This PDF file contains the front matter associated with SPIE Proceedings volume 7557, including the Title Page, Copyright information, Table of Contents, Introduction, and the Conference Committee listing.

It is generally believed that the inverse problem in diffuse optical tomography (DOT) is highly ill-posed and its solution is always under-determined and sensitive to noise, which is the main problem in the application of DOT. In this paper, we propose a method on image reconstruction for time-domain diffuse optical tomography based on panel detection and Finite-Difference Method, and introduce an approach to reduce the number of unknown parameters in the reconstruction process. We propose a multi-level scheme to reduce the number of unknowns by parameterizing the spatial distribution of optical properties via wavelet transform and then reconstruct the coefficients of this transform. Compared with previous traditional uni-level full spatial domain algorithm, this method can efficiently improve the reconstruction quality. Numerical simulations show that wavelet-based multi-level inversion is superior to the uni-level algebraic reconstruction technique.

Image reconstruction is one of the main challenges for fluorescence tomography. For in vivo experiments on small animals, in particular, the inhomogeneous optical properties and irregular surface of the animal make free-space image reconstruction challenging because of the difficulties in accurately modeling the forward problem and the finite dynamic range of the photodetector. These two factors are fundamentally limited by the currently available forward models and photonic technologies. Nonetheless, both limitations can be significantly eased using a signal processing approach. We have recently constructed a free-space panoramic fluorescence diffuse optical tomography system to take advantage of co-registered microCT data acquired from the same animal. In this article, we present a data processing strategy that adaptively selects the optical sampling points in the raw 2-D fluorescent CCD images. Specifically, the general sampling area and sampling density are initially specified to create a set of potential sampling points sufficient to cover the region of interest. Based on 3-D anatomical information from the microCT and the fluorescent CCD images, data points are excluded from the set when they are located in an area where either the forward model is known to be problematic (e.g., large wrinkles on the skin) or where the signal is unreliable (e.g., saturated or low signal-to-noise ratio). Parallel Monte Carlo software was implemented to compute the sensitivity function for image reconstruction. Animal experiments were conducted on a mouse cadaver with an artificial fluorescent inclusion. Compared to our previous results using a finite element method, the newly developed parallel Monte Carlo software and the adaptive sampling strategy produced favorable reconstruction results.

We present first results of a fluorescence optical diffusion tomography experiment coupled to a X-ray computed tomography reconstruction. An instrument, dedicated to the co-registration of optical and X-ray measurements, has been developed: specific acquisition protocol and reconstruction software have been developed for carrying out fluorescence diffuse optical tomography in a cylindrical geometry consistent with X-ray tomography. Actual animal geometry provided by the X-ray tomography is used to give animal boundaries to the diffuse optical tomography reconstruction algorithm. Experiments have been conducted on sacrificed mice and fluorescence reconstructions have been evaluated and are geometrically consistent with X-ray ones.

A gantry-based hybrid fluorescence and x-ray computed tomography (FT/CT) system is developed for quantitative molecular imaging. The performance of the dual modality FT/CT system is evaluated using an irregular shaped phantom with an inclusion containing Indocyanine-Green (ICG). The anatomical data from CT provides structural a priori information for the FT inverse problem. Although a 4.2 mm diameter inclusion can be resolved in the reconstructed concentration image without any a priori information, ICG concentration in the inclusion is recovered with 75% error. On the other hand, the error in the recovered ICG concentration reduces to 15% when a priori information from CT is utilized. The results demonstrate that accurate fluorophore concentration can only be obtained when x-ray CT structural a priori information is available.

Our goal is to assess the feasibility of a bi-functional contrast agent that is intravenously injected to an R3230 induced small animal breast tumor model. The MR/optical contrast agent was produced by GE Global Research, NY, and it was available in one size, Dp20. We used a combined frequency domain diffuse optical tomography (DOT) and a 4T magnetic resonance (MR) scanner to simultaneously measure the kinetics of the contrast agent in vivo. Both systems detected the signal change in the tumor and the non-tumor region. MR measurements served as a gold standard to validate the optical kinetics. We present both MR and DOT dynamic curves as well as the reconstructed optical absorption map.

Optical coherence tomography (OCT) and intravascular ultrasound (IVUS) are considered two complementary imaging techniques in the detection and diagnosis of atherosclerosis. OCT permits visualization of micron-scale features of atherosclerosis plaque, and IVUS offers full imaging depth of vessel wall. Under the guidance of IVUS, minimal amount of flushing agent will be needed to obtain OCT imaging of the interested area. We report on a dual-modality optical coherence tomography (OCT) - ultrasound (US) system for intravascular imaging. To the best of our knowledge, we have developed the first integrated OCT-US probe that combines OCT optical components with an ultrasound transducer. The OCT optical components mainly consist of a single mode fiber, a gradient index (GRIN) lens for light beam focusing, and a right-angled prism for reflecting light into biological tissue. A 40MHz PZT-5H side-viewing ultrasound transducer was fabricated to obtain the ultrasound image. These components were integrated into a single probe, enabling both OCT and ultrasound imaging at the same time. In vitro OCT and ultrasound images of a rabbit aorta were obtained using this dual-modality imaging system. This study demonstrates the feasibility of an OCT-US system for intravascular imaging which is expected to have a prominent impact on early detection and characterization of atherosclerosis.

To increase prostate cancer diagnosis sensibility, we propose to add an optical modality to an US biopsy tool to localize fluorophore marked tumors. Optical signals are acquired on a time-resolved acquisition chain composed by a 770 nm femtosecond laser source and a four channels TCSPC device. The fluorescence concentration is reconstructed by using intensity and mean time of flight acquired from each time-resolved source-detector signal. Validation experiments are performed on a phantom mimicking prostate both on its optical and ultrasound properties with 10 μmol/L ICG 1 cm deep double fluorescent inclusions to simulate marked tumors. An exhaustive search algorithm succeeded in reconstructing the two distinct fluorescence dots with correct locations.

Pressure ulcers have been identified as a public health concern by the US government through the Healthy People 2010 initiative and the National Quality Forum (NQF). Currently, no tools are available to assist clinicians in erythema, i.e. the early stage pressure ulcer detection. The results from our previous research (supported by NIH grant) indicate that erythema in different skin tones can be identified using a set of wavelengths 540, 577, 650 and 970nm. This paper will report our recent work which is developing a handheld, point-of-care, clinicallyviable and affordable, real time multispectral imager to detect erythema in persons with darkly pigmented skin. Instead of using traditional filters, e.g. filter wheels, generalized Lyot filter, electrical tunable filter or the methods of dispersing light, e.g. optic-acoustic crystal, a novel custom filter mosaic has been successfully designed and fabricated using lithography and vacuum multi layer film technologies. The filter has been integrated with CMOS and CCD sensors. The filter incorporates four or more different wavelengths within the visual to nearinfrared range each having a narrow bandwidth of 30nm or less. Single wavelength area is chosen as 20.8μx 20.8μ. The filter can be deposited on regular optical glass as substrate or directly on a CMOS and CCD imaging sensor. This design permits a multi-spectral image to be acquired in a single exposure, thereby providing overwhelming convenience in multi spectral imaging acquisition.

Although vertical cavity surface emitting lasers (VCSELs) have traditionally found their place in high-speed communication links, the recent advancements of VCSELs emitting in the visible spectrum has sparked interest for new applications in scanning and imaging. Compared to other lasers, VCSELs have many advantageous characteristics including compact size, low power requirements, low cost, and high reliability. VCSELs also offer the unique ability to be fabricated in one- or two-dimensional arrays, making it possible for multiple VCSELs on a single chip to perform the same function as a mechanically scanned beam. One such application is computed radiography (CR), which provides an efficient solution for digitizing and electronically storing x-ray images. In this work, we demonstrate a 1-inch prototype CR scanner based on high-density VCSEL arrays. The device is capable of generating very fast scans with no moving parts, and has the potential to increase throughput, stability, and image quality. In this paper, we discuss the design and performance of this scanner and demonstrate X-ray image acquisition with resolution exceeding 5 line pairs per millimeter (lp/mm).

The integration of x-ray imaging with optical imaging is becoming routine at the pre-clinical level, as both projection and tomography systems are now commercially integrated as packaged systems. Yet, the differences between their capabilities are wide, and there is still perhaps a lack of appreciation about how difference pre-clinical x-ray systems are from clinical x-ray systems. In this survey, the key advantages of each approach, x-ray and optical, are described, and the potential synergies and deficiencies are discussed. In simple terms, the major benefit of optical imaging is in the spectroscopic capabilities, which allow the potential for imaging fluorescent agents in vivo, and the future potential for imaging multiple species at a time with spectral discrimination or spectral fitting of the data. In comparison, multienergy x-ray systems are being realized in clinical use, or automated discrimination of soft versus hard tissues, and the combination of optical imaging with this type of dual-energy x-ray imaging will significantly enhance the capabilities of the hybrid systems. Unfortunately, the power of dual energy imaging is not as possible at the pre-clinical stage, because of the limitations of contrast-resolution and x-ray dose. This is discussed and future human systems outlined.

Accurate assessment of wound oxygenation and perfusion is important for evaluating wound healing/regression and guiding following therapeutic processes. However, many existing techniques and clinical practices are subjective and qualitative due to background bias, tissue heterogeneity, and inter-patient variation. To overcome these limitations, we developed a dual-modal imaging system for in vivo, non-invasive, real-time quantitative assessment of wound tissue oxygenation and perfusion. The imaging system integrated a broadband light source, a high-resolution CCD camera, a highly sensitive thermal camera, and a liquid crystal tunable filter. A user-friendly interface was developed to control all the components systematically. Advanced algorithms were explored for reliable reconstruction of tissue oxygenation and appropriate co-registration between thermal images and multispectral images. Dual-mode oxygenation and perfusion imaging was demonstrated on both benchtop models and human subjects, and compared with measurements using other methods, such as Laser Doppler and tissue oximeter. The test results suggested that the dual-modal imaging system has the potential for non-contact real-time imaging of wound tissue oxygenation and perfusion.

Near-infrared spectroscopy (NIRS) is based on the modified-Lambert-Beer's law that changes in absorbance are proportional to changes in hemoglobin parameters. Majority of the conventional measurement methods uses only two or three wavelengths. In this research, basic examination of NIRS measurement was approached by acquiring wide range of wavelength information. Arterial occlusion task was performed by using the blood pressure cuff around the upper arm. Pressure of 200mmHg was then applied for about 3 minutes. During the arterial occlusion, the spectrum of the lower arm muscles was measured every 15 seconds, within the range of 600 to 1100nm. The secondary derivative spectrum was calculated from the measured spectrum. Arterial occlusion is a task which changes the oxygenation level of the tissue. The change can be regarded as the change of the spectrum form, not as the change of the baseline. Furthermore, it was found that other wavelength bands hold information correlating to this arterial occlusion task.

The possibilities to perform multi-band spectral imaging by means of a consumer color camera without external filters have been studied. Images at up to 6 spectral bands may be extracted from a single color image after appropriate signal processing. The proposed technique was tested in pilot measurements of in-vivo skin hemoglobin maps and laser-excited autofluorescence images.

Fluorescence lifetime tomography (FLT) is an emerging imaging modality that seeks for recovering distributions of the fluorescent yield and lifetime inside in vivo tissues. This technique, mainly based on time-domain instrumentation, has found promising applications in small-animal imaging for studying tumor pathology and for drug development. As one of the model-based imaging methods, FLT can be finalized with inverting an underdetermined, ill-posed linear system with regard to both the parameters, for which several methods has been adopted. This paper concisely revises the main facts of three commonly-used linear inversions: algebraic reconstruction technique, truncated singular value decomposition and conjugate gradient descent, and presents a comparative investigation on these methods in terms of the image quality and noise robustness.

Non-contact scheme is prevalent to diffuse fluorescence tomography (FDT) since it facilitates instrumentation as well as simplifies experimental procedure. Although non-contact FDT generally uses CCD camera as detectors to achieve high throughput of data collection, a fiber-based implementation can make full use of well-established high-sensitive and time-resolved detection techniques. Therefore, a system that combines the fiber-based time-resolved detection and the non-contact geometry of optodes would be significantly attractive, which also means a more complex modeling of photon migration. This paper presents detailed computational aspects of the fiber-based non-contact DFT, including both the forward and inverse models. A pilot validation of the method is performed using simulated data for a two-dimensional case.

Time-domain optical mammography has attracted many attentions since it can diagnose early breast cancer by efficiently reconstructing optical parameter. However, the currently available image reconstruction algorithms for time-domain optical mammography are badly influenced by different Jacobian magnitudes of absorption coefficient and reduced scattering coefficient. To improve image quality, we proposed an efficient Jacobian scaling method with a relative data type based on generalized pulse spectrum technique. Our simulated and experimental reconstructions show that this Jacobian scaling method can efficiently enhancing the quality of reconstructed image.

Time-domain fluorescence diffuse optical tomography (FDOT) can provide information, not only concerning the localization of specific fluorophores, but also about the local fluorophore environment. We present a method based on linear inversion algorithm to reconstruct images of fluorescence yield and lifetime from time-resolved data. To provide efficient solutions, we convert the data type by Laplace transform and adapt normalized Born ratio for its advantages in fluorescence mode. The methodology is experimentally validated in reflection and transmittance measurements by use of time-correlation single photon counting system. We experimentally validate that the proposed scheme can achieve simultaneous three-dimensional reconstruction of the fluorescent yield and lifetime. The results show that for the positions, sizes and shapes of the targets, there are some deviation in reflection measurement, the quality in transmittance one is more satisfied.

In this paper we apply the shape-based approach to diffuse optical tomography (DOT) reconstruction, which aims to simultaneously recover the smooth boundaries of the tissue regions and the constant coefficients within them. An advantage of shape-based solutions is the reduction of the unknown parameters, which is especially important for nonlinear ill-posed inverse problems. We introduce a Fourier series representation of the closed region boundaries and a boundary element method (BEM) for the forward model. For inverse problem the Levenberg-Marquardt optimization process is implemented here. The performance of the proposed method is evaluated by simulations at different noise levels and phantom experiment which is embedded a single cylinder target. We can get reasonable reconstruction from both Gauss noise and real noise in the experimental study. The results illuminate that the methodology is very promising and of global convergence, the boundaries and the optical coefficients can both be recovered with good accuracy from the noisy measurements.

Diffuse optical tomography (DOT) has been increasingly studied in the past decades. In DOT, the radiative transfer equation (RTE) and its P1 approximation, i.e. the diffuse equation (DE), have been used as the forward models. Since the DE-based DOT fails where biological tissue has a void-like region and when the source-detector separation is less than 5 mean free pathlengths, as in the situations of small animal imaging, the RTE-based DOT methodology has become a focus of investigation. Therefore, the complete formalism of the RTE is attracting more and more interest. It is clear that the quality of the reconstructed image depends strongly on the accuracy of the forward model. In this paper, A FDM was developed for solving two-dimensional RTE in a 2cm×2cm square homogeneous tissue with two groups of the optical properties and different schemes of the spatial and solid angle discretization. The results of the FDM are compared with the MC simulations. It is shown that when the step size of the spatial mesh becomes small, more discretized angle number is needed.

Thermal imaging is an emerging method for early detection of female breast tumor. The main challenge for thermal imaging used in breast clinics lies in how to detect or locate the tumor and obtain its related parameters. The purpose of this study is to apply an improved method which combined a genetic algorithm with finite element thermal analysis to determine the breast tumor and its parameters, such as the size, location, metabolic heat generation and blood perfusion rate. A finite element model for breast embedded a tumor was used to investigate the temperature distribution, and then the influences of tumor metabolic heat generation, tumor location and tumor size on the temperature were studied by use of an improved genetic algorithm. The results show that thermal imaging is a potential and effective detection tool for early breast tumor, and thermal simulation may be helpful for the explanation of breast thermograms.

It is well known that the reconstruction problem in fluorescence diffuse optical tomography is badly conditioned and requires the knowledge of medium optical properties. Its principle is to measure the fluorescence light emerging at different positions of the surface when the biological medium is excited with point sources. In this paper, we evaluate the influence of the medium optical properties and the noise on the fluorescence reconstruction, and we introduce a new regularized fluorescence reconstruction method using an a priori on contours. The fluorescence reconstruction improvement is studied when using this method.

We report a combined swept source optical coherence tomography (SSOCT) and fluorescence spectroscopy (FS) system for multimodal measurements of biochemical information and internal structures with fast imaging speed. The FS probe composed of a double clad fiber (DCF) coupler is adopted into the single mode fiber based SSOCT system. Here, the DCF coupler acts as a FS system as well as a multifunctional probe for the SSOCT-FS system. The performances of the single-unit SSOCT-FS system is confirmed by monitoring the fluorescence signal from photosensitizer in cancerous region of in-vivo rat and by imaging the internal structures of same region with speed of 40 frames/sec.

We provide theoretical validation of the brain-functional detection using multidistance probe arrangement based on Monte Carlo simulations of five-layered model in which both scattering and absorption changes occur. It shows that optimized multidistance probe arrangement can be effective in removing interferences by scattering and absorption changes in upper layers and extracting absorption change in the gray matter layer. Using newly designed probes and their holder system, both conventional and proposed fNIRS measurements were implemented with non-functional (body and head movements and respiratory change) and functional (finger opposition) tasks. Artifacts, even if it correlate with task sequence, were well reduced. Functional signals were well localized at lateralized cerebral functional area.

Blood samples are frequently analyzed for the blood disorders or other diseases in the research and clinic applications. Most of the analyses are related to blood cell counts and blood cell sizes. In this paper, the line scan CCD imaging system is developed, which is based on the Texas Instruments' TMS320C6416T (DSP6416), a high performance digital signal processor and Altera's Field programmable Gate Array (FPGA) EP3C25F324. It is used to acquire and process the images of blood cells for counting the number of cells, sizing and positioning them. The cell image is captured by line scan CCD sensor and then the digital image data converted by Analogue Front-End (AFE) are transferred into FPGA, after pre-processing they are transferred into DSP6416 through the interface of First In First Out (FIFO) in FPGA and External Memory Interfaces (EMIF) of DSP6416. Then the image data are processed in DSP6416. Experimental results show that this system is useful and efficient.

The lack of clear consensus over the utility of multispectral imaging (MSI) for bright-field imaging prompted our team to investigate the benefit of using MSI on breast tissue microarrays (TMA). We have conducted performance studies to compare MSI with standard bright-field imaging in hematoxylin stained breast tissue. The methodology has three components. The first extracts a region of interest using adaptive thresholding and morphological processing. The second performs texture feature extraction from a local binary pattern within each spectral channel and compared to features of co-occurrence matrix and texture feature coding in third component. The third component performs feature selection and classification. For each spectrum, exhaustive feature selection was used to search for the combination of features that yields the best classification accuracy. AdaBoost with a linear perceptron least-square classifier was applied. The spectra carrying the greatest discriminatory power were automatically chosen and a majority vote was used to make the final classification. 92 breast TMA discs were included in the study. Sensitivity of 0.96 and specificity of 0.89 were achieved on the multispectral data, compared with sensitivity of 0.83 and specificity of 0.85 on RGB data. MSI consistently achieved better classification results than those obtained using standard RGB images. While the benefits of MSI for unmixing multi-stained specimens are well documented, this study demonstrated statistically significant improvements in the automated analysis of single stained bright-field images.

Breast cancer is a significant cause of mortality and morbidity among women with early diagnosis being vital to successful treatment. Diffuse Optical Tomography (DOT) is an emerging medical imaging modality that provides information that is complementary to current screening modalities such as MRI and mammography, and may improve the specificity in determining cancer malignancy. Using high-resolution anatomic images as a priori information improves the accuracy of DOT. Measurements are presented characterizing the performance of our system. Preliminary data is also shown illustrating the use of a priori MRI data in phantom studies.ä

Multispectral bioluminescence tomography (BLT) is one of the seemingly promising approaches to recover 3D tomographic images of bioluminescence source distribution in vivo. In bioluminescence tomography, internal light source, such as luciferase is activated within a volume and multiple wavelength emission data from the internal bioluminescence sources is acquired for reconstruction. The underline non-uniqueness problem associated with non-spectrally resolved intensity-based bioluminescence tomography was demonstrated by Dehghani et al. and it also shown that using a spectrally resolved technique, an accurate solution for the source distribution can be calculated from the measured data if both functional and anatomical a priori information are at hand. Thus it is of great desire to develop an imaging system that is capable of simultaneously acquiring both the optical and structural a priori information as well as acquiring the bioluminescence data. In this paper we present our first combined optical tomography and CT system which constitutes with a cool CCD camera ( perkin elmer "cold blue"), laser launching units and Xray CT( Dxray proto-type). It is capable of acquiring non contact diffuse optical tomography (DOT) data which is used for functional a priori; X-ray CT images which yields the structure information; and BLT images. Physical phantom experiments are designed to verify the system accuracy, repeatability and resolution. These studies shows the feasibility of such imaging system and its potential.

Angular domain imaging (ADI) generates a projection image of an attenuating target within a turbid medium by employing a silicon micro-tunnel array to reject photons that have deviated from the initial propagation direction. In this imaging method, image contrast and resolution are position dependent. The objective of this work was to first characterize the contrast and resolution of the ADI system at a multitude of locations within the imaging plane. The second objective was to compare the reconstructions of different targets using filtered back projection and iterative reconstruction algorithms. The ADI system consisted of a diode laser laser (808nm, CW, ThorLabs) with a beam expander for illumination of the sample cuvette. At the opposite side of the cuvette, an Angular Filter Array (AFA) of 80 μm x 80 μm square-shaped tunnels 1 cm in length was used to reject the transmitted scattered light. Image-forming light exiting the AFA was detected by a linear CCD (16-bit, Mightex). Our approach was to translate two point attenuators (0.5 mm graphite rod, 0.368 mm drill bit) submerged in a 0.6% Intralipid^TM dilution using a SCARA robot (Epson E2S351S) to cover a 37x37 and 45x45 matrix of grid points in the imaging plane within the 1 cm path length sample cuvette. At each grid point, a one-dimensional point-spread distribution was collected and system contrast and resolution were measured. Then, the robot was used to rotate the target to collect projection images at several projection angles of various objects, and reconstructed with a filtered back projection and an iterative reconstruction algorithm.

Diffuse Optical Tomography (DOT) is a new and promising medical imaging modality which uses near-infrared light to probe tissue properties. Using multiple wavelengths of light can provide important information about tissue metabolism and cancer malignancy. Unfortunately, in most DOT acquisition schemes, acquiring data for each wavelength has a multiplicative effect on the overall imaging time. In this paper, we evaluate a new multiple wavelength laser module (Praevium Research Inc.) with 12 laser diodes all coupled to a single output fiber. When used in conjunction with a cooled spectrometer, it allows simultaneous multi-wavelength data acquisition and hence, higher temporal resolution.

A crucial parameter in Diffuse Optical Tomography (DOT) is the construction of an accurate forward model, which greatly depends on tissue boundary. Since photon propagation is a three-dimensional volumetric problem, extraction and subsequent modeling of three-dimensional boundaries is essential. Original experimental demonstration of the feasibility of DOT to reconstruct absorbers, scatterers and fluorochromes used phantoms or tissues confined appropriately to conform to easily modeled geometries such as a slab or a cylinder. In later years several methods have been developed to model photon propagation through diffuse media with complex boundaries using numerical solutions of the diffusion or transport equation (finite elements or differences) or more recently analytical methods based on the tangent-plane method . While optical examinations performed simultaneously with anatomical imaging modalities such as MRI provide well-defined boundaries, very limited progress has been done so far in extracting full-field (360 degree) boundaries for in-vivo three-dimensional DOT stand-alone imaging. In this paper, we present a desktop multi-spectrum in-vivo 3D DOT system for small animal imaging. This system is augmented with Technest's full-field 3D cameras. The built system has the capability of acquiring 3D object surface profiles in real time and registering 3D boundary with diffuse tomography. Extensive experiments are performed on phantoms and small animals by our collaborators at the Center for Molecular Imaging Research (CMIR) at Massachusetts General Hospital (MGH) and Harvard Medical School. Data has shown successful reconstructed DOT data with improved accuracy.