The Section on Biomedical Stochastic Physics (SBSP) works on light-tissue interactions related to non-invasive optical imaging of biological targets. The objectives of the SBSP are to devise quantitative theories, develop methodologies, and design instrumentation to study biological phenomena whose properties are characterized by elements of randomness in both time and space. The main focus of the section has been to develop quantitative theories applicable to in vivo quantitative optical spectroscopy and tomographic imaging of tissues. This requires analysis of different optical sources of contrast such as endogenous or exogenous fluorescent labels, absorption (e.g., hemoglobin or chromophore concentration), and/or scattering. In many instances, we design and conduct experiments and computer simulations to validate theoretical findings. In addition, we conduct collaborative research both nationally, internationally, and within the Intramural Research Program of the NIH to investigate physiological sites where optical techniques might be clinically practical and offer new diagnostic knowledge and/or less morbidity than existing diagnostic methods. In addition, we have started a project to study tumor-induced angiogenesis. Angiogenesis plays an essential role in the establishment of tumor malignancy. To gain a better understanding of the mechanisms underlying angiogenesis, we have developed a stochastic model for qualitative analysis of the biological events that constitute angiogenesis.
Members of SBSP start at the desk to develop methodologies applicable to in vivo quantitative tissue spectroscopy and tomographic imaging. At the bench, we are designing and conducting experiments on tissue-like phantoms and running computer simulations to validate our findings. With successes at the bench, we are bringing our imaging and spectroscopic devices to the bedside.
SBSP is involved in several clinical studies, including three NCI-sponsored clinical trials to evaluate the effectiveness of anti-angiogenesis drug treatment for Kaposi's sarcoma, and an NIDCR/NINDS clinical study to evaluate the drug response of patients experiencing complex regional pain syndrome. SBSP has also started to model the stochastic process of tumor-induced angiogenesis in the extracellular matrix.
For more information on the NICHD Satellite Mouse Imaging Area visit the Gandjbakhche Lab site.
For more information on NIH Optical Imaging Workshop visit the Gandjbakhche Lab site.
Scientists in the Section on Biomedical Stochastic Physics are currently involved in several projects that are in various stages from the bench to the bedside. A list of some projects is listed below.
Accurate, simple, and reliable techniques are needed to perform sensitive and specific in vivo determinations of abnormalities within a given tissue. Successful, noninvasive "optical biopsies" might replace invasive, destructive biopsies, providing advantages of smaller sampling errors, reduction in cost and time for diagnosis, and resulting in easier integration of diagnosis and therapy by following progression of disease or its regression in response to therapy.
Clinically practical fluorescence imaging techniques must meet several requirements. First, the pathology under investigation must not lie at such a depth that the attenuation of the signal gives poor signal-to-noise ratio and resolvability. Second, the specificity of the marker must be high enough so that one can clearly distinguish between normal states and abnormal lesions. Finally, one must have a robust image reconstruction algorithm, which enables one to quantify the fluorophore concentration at a given depth. We have chosen Sjøgren's Syndrome (SS), as an appropriate test case for developing this noninvasive optical biopsy. SS is an autoimmune disease affecting minor salivary glands that are near (0.5 to 3.0 mm below) the oral mucosal surface. We are attempting to use exogenous fluorescent ligands (e.g., antibodies to CD4+ T cell-activated lymphocytes infiltrating the salivary glands) to provide high optical contrast in a quantitative relationship that will allow assessment of the stage of the disease process. We have developed a theoretical description of diffuse fluorescence photon migration for such a system, and have tested it with data acquired from a simulation experiment. Small (1mm3) rhodamine targets, which mimic diseased minor salivary glands labeled with fluorescent antibodies to infiltrating lymphocytes in SS, were embedded in a highly scattering tissue-like phantom at various depths. Excellent agreement between the experimentally measured surface profiles and those predicted by our theory was obtained.
We have also been investigating whether fluorescent lifetime imaging may be used to obtain functional information regarding local concentrations of specific substances such as O2, or information about environmental conditions such as temperature and pH. Quantification requires a fluorophore with a known dependence on the environmental factor. For deeply embedded sites in a turbid medium such as tissue, measuring the photon arrival delay caused by a specific fluorophore lifetime is made difficult by the photon arrival delays caused by multiple scattering of the photon on its transit through the tissue. However, RWT is well suited to solving this type of problem. The delay of photons that results from the excitation and later emission by a fluorophore can be modeled in RWT in the same way as transit delays due to multiple scattering. Hence, we have derived a RWT closed-form solution for time-resolved fluorescent lifetime imaging. This is now being tested with fluorophore-impregnated agarose phantoms. We have entered into two collaborative projects that will make use of the results of these studies. In collaboration with the Radiation Biology Program of the NCI, we have begun a mouse imaging study that will utilize malignant cells transfected with the gene(s) for synthesis of fluorescent proteins. In a second project we are investigating the use of IR-dependent fluorescent detection methods as alternatives to radionucleotide detection to determine the locations of sentinel lymph nodes for surgical treatment of cancer.
Angiogenesis, the transient formation of new blood vessels, is mainly observed under certain physiological conditions in the adult. Although recent in vitro models have been crucial in determining the effects of growth factors and inhibitors on cell migration and sprouting, the fundamental processes of cell adhesion kinetics and the interaction between endothelial cells (ECs) and extracellular matrix (ECM) during tumor induced angiogenesis have not been quantified. During early stages of angiogenesis, ECs migrate from pre-existing vessels through the ECM, adhere to each other, and form a specific pattern. Following this initial event, many ECs form new, small finger-like capillaries. The sprouts grow in length with the migration and further recruitment of endothelial cells, and move toward the tumor.
To study the network formation of endothelial cells (ECs) in an extracellular matrix (ECM) environment, we have devised an EC aggregation-type model based on a diffusion-limited-cluster-aggregation model (DLCA), where clusters of particles diffuse and stick together upon contact. We use this model to quantify EC differentiation into cord-like-structures by comparing experimental and simulation data. Approximations made with the DLCA model, when combined with experimental kinetics and cell concentration results, not only allow us to quantify cell differentiation by a pseudo diffusion coefficient, but also measure the effects of tumor angiogenic factors (TAF) on the formation of cord-like structures by ECs.
We have tested our model by using an in vitro assay, where we record EC aggregation by analyzing time-lapse images that provide us with the evolution of the fractal dimension measure through time. We performed these experiments for various cell concentrations and TAF (e.g. EVG, FGF-b, and VEGF). During the first six hours of an experiment, ECs aggregate quickly. The value of the measured fractal dimension decreases with time until reaching an asymptotic value that depends solely on the EC concentration. In contrast, the kinetics depend on the nature of TAF. The experimental and simulation results correlate with each other in regards to the fractal dimension and kinetics, allowing us to quantify the influence of each TAF by a pseudo diffusion coefficient.
We have shown that the shape, kinetic aggregation, and fractal dimension of the EC aggregates fit into an in vitro model capable of reproducing the first stage of angiogenesis. We conclude that the DLCA model, combined with experimental results, is a highly effective assay for the quantification of the kinetics and network characteristics of ECs embedded in ECM proteins. Finally, we present a new method that can be used for studying the effect of angiogenic drugs in in vitro assays.
Another project is the use of stochastic modeling to analyze the influence of the ECM's heterogeneity on tumor induced angiogenesis. Cell migration during angiogenesis guides the formation of new blood vessels. This migration is mainly directed by a chemotactic flux via a gradient of growth factors between the tumor and the cell. New studies have shown that the ECM substrate can influence cell migration and apoptosis. We have performed experiments showing the correlation between endothelial cell migration and the fiber distribution in the ECM. The endothelial cells follow the pathways created by the fibers in the ECM. We have seen through stochastic modeling that the heterogeneity of the fibers in the ECM create obstacles for the blood vessels to perfuse the tumor.
We are currently using fluorescence correlation spectroscopy (FCS) to measure locally the degradation of the ECM due to the ECs. During angiogenesis, when ECs migrate through the tumor to form new blood vessels, the ECs degrade the ECM by cell locomotion and the release of matrix metalloproteinase (MMP). FCS measures tiny variations in gel structures by measuring the diffusion coefficient. Some preliminary experiments have shown promising results. We are going to do a cordlike structure assay to measure how the cell migration and aggregation change locally the structure of the gel. Preliminary results show a variation in the gel structure during the cell aggregation. In the future we intend to study how MMP and anti-MMP could affect the degradation of the gel and quantify the cause of this degradation (cell locomotion and MMP).
Kaposi's sarcoma (KS) is a common AIDS-related vascular tumor. Angiogenesis can play an important role in the development and progression of KS. No non-invasive standard technique is available for qualitatively or quantitatively evaluating the effect of anti-angiogenesis based therapy on blood flow in KS tissue. In this study, we are investigating the usefulness of several non-invasive techniques to assess the response of anti-angiogenetic drugs for treatment of KS patients. These techniques include infrared imaging, laser Doppler imaging (LDI) and near-infrared multi-spectral imaging.Thermography and LDI images of a representative KS lesion were recorded in 16 patients and compared to normal skin either adjacent to the lesion or on the contralateral side. Eleven of the 16 patients had greater than 0.5 oC increased temperature and 12 of 16 patients had increased flux (measured by LDI) as compared to normal skin. There was a strong correlation between these two parameters (R = 0.81, p < 0.001). In ten patients, measurements were obtained prior to therapy and after receiving a regimen of liposomal doxorubicin and interleukin-12. After 18 weeks of therapy, temperature and blood flow of the lesions were significantly reduced from the baseline (p = 0.004 and 0.002 respectively). These techniques hold promise to assess physiologic parameters in KS lesions and their changes with therapy.
Figure 1. Typical example of lesion obtained from a subject with KS (A) before, and (B) after the treatment. The lesion becomes normal as assessed by the thermal or LDI images 18 weeks.
Near-infrared spectroscopy is a non-contact and noninvasive method of monitoring changes in concentrations of blood volume and oxygenated- and deoxygenated-hemoglobin. Assessing these analytes is complicated by other pigments in the skin, i.e. melanin and hemosiderin. However, it is possible to correct for such pigments, and NIR spectroscopy has the potential to aid in assessing the pathogenesis of the status and changes of KS lesions during therapy.
Near-infrared spectroscopy is most closely related to visual assessment. With S. Demos at the Lawrence Livermore National Laboratory, a spectral imaging system was designed that captures images with a high-resolution CCD portable camera at six near-infrared wavelengths (700, 750, 800, 850, 900, and 1000 nm) was designed. A white light held approximately 15 cm from tissue illuminates the tissue uniformly. Using optical filters, images are obtained at the six wavelengths and the intensity images are used in a mathematical optical model of skin with epidermis and much thicker, highly scattering dermis. Each layer contains major chromophores that determine absorption in the corresponding layer and the layers together determine the total reflectance of the skin. Local variations in melanin, oxygenated hemoglobin (HbO2) and blood volume can be reconstructed through a multivariate analysis. For the mathematical optical skin model, the effect of the thin epidermis layer on the intensity of the diffusely reflected light was determined by the effective attenuation of light. The epidermis absorption coefficient was determined by the percentage of melanin, the absorption coefficient of melanin and the absorption coefficient of normal tissue. The influence of the much thicker, highly scattering dermis layer on the skin reflectance should be estimated by a stochastic model of photon migration, e.g., random walk theory. Fitting the known random walk expression for diffuse reflectivity of the turbid slab yields a formula that depends on the reduced scattering coefficient and dermis absorption coefficient. The dermis absorption coefficient is based on the volume of blood in the tissue and hemoglobin oxygenation, i.e. relative fractions of HbO2 and deoxygenated hemoglobin (Hb). At wavelengths greater than 850 nm, the contribution of water and lipids should be taken into account. The absorption coefficient of blood was calculated by the volume fraction of HbO2 times the absorption coefficient of HbO2 plus the volume fraction of Hb times the absorption coefficient of Hb. In the dermis, large cylindrical collagen fibers are responsible for Mie scattering, while smaller scale collagen fibers and other micro-structures are responsible for Rayleigh scattering. A best-fit procedure was used to reconstruct for the melanin volume, HbO2 fraction, and blood volume fraction.
Figure 2. Example of multi-modality images obtained from a KS patient before cytotoxic/antio-angioigenesis treatment: (a) digital, (b) reconstructed oxygenated hemoglobin, (c) reconstructed tissue blood volume, (d) thermal, and (e) laser Doppler.
Figure 3. A thermal image from our Kaposi's sarcoma research appears on the cover of a new book entitled Medical Infrared Imaging.
Reflex Sympathetic Dystrophy (RSD), currently known as Complex Regional Pain Syndrome type I (CRPS-I) has long been recognized clinically. However, RSD is very difficult to diagnosis and track. In this study, the nature of chronic neuropathic pain is assessed using three new techniques: (a) asymmetric thermal pattern between pain and contralateral pain free sites; (b) comparisons of blood flow measured by laser Doppler imaging (LDI) between the two sides; and (c) sympathetic response of blood flow and temperature patterns in the two sides following cold stimulation.The main purpose of this study is to investigate the applicability of thermography and LDI for tracking the physiological parameters associated with chronic pain before, during and up to five weeks after therapeutic treatment. Asymmetric thermal patterns and sympathetic responses on the two sides after applying cold stimulations are observed. On the other hand, no remarkable difference is observed between the two sides in LDI blood velocity images. The thermal pattern on the most painful side before drug or placebo treatment is typically warmer by at least 1oC spread over a large area. After treatment, thermography shows cooler temperatures on the painful side than the contralateral side.
Typical example of thermal patterns of arms, legs and feet of a normal volunteer.
Figure 5. Typical example of thermal patterns of feet and legs of an RSD patient.
Radiation-induced damage to skin and soft tissue continues to be a persistent problem despite advances in physics (conformal treatment and intensity modulation). The consequence of radiation-induced damage is that the individual suffers varying degrees of discomfort and dysfunction as a function of where there is damage and the degree of damage.At this time there still is no established means of assessing the degree of radiation-induced fibrosis. Range of motion studies, ultrasound, magnetic resonance imaging all contribute to assessment, but neither any individual diagnostic modality nor the combination of modalities provides insight that would permit scientific evaluation of the degree of radiation or impact on treatment. In addition, because irradiated skin and soft tissues are prone to repair poorly if injured, biopsies for evaluation are contraindicated. Hence, there is a definite need to develop non-invasive modalities to pursue diagnostic and treatment assessment. In this protocol non-invasive optical imaging will be tested as a means of providing a scientifically reproducible means of assessing radiation-induced fibrosis. Thermal imaging, laser Doppler imaging and near-infrared multi-spectral imaging are used in this study. It is the explicit intent of this work to be translated to use in human being so that an individual can be diagnosed and treatment intervention can be assessed.
Anisotropy of mouse and human skin is investigated in vivo using polarized videoreflectometry. An incident beam (linearly polarized, wavelength 650 nm) is focused at the sample surface. Two types of tissuelike media are used as controls to verify the technique: isotropic delrin and highly anisotropic demineralized bone with a priori knowledge of preferential orientation of collagen fibers. Equi-intensity profiles of light, backscattered from the sample, are fitted with ellipses that appear to follow the orientation of the collagen fibers. The ratio of the ellipe semiaxes is well correlated with the ratio of reduced scattering coefficients obtained from radial intensity distributions. Variation of equi-intensity profiles with distance from the incident beam is analyzed for different initial polarization states of the light and the relative orientation of polarization filters for incident and backscattered light. For the anisotropic media (demineralized bone and human and mouse skin), a qualitative difference between intensity distributions for cross- and co-polarized orientations of the polarization analyzer is observed up to a distance of 1.5 to 2.5 mm from the entry point. The polarized videoreflectometry of the skin may be a useful tool to assess skin fibrosis resulting from radiation treatment.
Figure 6. Depolarization of initially linearly polarized light with distance from the entry point (light is diffusively reflected by mouse skin).
Figure 7. The dependence ln[r2/(r)] for (a) demineralized bone and (b) human forearm skin, measured along the main axes of the external ellipse. Mean slopes of linear fits are shown for both vertical and horizontal axes.
Figure 8. Equi-intensity profiles of the pencillike probe beam diffusively reflected from the sample of demineralized bone and their elliptical fits. The collagen fibers of the sample were assembled vertically for (a) Hv and (b) Hh configurations of analyzer-polarizer, respectively.
The skin of athymic nude mice was irradiated with a single dose of X-ray irradiation that initiated fibrosis. Digital photographs of the irradiated mice were taken by illuminating the mouse skin with linearly polarized probe light of 650 nm. The specific pattern of the surface distribution of the degree of polarization allowed for the detection of initial skin fibrosis structures that were not visually apparent. Data processing of the raw spatial distributions of the degree of polarization based on Fourier filtering of the high frequency noise improved subjective perception of the revealed structure in the images. In addition, Pearson correlation analysis provided information about skin structural size and directionality.
Figure 9. A typical image of untreated area of a mouse.
Figure 10. The Fourier transformed image of degree of polarization data.
Spectral imaging techniques, using visible and near-infrared radiation, are being investigated for use in detecting abnormal regions deeply embedded within normal tissue. These involve time-resolved imaging techniques that enhance spatial resolution by detecting the time-of-flight of photons within the tissue. To evaluate the performance of such time-resolved transillumination techniques, we used random walk theory (RWT). Our evaluation showed that the strong scattering properties of tissues prevent direct imaging of abnormalities. We thus derived theoretical constructs to separate the effects of scattering from absorption, thus allowing us to map the optical coefficients as spectroscopic signatures of an abnormal target. By utilizing our method for different wavelengths, one can obtain diagnostic information (for example, estimates of blood oxygenation of the tumor) from corresponding absorption coefficients that no other imaging modality can directly provide. We have verified our methodology by applying it to two transillumination and fan geometry data sets provided by collaborators at University College, London. The data were obtained from an experimental phantom that had the optical properties, thickness, and characteristics of the embedded, relatively small, abnormal target that were similar to those measured in human breast. To improve performance of our algorithm, we have extended the analysis to non-localized abnormalities. We successfully applied the RWT approach to quantification of the optical characteristics and dimensions of larger inclusions (increased scattering and/or absorption) that were realized in the tissue-like phantoms of our collaborators at the Politecnico di Milano. We also have begun an analysis of real clinical data, involving tumors embedded inside normal tissues. This work is a joint project with researchers at Physikalisch-Technische Bundesanstalt of Berlin (PTB), who have designed a clinically practical optical imaging system capable to implementing time-resolved in vivo measurements on human breast. Our first goal is to quantify the optical parameters at several wavelengths and thereby estimate blood oxygen saturation of the tumor and surrounding tissue. After data processing that includes filtering and deconvolution of the raw time-resolved data, we create linear contrast scans passing through the tumor center and analyze these scans using our algorithms. Preliminary results of data obtained from a patient with invasive ductal carcinoma suggest that the tumor tissue is in a slightly deoxygenated state, with higher blood volume than surrounding tissue. These results have encouraged us to extend our collaboration with the Berlin group and to start a study with the Oncology Radiology Program at the NCI, using a new imaging device that will be installed at the NIH campus.
Chronic oral epithelial tissue inflammation is considered a potential diagnostic precursor to oral cancer. A key step in the inflammation response of epithelial tissue is the generation of cyclooxygenase (Cox), which has been associated with immunosuppression and carcinogenesis. New chemoprevention drugs that inhibit Cox-2 production and lead to decreases in PGE2 levels are now available, but evaluating the effectiveness of these chemoprevention drugs requires quantitative measures of epithelial inflammation. These are traditionally obtained from the histology of punch biopsies, which cause patient discomfort and morbidity, thereby limiting the options for repeat measurements to monitor patient progress. In principle, by using the visible light spectrum, information on numerous aspects of a tissue milieu and its biochemistry can be ascertained by non-invasive means. However, for such subsurface spectroscopy, special techniques are needed to remove the effects of scattering, which otherwise would confound the spectral analysis of relevant analytes.
We have designed an optical reflectance spectroscopy (ORS) device for measuring the thickness of the epithelial layer, and have developed an assessment technique based on oblique angle reflectance spectroscopy that allows us to assess the scattering and absorption properties of the epithelium and stroma. Our ORS device is being tested in a randomized, double blind, placebo controlled, phase IIB trial of Ketorolac mouth rinse on oropharyngeal leukoplakia, currently being conducted by NCI, NIDCD, and NIDCR. In the trial, the reduction of inflammation in the epithelial layer over a three-month period is measured using invasive punch biopsies and immunohistochemistry, combined with non-invasive OCT, and ORS. We have found encouraging qualitative agreements between histological and optical data collected from both normal and patient subjects. Based on the information obtained thus far, we are designing an optical probe with finer resolution and more channels of acquired data (eight vs. the current four). We plan to collect data from several more patients with this device, and will make repeat measurements on the same patients to allow us to look for indications of the effectiveness of treatment.
Figure 11. Oblique Angle Spectroscopy (Design and Patented by SBSP): Monitoring the Effectiveness of Chemopreventative (Cox 2 inhibitors) Drugs in Leukoplakia Patients
The algorithm developed in the Section of Biomedical Stochastic Physics enhances diffraction-limited images based on pixel-to-pixel correlations introduced by the finite width of the Point Spread Function (PSF). We simulate diffraction-limited images of point sources by convolving the PSF of a diffraction-limited lens with simulated images, and enhance the blurred images with our algorithm. Our algorithm reduces the PSF width, increases the contrast, and reveals structure on a length scale half of that resolvable in the unenhanced image. Our enhanced images compare favorably with images enhanced by conventional Tikhonov regularization.Click here to download a zip file that includes Matlab functions to run the image enhancement algorithm, a manual that describes the algorithm and the Optics Express paper that describes the work.
All related news