SCB | SMB | STBB | UBSP | Main Page
APPLICATIONS OF PHOTON MIGRATION TO TISSUE
TOMOGRAPHY AND SPECTROSCOPY
|
Amir H. Gandjbakhche, PhD, Head, Unit on Biomedical Stochastic Physics Victor Chernomordik, PhD, Senior Research Fellow |
|
| Our goal is to devise new approaches for noninvasive, quantitative optical spectroscopic and tomographic imaging of deep tissue structures for clinical screening and monitoring of physiological parameters. To achieve this goal, we have undertaken a multifaceted theoretical, computational, experimental, and clinical research program that includes time-resolved transillumination of thick tissue applied to quantitative spectroscopy of breast tumors; studies of the use of specific fluorescent markers (e.g., ligands) for identifying the molecular biology of disease processes applied to noninvasive biopsy of Sjøgren's syndrome; and lymphatic imaging for sentinel node detection. To bring these methodologies from bench to bedside, we are using rodents as test cases for our quantitative imaging concepts. We are involved in several clinical studies, including an NCI protocol to use oblique angle reflectometry for noninvasive monitoring of inflammation in the oral cavity, an NIDCR/NINDS clinical study to evaluate the drug response of patients experiencing complex regional pain syndrome, and another NCI-sponsored clinical trial to evaluate the effectiveness of anti-angiogenesis drug treatment for Kaposi's sarcoma. We have started to model the stochastic process of tumor-induced angiogenesis in the extracellular matrix. Time-resolved tomography of thick tissue Chernomordik, Hattery, Gandjbakhche; in collaboration with Cubbedu, Rinneberg, Weiss, Zaccanti We continued our joint project with the Physikalisch-Technische Bundesanstalt of Berlin and Politecnico di Milano aimed at quantifying optical properties of breast abnormalities (e.g., tumors) by using time-domain scanning mammographs designed by the Berlin and Milan groups. Our analysis of their multispectral in vivo data, obtained at several projections, is based on the random-walk model of time-resolved contrast functions. It results in estimates of the blood volume and oxygen saturation of the tumors and surrounding tissues. In most cases, we were also able to reconstruct sizes of the abnormalities. Preliminary findings on the small sample of patients indicate that the tumors (invasive ductal carcinoma) are hypoxic and have increased blood volume as compared with surrounding normal tissue. We have started a joint project with the Canadian company Advanced Research Technology to install its time-resolved imaging system at NIH. We will use the analysis of the in vivo data obtained with the device to characterize tumors and study the correlation between the spectroscopic signatures of the breast tumors and the risk associated with the outcome of therapy in several clinical protocols. Using Monte Carlo simulations and phantom experiments, we are also continuing to expand and substantiate our analysis of photon migration inside turbid tissues, especially for decreased scattering abnormalities that require special treatment, as some tumors as well as cysts do show decreased scattering. We have devised a theory that accounts for different diffusion coefficients for structured tissue such as white matter of the brain or muscle. In this formalism, photons entering the tissue will diffuse mostly in the direction of the bers rather than in the perpendicular directions. We will design special phantoms to test the theory. Chernomordik V, Hattery D, Gannot I, Zaccanti G, Gandjbakhche AH. Analytical calculation of the mean time spent by photons inside an absorptive inclusion embedded in a highly scattering medium. J Biomed Opt 2002;7:486-492. Chernomordik V, Hattery D, Grosenick D, Wabnitz H, Rinneberg H, Thomas Moesta K, Schlag PM, Gandjbakhche AH. Quantification of optical properties of a breast tumor using random walk theory. J Biomed Opt 2002;7:80-87. Dagdug L, Weiss GH, Gandjbakhche AH. Effects of anisotropic optical properties in photon migration in structured tissues. Phys Med Biol 2003;48:1-10. Three-dimensional reconstruction of localized in vivo fluorescence Chernomordik, Hattery, I. Gannot, Gandjbakhche; in collaboration with G. Gannot, Russo, Smith The development of specific fluorescently labeled cell surface markers has opened the possibility of specific and quantitative noninvasive diagnosis of tissue changes. We are pursuing the development of a fluorescence scanning imaging system that can perform a "noninvasive optical biopsy" of Sjøgren's syndrome (SS) and replace the currently used histological biopsy. In our animal experiments, we have devised an imaging system to quantify the concentration of fluorescinated antibodies binding to B or T tumor cells previously injected in the oral cavity of mice. Preliminary results show that our imaging system, along with our theoretical algorithm, is able to monitor noninvasively the concentration of fluorophore signals as well as the pharmacokinetics associated with wash-out of the antibodies. We are also continuing our study of infrared-dependent fluorescent detection methods to determine the position of sentinel lymph node(s) to replace currently used detection by radioactive particles. We have extended our analysis of different phantom data and ex vivo tissue to confirm the potential of the random walk-based theoretical approach to reconstruct with good accuracy three-dimensional positions of deeply embedded fluorophores. In a series of animal experiments, we were able to use near-infrared nanoparticles to study the uptake of the fluorophores through the blood circulation and in tumors. We were able to image these dynamics in real time. We are also investigating whether our model of diffuse fluorescent photon migration is able to separate the effects of light diffusion at a given depth from the actual distributions of the fluorescent antibodies. In one experiment, we first injected the tongues of Balb-C mice with squamous carcinoma cells that are CD3- and CD19-positive and then injected the tongues with FITCI conjugated antibodies to CD3 or CD19. To assess the pharmacokinetics of the fluorescent antibodies, we imaged the tongues at different time intervals after injection of the fluorescent antibodies. In a separate experiment, we used a similar procedure except that we covered the tongues with slabs of agarose gel, whose properties mimic those of human tissue. We used two pertinent parameters, i.e., change in peak intensities of the fluorescent signals and their full width at the half maximum (FWHM) as a function of time, to study the dynamics of labeled antibodies. As expected, the results of the two experiments exhibited different functional forms as a consequence of the "degeneration" of the signal by tissue scattering. Using our model of diffuse fluorescent photon migration, we are able to "deconvolve" the signal coming from deep structures and retrieve the true pharmacokinetics of tumor cells. The results show clearly that our mathematical model is able to quantify not only the concentration of fluorescently labeled antibodies but also their clearance and diffusion through tissue. Eidsath A, Chernomordik V, Gandjbakhche AH, Smith P, Russo A. 3-D localization of fluorescent masses deeply embedded in tissue. Phys Med Biol 2002;47:4079-4092. Gannot I, Gannot G, Garashi A, Gandjbakhche AH, Buchner A, Keisari Y. Laser activated fluorescence measurements and morphological features-an in vivo study of clearance time of FITC tagged cell markers. J Biomed Optics 2002;7:14-19. Gannot I, Garashi A, Gannot G, Chernomordik V, Gandjbakhche A. Quantitative 3-D imaging of tumor labeled with exogenous specific fluorescence markers. Appl Opt 2003;42:3073-3080. Oblique angle reflectometry for noninvasive monitoring of inflammation: application to chemopreventive drugs Hattery, Hekmat, Gandjbakhche; in collaboration with Mulshine Inflammatory cell populations produce cytokines that can specifically stimulate growth of evolving cancer clones. Given that normal epithelium shares many biological properties with cancer cells, it also responds to the chronic presence of mitogenically active cytokines with accelerated growth (hyperplasia); normal cell hyperplasia provides a measure of the promotional environment of a cancer clone. At NCI, a Phase II trial is under way to determine the effectiveness of a cyclooxygenase (COX) inhibitor on oral leukoplakia. The effectiveness will be monitored by surgical biopsy with its associated morbidity. As a less morbid approach to monitoring patient response, we have developed a noninvasive quantitative epithelial inflammation-measuring device that may be used to evaluate the general state of the oral mucosa and monitor the effectiveness of chemopreventive treatment regimes. The device measures diffuse reflectance from a low-power optical wavelength source, which inserts photons at specific angles into the tissue. As the angle of insertion approaches parallel with the epithelial surface, the mean penetration of the photons becomes smaller and the photons spend more time in the epithelial layer. Wavelengths of interest are those with high tissue absorption, thus limiting detected photons to those with few scattering events and conning to the surface layers the tissue volume that has been interrogated. From a theoretical point of view, this approach permits the use of integral equations to describe photon migration in a two-layer tissue model. We performed several phantom experiments by using acrylamide gel with high absorption covered with a thin liquid layer to simulate the epithelial layer of tissue and compared the data with the theoretical predictions. Using our theoretical model, we collected and analyzed 10 sets of data from five patients and two controls. Initial results indicate that the model is effective in indicating the level of inflammation in patients. An improved device using flexible imaging fiber bundles is also under construction. In this new design, an imaging bundle replaces a set of optical fibers. Instead of obtaining output readings from each fiber sequentially, a CCD camera will record light distribution instantaneously, promising shorter acquisition time and higher spatial resolution. We have run high-resolution Monte Carlo simulations on a high-performance computing cluster for comparison with data obtained from phantoms. Analysis of the data indicates that epithelial thickness may be quantified with an uncertainty of approximately 30 percent, even if patient variations in scattering are as high as a factor of four and variations in absorption are as high as a factor of 10 Thus, the new technique should be particularly robust as well as clinically relevant. A company is currently evaluating the technology for licensing. Hattery D, Gerdelman K, Hekmat F, Chernomordik V, Smith P, Eidsath A, Atkinson J, Mulshine J, Gandjbakhche A. Using diffuse reflectance spectroscopy to quantify inflammation of the oral epi-thelium in vivo. In: Alfano R, ed. Proceedings of the SPIE Photonics West Conference, Progress in Biomedical Optics and Imaging: Optical Biopsy IV. 2002;3:59-70. Hattery D, Hassan M, Chernomordik V, Mulshine J, Gandjbakhche A. Measuring oral inflammation in vivo with diffuse reflectance spectroscopy. Proceedings of the IEEE EMBS/BMES Conference. Houston, 2002;2243-2244. Blood circulation in Kaposi's sarcoma and complex regional pain syndrome type I Hassan, Hattery, Hekmat, Vogel, Byram, Gandjbakhche; in collaboration with Dionne, Yarchoan We are investigating three imaging modalities to quantify different parameters associated with blood circulation. In the first case, thermography provides a two-dimensional image of superficial skin temperatures. The concept is that higher temperatures occur in the skin superficial to veins that are involved in active transport of blood. In the second case, laser Doppler imaging (LDI) produces two-dimensional images of blood flow over a defined area at 690nm and 780nm. In the third case, multispectral imaging, designed in our unit, produces two-dimensional images at six wavelengths (700 to 1000nm) by using a differential absorption algorithm (devised by our unit). The multispectral imaging technique permits determination of parameters related to the oxy-/deoxy-hemoglobin ratio (i.e., oxygenation) and total blood content as well as of changes in other metabolic and disease-related compounds such as cytochromes and ferritins. Two studies using the techniques are now under way. An NCI-sponsored clinical trial is evaluating the effectiveness of anti-angiogenetic drug treatment in Kaposi's sarcoma (KS), a highly vascular tumor, in whose development and progression angiogenesis and capillary permeability can play an important role. Responses to KS therapy are now generally assessed visually by measuring and palpating the numerous lesions and using complex response criteria; no noninvasive standard technique is available to assess the effect of anti-angiogenetic therapy on blood flow. The purpose of the clinical trial is to test three noninvasive methods to assess of vascularity and the vascular changes associated with KS. For 20 patients, we recorded thermography, LDI, and multispectral images over the lesion for comparison with normal skin either adjacent to the lesion or on the contralateral side both before therapy and then after administration of a regimen of liposomal doxorubicin and interleukin-12 for 18 weeks. Comparative image analysis between LDI and thermography showed a strong correlation between results obtained with the two techniques. Measured by LDI, the KS lesions generally exhibited increased temperature and blood flux compared with normal skin. Interestingly, we observed that few patients had lower temperature in the lesion than normal. The multispectral data show a large increase in blood volume as well as intriguing patterns of cytochrome concentration changes within the tumor. These patterns may yield information on the detailed effects of drugs on specific tumors. Our techniques are objective, easy to perform, and appear to be highly sensitive in assessing improvement in the lesions upon administration of therapy; thus, they may have utility in monitoring trials of anti-angiogenesis therapy in KS patients. Although complex regional pain syndrome type I (CRPS-I), formerly known as reflex sympathetic dystrophy (RSD), has long been recognized clinically, its pathogenesis is not understood. We are investigating the sensitivity and specificity of infrared thermography and LDI for tracking physiological parameters associated with CRPS-I before, during, and up to five weeks after therapy. Subjects are referred with a diagnosis of CRPS in one limb only. All subjects received an experimental drug or placebo for five weeks. All treatments were double-blinded. For six patients, we compared thermograms of the painful areas with adjacent and nonpainful contralateral regions to determine the differential temperature pattern. We acquired the LDI images consecutively with thermographic imaging. The thermal pattern on the painful side before drug or placebo treatment was warmer or cooler by at least 1ºC and spread over a large area while no remarkable difference in LDI blood velocity images was observed between the two sides. By applying cold stimulation to an extremity for 10 seconds, the temperature dropped suddenly on both sides and increased over time; thermal recovery time was longer on the painful side than the pain-free side. After the drug treatments, thermography showed smaller temperature differences between pain and pain-free sides and an improved thermal recovery time. To date, the sample size is too small to draw firm conclusions. However, the results suggest that thermography can be used to locate and track the response to treatment of neuropathic dysfunction as the cause of chronic pain. Hassan M, Hattery D, Chernomordik V, Toda K, Fukuhara K, Mittal D, Rowan J, Shah J, Gerber L, Dionne R, Kopin I, Gandjbakhche A. Infrared thermographic imaging for the assessment of temperature asymmetries in reex sympathetic dystrophy. Proceedings of IEEE EMBS conference. Cancun, Mexico, 2003; in press. Hassan M, Hattery D, Vogel A, Chernomordik V, Hekmat F, Aleman K, Wyvill K, Merced L, Little R, Yarchoan R, Gandjbakhche A. Multi-modality imaging techniques to assess Kaposi's sarcoma associated with angiogenesis. Proceedings of 7th International Conference on Malignancies in AIDS and Other Immunodeficiencies: Basic, Epidemiologic and Clinical Research. Bethesda MD, 2003;22. Hassan M, Hattery D, Vogel A, Chernomordik V, Hekmat F, Aleman K, Wyvill K, Merced L, Little R, Yarchoan R, Gandjbakhche AH. Multi-modality imaging techniques to study angiogenesis associated with Kaposi's sarcoma. Proceedings of IEEE EMBS/BMES Conference. Houston, 2002;1139-1140. Hattery D, Hassan M, Demos S, Gandjbakhche A. Hyperspectral imaging of Kaposi's sarcoma for disease assessment and treatment monitoring. Proceedings of Applied Imagery Pattern Recognition. 2002;124-132. Stochastic modeling of tumor-induced angiogenesis in a heterogeneous medium: the extracellular matrix Amyot, Pourzand, Gandjbakhche; in collaboration with Camphausen Angiogenesis plays an essential role in the establishment of malignant tumors. Tumors stimulate angiogenesis by releasing tumor angiogenic factors (TAF), such as vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF). In response to these angiogenic stimuli, endothelial cells need to degrade the extracellular matrix (ECM), migrate through the extracellular matrix, proliferate, and organize themselves into a patent blood vessel. The ECM constituents play an important role during both angiogenesis and vascular morphogenesis, and its cytoskeleton gives support to the cell, allows it to grow, and helps self-organization. During endothelial cell (EC) migration through the ECM, the cells secrete adhesive proteins that permit the ECs to adhere to each other and to the ECM meshwork. This adhesion aids in the construction and extension of new vessels, which organize around the meshwork and guide the formation of the future vascular network. We have started a program to model tumor-induced angiogenesis. Our stochastic model begins with the secretion of TAF from cells of a solid tumor. The TAF molecules diffuse in the ECM until the endothelial cells absorb them. In response to the TAF, recruitment and cell migration initially power sprout formation and elongation. Sprout movement is a chemotactic response to the tumor-induced TAF gradient; the locomotion of the sprout is governed by the directionality of the endothelial cells that constitute the cell tip. Additional sprout extension takes place when cell division occurs in a narrow region juxtaposed to the inactive sprout tip. The vascular network generated in numerical simulations is similar to networks observed experimentally. Our model tests the influence of the distribution of ECM material on the formation of the vascular network. We have calculated the vascular density for different ECM network densities and have found that ECM heterogeneity has a barrier effect on angiogenesis: below a given value of heterogeneity, the sprouts cannot perfuse the tumor and a vascular network cannot form. Our simulation results indicate that the ECM network function is similar to a percolation medium between the tumor-secreting TAF and the vascular sprout. Depending on the ECM heterogeneity, a threshold for perfusion, comparable to the classical value of an ECM percolation threshold, exists, which allows the tumor to become vascularized. In collaboration with our colleagues at NCI, we are devising cell culture experiments to relate our theoretical finding to quantifiable angiogenesis processes. COLLABORATORS Kevin Camphausen, MD, Radiation Oncology Branch, NCI, Bethesda MD Rinaldo Cubeddu, PhD, Politecnico di Milano, Italy Raymond Dionne, DDS, PhD, Pain and Neurosensory Mechanisms Branch, NIDCR, Bethesda MD Gallya Gannot, DDS, PhD, Laboratory of Pathology, Center for Cancer Research, NCI Bethesda MD James Mulshine, MD, Cell and Cancer Biology Branch, NCI, Bethesda MD Herbert Rinneberg, PhD, Physikalisch-Technische Bundesanstalt Berlin, Germany Angelo Russo, MD, PhD, Radiation Biology Branch, NCI, Bethesda MD NIH, Bethesda MD George H. Weiss, PhD, Physical Sciences Laboratory, CIT, NIH, Bethesda MD Robert Yarchoan, MD, HIV and AIDS Malignancy Branch, NCI, Bethesda MD Giovanni Zaccanti, PhD, University of Florence, Italy For further information, contact amir@helix.nih.gov |
|
