained by three arrangements of optical fibers was analyzed through simulation. Three arrangements, a "lattice arrangement" (LA), "double-density arrangement" (DA), and "quadruple density arrangement" (QA) were investigated. The density of spatial-sapling points is higher in the DA and QA than in the LA, i.e. the distance between sampling points for these arrangements were 21, 15 and 11 mm, respectively. The efficacy of these arrangements was evaluated. An adult head-structure phantom was prepared. The absorption coefficient in the phantom was varied to simulate brain activation in the cortex, and the resulting absorbance change (ΔOD) was thus obtained. The 'activated' area in the overall measurement area was fixed and the ΔOD at each of the sampling point in each arrangement was obtained. The resulting distributions of ΔOD were spatially interpolated to obtain topographical images. The spatial resolution and locational accuracy was obtained for each of the images; the results indicated that the DA is the most efficacious of the three arrangements. An experimental DA-OT system was then built. Topographical images of motor-function activation obtained by this system and a commercial LA-based system were compared; the DA-OT system provided the higher spatial resolution.

We utilized the nude mouse model bearing the human hepatoblastoma G2 (HepG2) tumor and B-16 Murine Melanoma tumor to study the delivery and detection of the reconstituted Pyropheophorbide Cholesterol Oleate (r-pyroCE) molecular beacon. The delivery vehicle, low-density lipoprotein (LDL), labeled with the porphyrin derivative, was employed in response of the overexpression of LDL receptors in the HepG2 tumor. The B-16 melanoma tumor was also observed in this study for its overexpression of the LDL receptors. The tumors were imaged using the 3D low temperature scanner to produce images throughout several sliced sections of each tumor. The fluorescence signal of the pyropheophorbide was detected at 720nm when excited at 670nm in the tumor tissue. The uniform distribution of the signal in the HepG2 tumor shows extravasation of the beacon from the blood vessels. The B-16 tumor did not exhibit strong fluorescent signals and successful delivery as the HepG2 tumor outside the blood vessels and into the tumor tissue.

Proteases are involved in the invasion and metastasis of tumor cells. Cathepsin B overexpression has been shown in some neoplastic tissues. This study assesses the expression of Cathepsin B in the human fibrosarcoma (HT1080) in the mouse model by near-infrared (NIR) imaging. The nude mice were intravenously injected "a stealth probe" - an activable Cathepsin B sensing near-infrared fluorescence (NIRF) probe (24 hours before sacrifice) and the dye Cy5.5 (30 seconds before sacrifice). The animals were freeze-trapped and NIR images were obtained by the low temperature NIR scanner at the following excitation-emission wavelength pairs: 366, 450nm (NADH), 436, 520nm (FAD), and 670, 695nm (Cathepsin B probe). After imaging, the samples were submitted for histopathological evaluation. The tumor redox ratio NADH/(NADH+FAD) increased significantly because of the hypoxic state of tumor tissue with respect to normal tissue. The Cathepsin B probe was uniformly distributed throughout the tumor. This study indicated the efficient usage of the Cathepsin B probe in the molecular imaging for the detection of the early stage tumors.

Many biomedical applications of time-resolved transmittance require precise measurements of optical properties of scattering samples. Recently, random walk theory has been proposed to accurately model the time-resolved transmittance and reflectance across uniform slabs of scattering media as well as slabs containing isolated targets of different scattering or absorbing properties. In this paper random walk description of light propagation in turbid media has been used to analyse transmittance measurements for optical characterization of scattering and absorbing media. A femtosecond Ti:Sa laser working at 800 nm and a streak camera with a time resolution of a few picoseconds have been employed. Various Intralipid suspensions with and without absorbing ink have been investigated. These studies have experimentally confirmed that random walk theory can be usefully employed to discriminate small changes in absorbing and scattering properties. An appropriate fitting procedure, without arbitrary constrains, have been checked and used. In addition, the detailed investigation here reported can help in defining the applicability range of random walk description. This analysis can be important since random walk is a powerful and versatile theory as fast and efficient solution of the inverse problem in optical tomography also for sophisticated geometry and for a wide range of optical properties.

The primary objective of this work was to use FT-Raman spectroscopy to detect spectral changes between benign and basal cell carcinoma (BCC) skin tissues. Those spectral changes can give us important information about the biochemical alterations between these two types of tissues. The early cancer detection stills a great challenge in clinical oncology. Recently, Raman spectroscopy has been used for skin lesion detection. FT-Raman spectroscopy is a modern analytical tool and its use for cancer diagnosis will lead to several advantages for the patient as, for example, real time and less invasive diagnosis. We have analyzed by FT-Raman eight sets of samples histopathologically diagnosed as BCC and made a comparison with five sets of samples diagnosed as benign tissue. We have found that the main spectral differences between these samples were in the shift region of 1220-1300cm^-1 and 1640-1680cm^-1. These vibration bands correspond to the amide III and to the amide I vibrations, respectively. Principal component analysis applied over all 13 samples could identify tissue type with 100% of sensitivity and specificity.

The optical noninvasive diagnostic systems are now widely applied and investigated in different areas of medicine. One of the such techniques is the noninvasive spectrophotometry, the complex diagnostic technique consisting on elastic scattering spectroscopy, absorption spectroscopy, fluorescent diagnostics, photoplethismography, etc. Today a lot of real optical diagnostic systems indicate the technical parameters and physical data only as a result of the diagnostic procedure. But, it is clear that for the medical staff the more convenient medical information is needed. This presentation lights the general way for development a diagnostic system’s software, which can produce the full processing of the diagnostic data from a physical to a medical level. It is shown, that this process is a multilevel (3-level) procedure and the main diagnostic result for noninvasive spectrophotometry methods, the biochemical and morphological composition of the tested tissues, arises in it on a second level of calculations.

In many studies on functional neuroimaging, change in local cerebral blood flow (CBF) induced by sensory stimulation is used as a substitutive marker for change in cortical neuronal activity by Roy and Sherrington’s postulation. The purpose of this study is to evaluate the close temporal relationship between evoked local CBF and neuronal activity through utilizing temporal clustering analysis (TCA) method to analyze the dataset obtained by high-resolution laser speckle imaging (LSI). We mapped a relatively large somatosensory area of cerebral cortex in successively and followed variations of cerebral blood flow under sciatic nerve stimulation in anesthetized rats using LSI technique. Then we illustrated TCA method on the optical imaging signals and investigated the temporal characteristics of CBF activation. Contralateral somatosensory cortical microflow was activated to increase promptly in less than 1 second after the onset of 2 seconds electrical stimulation (pulse parameters: 0.5ms, 5Hz, 350mV). The data of the present study may possibly shed light on the development question of functional magnetic resonance imaging (fMRI) whether the ultimate spatial resolution attainable by fMRI is limited by the physiology of the cerebral blood flow response to activation or by the hardware and signal-to-noise ratio.

We have measured relative diffuse reflectance images on tissue-like liquid phantoms using a laser diode (784 nm, 40 mW) as a broad-beam source and a CCD camera detector. We have found that, to a good approximation, the result that the reflectance R only depends on the ratio of the absorption and reduced scattering coefficients of the medium (μ_a/μ_s') can be extended to geometrical configurations that are more general than those considered by theoretical studies. In fact, over the range 0.0028<μ_a/μ_s'<0.058 considered by us, the dependence of R on μ_a/μ_s' varies by no more than 8% by changing the incidence angle of illumination over the range 0-30°. The relative diffuse reflectance vs. μ_a/μ_s was well described by introducing an arbitrary normalization factor K in the analytic relationship for the absolute reflectance reported by S.L. Jacques [http://omlc.ogi.edu/news/may99/rd/index.html]: R=Kexp{-7.8/[3(1+μ_s'/μ_a)]}. The ratio of the relative diffuse reflectance values at two locations (for spatial changes) or at two times (for temporal changes) cancels out the unknown normalization factor K, so that this ratio can be used to obtain quantitative estimates of the corresponding change in μ_a/μ_s'. We have experimentally estimated that in our experimental conditions R is sensitive to a superficial volume about 2 mm deep and 4 mm wide.

Near Infrared Spectroscopy (NIRS) has been widely used in cancer imaging spectroscopy because cancer tissue has more blood volume and less oxygen than normal tissue. The neck squamous cell carcinoma is ideal for our study because it’s a surface mass can be easily feel through the skin. We use a simple homodyne phase modulation system, In-phase and Quadrature Phase (I&Q) Detection system, to study the curative effect of therapy in neck cancer patients. Clinical treatment includes mainly chemical and radiation therapy, both of which alter the blood volume and the oxygenation saturation in cancer tissue. The I&Q detection system is capable of the quantification measurement of those biological changes. In this paper, we will simply introduce the I&Q detection system’s principle and constitution, and mainly explain the analysis of patients’ data.

Many investigations in the literature have studied the migration of photons in fluorescent and non-fluorescent applications, both in the steady-state and time-resolved, within a tissue-simulating medium. However, none have addressed the specific issue of quantified the subsequent migration path of the emitted photons. This is important since fluorescent spectroscopy has been gaining acceptance as an important diagnostic tool. In this paper we show the migration patterns of fluorescent photons with respect to time in a scattering medium such as tissue. The images produced are of the paths of the emission photons that reach the detector. We are able to observe how they migrate at different times. We investigate the effect of the absorption and scattering coefficients on the migration patterns, and how it could give us information about methods to detect inhomogeneities in the medium. The images produced give us information about the accuracy of the estimation of the optical properties in particular medium. Finally, we study how the detector position and absorption and scattering coefficients affect the source of the photons that reach the detector. We discover that during the rising time of the temporal spectrum most of the detected fluorescent photons are being generated very close to the source and the path followed by these photons are localized near the detector. Therefore, we could explore this for the best orientation and location of the detector and source positions to locate inhomogenieties and also source of fluorescence photons within the medium.

The approximate extended Kalman filter (AEKF) has been suggested as an appropriate inverse method for reconstructing fluorescent properties in large tissue samples from frequency domain data containing measurement error. The AEKF is an “optimal” estimator, in that it seeks to minimize the predicted error variances of the estimated optical properties in relation to measurement and system errors. However, due to non-linearities in the recursive estimation process, the updates are actually suboptimal. Furthermore, the computational overhead is large for the full AEKF algorithm when applied to large datasets. In this contribution we developed three hybrid forms of the AEKF algorithm that may improve the performance in frequency domain fluorescence tomography. Numerical results of image reconstruction from actual frequency domain emission data show that one hybrid form of the AEKF outperforms the traditional full AEKF in both image quality and computational efficiency for the two cases tested.

Time-Resolved Laser-Induced Fluorescence Spectroscopy (tr-LIFS) offers the potential for intra-operative diagnosis of primary brain tumors. However, both the intrinsic properties of endogenous fluorophores and the optical properties of brain tissue could affect the fluorescence measurements from brain. Scattering has been demonstrated to increase, for instance, detected lifetimes by 10-20% in media less scattering than the brain. The overall goal of this study is to investigate experimentally and computationally how optical properties of distinct types of brain tissue (normal porcine white and gray matter) affect the propagation of the excitation pulse and fluorescent transients and the detected fluorescence lifetime. A time-domain tr-LIFS apparatus (fast digitizer and gated detection) was employed to measure the propagation of ultra-short pulsed light through brain specimens (1-2.5-mm source-detector separation; 0.100-mm increment). A Monte Carlo model for semi-infinite turbid media was used to simulate time-resolved light propagation for arbitrary source-detector fiber geometries and optical fiber specifications; and to record spatially- and temporally resolved information. We determined a good correlation between experimental and computational results. Our findings provide means for quantification of time-resolved fluorescence spectra from healthy and diseased brain tissue.

The method consists in measuring the perturbation provoked by a small volume (0.2-1 ml) of tissue on light propagating through a medium of known optical properties. The absorption and the reduced scattering coefficient of the medium are retrieved from multidistance continuous wave measurements of transmittance. The inversion procedure is based on the solution of the diffusion equation obtained with a perturbative approach. The method has been validated with Monte Carlo results. Experimental results are reported for various kind of tissue. The results obtained shows that the method is particularly effective for low absorbing tissues.

In this research, we investigate the imaging properties of tansillumination laser CT system based on the coherent detection imaging (CDI) method, for early diagnosis of rheumatoid arthritis, from the viewpoints of morphological and quantitative imaging. First, to investigate the morphological imaging ability, the index finger PIP joint of a healthy volunteer was imaged with the laser CT, X-ray CT, and MRI. By comparison of these images, we present that the laser CT delineates the bone region with high spatial resolution of sub-millimeters. Next, to investigate the quantitative imaging, three kinds of the physical phantoms simulating bone, made from polyurethane including different concentrations of CaCO₃, were imaged. We present that quantitative measurement is possible by demonstrating a satisfactory linear relationship between the averaged pixel value of the reconstructed images and the actual concentrations. These results demonstrate the feasibility of early diagnosis for rheumatoid arthritis.

The equation of radiative transfer can take into account the anisotropic scattering behavior of photons and anisotropic sources for modeling the light propagation in tissue. This is an important aspect when small tissue geometries are considered. In this case the solutions of the commonly applied diffusion approximation may provide only insufficiently accurate results. We numerically solve the equation of radiative transfer by means of a finite-difference discrete-ordinates technique. However, strong anisotropically scattering media require many discrete ordinates, which lead to a large computational burden. In this study we implemented a Delta-Eddington method that allows using only a small number of discrete ordinates, and the solution can be obtained at a lesser computational costs.

Diffuse optical imaging is an emerging modality that uses Near Infrared (NIR) light to reveal structural and functional information of deep biological tissue. It provides contrast mechanisms for molecular, chemical, and anatomical imaging that is not available from other imaging modalities. Diffuse Optical Tomography (DOT) deals with 3D reconstruction of optical properties of tissue given the measurements and a forward model of photon propagation. DOT has inherently low spatial resolution due to diffuse nature of photons. In this work, we focus to improve the spatial resolution and the quantitative accuracy of DOT by using a priori anatomical information specific to unknown image. Such specific a priori information can be obtained from a secondary high-resolution imaging modality such as Magnetic Resonance (MR) or X-ray. Image reconstruction is formulated within a Bayesian framework to determine the spatial distribution of the absorption coefficients of the medium. A spatially varying a priori probability density function is designed based on the segmented anatomical information. Conjugate gradient method is utilized to solve the resulting optimization problem. Proposed method is evaluated using simulation and phantom measurements collected with a novel time-resolved optical imaging system. Results demonstrate that the proposed method leads to improved spatial resolution, quantitative accuracy and faster convergence than standard least squares approach.

This study imaged the metabolic state of a growing tumor and the relationship between energy metabolism and the ability of glucose uptake in whole tumor tissue with cryo-imaging at 77° K. A MTB/TOM mouse model, bearing c-MYC-induced mammary tumor, was very rapidly freeze-trapped 2 hrs post Pyro-2DG injection. The fluorescence signals of oxidized flavoprotein (Fp), reduced pyridine nucleotide (PN), pyro-2DG, and the reflection signal of deoxy-hemoglobin were imaged every 100 μm from the top surface to the bottom of the tumor sequentially, 9 sections in total. Each of the four signals was constructed into 3D images with Amira software. Both Fp and PN signals could be detected in the growing tumor regions, and a higher reduction state where was shown in the ratio images. The necrotic tumor regions displayed a very strong Fp signal and weak PN signal. In the bloody extravasation regions, Fp and PN signals were observably diminished. Therefore, the regions of high growth and necrosis in the tumor could be determined according to the Fp and PN signals. The content of deoxy-hemoglobin (Hb) in the tumor was positively correlated with the reduced PN signal. Pyro-2DG signal was only evident in the growing condition region in the tumor. Normalized 3D cross-correlation showed that Pyro-2DG signal was similar to the redox ratio. The results indicated that glucose uptake in the tumor was consistent with the redox state of the tumor. And both Pyro-2DG and mitochondrial NADH fluorescence showed bimodal histograms suggesting that the two population of c-MYC induced mammary tumor, one of which could be controlled by c-MYC transgene.

A system for rapid and precise localization of breast cancer using low frequency phased array near-infrared technique has been developed. In this system, in-phase and out-of-phase of sine-wave signals at 3 kHz modulate two light emitting diode (LED) sources respectively. And amplitude cancellation of two photon density waves will occur in the silicon-detector at the mid-plane of two sources reference. Cancellation signal is displayed by an embedded amplitude and phase detector and is displayed acoustically. The whole system is pocket-size, handheld with the dimension of 6.0cm×5.0cm×2.5cm and the weight is 140g. Experiments with phantom and animals show that the system can provide real time detection and localization of small hidden absorbing-fluorescent objects inside the highly scattering medium at the depth of 2cm with high accuracy of ±1 ~ 2mm. The limit of object detection with 20 mA LED current is 1cm and with 40 mA is 2cm. The potential application is that it can be used for early breast cancer detection as a convenient self-examination device.

We have developed a frequency-domain near-infrared device suitable for physiological studies in human. In this work, a four-channel configuration of the instrument is applied to monitor hemodynamic and oxygenation changes in the frontal cortex of volunteers during different ventilation tasks. We use four different source-receiver separations (2, 3, 4, and 5 cm) and three wavelengths (760, 808, and 830 nm) to test the sensitivity of these parameters to cardiovascular and metabolic changes. Low-frequency oscillations (~ 0.02 Hz) and variations in heart rate during different ventilation tasks are investigated as well. We also study physiological changes during natural sleep using the frequency-domain instrument simultaneously with a polysomnography system containing a pulse oximeter. Our results indicate that hemodynamic and oxygenation changes in the frontal cortex during natural sleep can be detected using near-infrared measurements.