Research

    Research projects available to graduate students cover a broad range of Medical Physics topics. The following is a list of faculty research interests encompassing both theoretical and experimental approaches.

    Innovative Imaging Techniques
    (Pan, LaRiviere) MRI
    We are investigating innovative imaging techniques such as thermoacoustic tomography, diffraction tomography, reflectivity tomography, phase-contrast imaging, and other wave imaging techniques. Such innovative imaging techniques have becoming increasingly important for studying physiological processes as well as anatomic structures in small animals. The hope is that some of these may ultimately find use in humans as well. We have made considerable contributions to the development of imaging theory and algorithms for these imaging techniques. For example, much effort has been devoted to developing x-ray imaging techniques that rely upon contrast mechanisms other than absorption. Phase-contrast imaging is one such technique that exploits differences in the real part of the refractive index distribution of an object to form an image using a spatially coherent source. We have developed algorithms to reconstruct and process phase-contrast images from data acquired by use of the Advance Photon Source, a third-generation synchrotron at Argonne National Laboratory. We have also made significant contributions to the theory and practice of x-ray fluorescence computed tomography, a stimulated emission tomography imaging modality that allows for mapping the distribution of elements in a small, intact samples using synchrotron radiation.

    Evaluation and Assessment of Imaging Technologies and Diagnostic Performance
    (Metz, Armato, Chen, Giger, Jiang, Nishikawa, Doi, O'Brien-Penney, Pan)
    We have pioneered the use of receiver operating characteristic (ROC) analysis to assess the performance of radiologists and medical imaging systems, to optimize image quality, and to compare objectively and quantitatively the image quality that is available from different systems and technologies. We are also developing evaluation tools such as computationally efficient model observers that predict human observer performance in tasks of interest, and we have used these approaches to evaluate a number of clinical cases of SPECT, helical cone-beam CT, and tomosynthesis. In addition to its methodological contributions, our group creates, refines and distributes computer software for ROC analysis that is used by more than 10,000 researchers around the world.

    Diagnostic Radiography
    (Nishikawa, Giger, Doi, Metz, Pan)
    We have also pioneered the use of linear systems analysis in the evaluation of x-ray imaging systems. Many new imaging technologies such as breast CT, breast tomosynthesis, and phase contrast imaging are now characterized using concepts such as the modulation transfer function (MTF), noise power spectra, noise equivalent quanta (NEQ) and detective quantum efficiency (DQE). Through the use of these metrics, imaging systems can be optimized (in terms of spatial resolution, contrast, noise and patient dose), and rigorous comparisons of different systems can be performed. We have developed analytical and Monte Carlo-based models to simulate the photon transport in radiography to model the physics of x-ray detectors. These models are used to understand the performance limitations of CT, tomosynthesis, digital radiography, and screen-film systems.

    System and Algorithm Development of Micro-CT and Its Applications
    (Pelizzari, Chen, Pan)
    We have developed a micro-CT system with potential applications in specimen and small animal imaging. The developed system can provide 3D images with isotropic resolution around 100µm. We have used this system to perform imaging studies of lung cancer, bone cancer, and vascular vessels in small animals from various laboratories in the Biological Sciences Division. This system also serves as an experimental testbed to perform a wide variety of studies aiming at understanding the physical properties of cone beam CT systems and reconstruction algorithms.

    Advanced Cone-Beam CT
    (Pan, Pelizzari)
    In recent years we have made significant breakthroughs in CT research and our CT image reconstruction group enjoys a reputation as one of the world leaders in this area. We have developed innovative scanning strategies, theory, and algorithms for image reconstruction from data acquired with helical and other imaging configurations and for obtaining images within regions of interest (ROIs) from much-reduced data. The new concepts and theories would allow the design of innovative imaging protocols that may have significant clinical implications, such as dose reduction or image quality improvement in breast imaging, liver imaging, and cardiac imaging. One of the important advances made in the past year was the development of algorithms for image reconstruction from highly incomplete data, which are expected to find significant applications to a wide variety imaging problems in medicine, security scan, and other areas. We have also established close collaborations with clinicians in applications of the advanced imaging methods to clinical problems and with leading industrial companies on investigating advanced imaging systems.

    Improving Image Quality in Low-dose and Low-contrast CT
    (LaRiviere)
    We are working to develop computationally efficient sinogram-domain approaches that help improve image quality in low-dose and non-contrast CT exams. Continuing technological innovations have spurred growth in the number of CT exams performed annually, which has prompted concern over the associated rise in CT-related radiation dose to the population. The ability to reduce CT dose without sacrificing image quality would have significant beneficial public-health consequences. This is especially pertinent given that widespread use of CT in screening for lung and colon cancer is currently being considered. The data measured in a low-dose scan is necessarily noisier than data measured in a standard scan, and this can lead to an overall mottled appearance in the images as well as to noise-induced streak artifacts than can obscure subtle lesions. In the past few years, our group has developed a novel strategy for CT data processing that entails application of a statistically principled penalized likelihood estimation of the ideal, low-noise line integrals needed for image reconstruction from the noisy degraded transmission measurements. Our preliminary results suggest that dose reductions of 30% or more might be achievable relative to current clinical standard-exposure scans, and even larger dose reductions may be possible for the low-dose screening scans.

    Magnetic Resonance Imaging and Spectroscopy
    (Karczmar, Gao, Roman) MRI
    We develop computer graphics software for creation of 3D, rotating, shaded surface displays of organs and lesions detected in the usual tomographic MR images, special radio frequency pulse sequences and data manipulation software for enhancement of contrast in tomographic images, development of fast MR imaging methods that utilize prior knowledge of anatomy, development of new methods of acquiring and analyzing functional MR images, use of MR to measure tumor metabolism, blood flow, and response to therapy. We also develop of an in-vivo current density MRI technology to measure the distribution of the current density within the body when an external current source is applied.

    Functional Magnetic Resonance Imaging (fMRI)
    (Gao)
    The development of functional MRI (fMRI) technology is aims for the understating of the neuronal activity of the human brain in the control of the body movements and in performing various cognitive tasks. The experimental platform for our fMRI research program is based on a high filed and research-dedicated MRI scanner. We develop and optimize both fMRI data acquisition and data analysis strategies to better localize the neuronal activity accurately. Cerebral blood flow, blood volume, and tissue oxygenation during brain stimulation has been measured and evaluated. Diffusion tensor imaging is being developed as one of biomarkers for determination of brain disorder that associated with white matter damage.

    Electron Paramagnetic Resonance Imaging
    (Halpern, Pelizzari, Karczmar, Pan)
    We are actively engaged in development of EPR oxygen imaging with application to tumor physiology and response to therapy. We are also investigating EPR based techniques to image molecular biologic cell signaling. Active areas of investigation in instrument design include rapid scanning continuous wave techniques; magnet design, construction, and evaluation; novel techniques for pulsed EPR projection acquisitions; resonator design, construction, and performance evaluation. In collaboration with chemistry colleagues, we are pursuing development of novel injectable spin probes with sensitivity to various aspects of body fluids with distribution in various (controlable) fluid compartments. We are also researching novel tomographic and non-tomographic image acquisition strategies, and the scaling of EPR imaging technology to larger biologic objects.

    Algorithm Development of Fast Spin Resonance Imaging
    (Karczmar, Roman, Halpern, Pan) MRI
    We are conducting research on fast spin-resonance-based imaging. Nuclear magnetic resonance imaging (MRI) has become one of the dominant clinical and research imaging techniques in recent years. We have been investigating, developing, and evaluating a variety of efficient and accurate sampling schemes such as the spiral sampling schemes for fast spatial-spectral MRI. We are also developing algorithms for optimal suppression of data noise and other artifacts that are likely to be contained in rapidly acquired data.

    Electron paramagnetic resonance imaging (EPRI) detects spin resonances of unpaired electrons of free radicals in samples and determines the spatial distributions of parameters of physiologic significance. We are investigating innovative sampling schemes for efficient and complete acquisition of EPRI data and are developing algorithms for optimally processing the acquired data and for accurately reconstructing EPRI images. One of the important breakthroughs that they have made in the area is that we have proposed algorithms for exact reconstruction of ROI images from reduced EPRI data.

    Nuclear Medicine Imaging
    (Kao, Chen, Pan, Metz)
    We conduct pioneering research in several areas in positron emission tomography (PET) and single-photon emission computed tomography (SPECT). We investigate innovative system design approaches for building high-performance PET systems and have developed a small-animal PET scanner (microPET) that is about an order of magnitude more sensitive than other microPET systems. We collaborate with high-energy physicists from the University, Argonne National Laboratory and Fermi National Accelerator Laboratory in evaluating new photo-detectors, investigating non-conventional detector designs and developing digital data-acquisition architectures for PET and time-of-flight (TOF) PET. We co-develop an ultra-high resolution SPECT scanner having resolution better than 100 microns. We develop novel reconstruction theories and methods for PET and SPECT, including spatial-temporal image reconstruction, list-mode image reconstruction, efficient methods for correcting scatter and other data degradations, incomplete-data image reconstruction, multi-modality image reconstruction, and TOF PET image reconstruction. We also develop and evaluate methodologies for quantitative PET and SPECT data analysis, including kinetic modeling and data-driven statistical image analysis. Other research areas include: 3D image visualization; multi-modality image registration; application of artificial intelligence and neural networks to nuclear medicine imaging; application of system theory for analysis of the radionuclide imaging process; and task-based system performance evaluation.

    Computer-Aided Diagnosis (CAD) and Quantitative Image Analysis
    (Armato, Doi, Giger, Jiang, Nishikawa, Suzuki, Metz) MRI
    As one of the leading research groups in the field, we pioneered and are developing computer-aided detection (locating suspicious regions in an image) and computer-aided diagnosis (classifying a suspicious region into different types of disease states) for a wide variety of applications including: detection and diagnosis of lung nodules in chest radiography and thoracic CT, detection and diagnosis of lesions and assessment of cancer risk on multi-modality breast images (mammography, ultrasound, MRI); detection and diagnosis of lesions on breast tomosynthesis images and breast CT scans; automated analysis of lung texture patterns in chest images; computerized analysis of bone structure; detection of polyps in CT colonography; MRI detection and characterization of liver tumors in dynamic hepatic CT; and analysis of cardiac CT images. We also conduct theoretical and experimental studies on the effects of training, size and composition of image databases, and the effects of scoring criteria on CAD performance. We also conduct studies on optimizing the human-computer interface for CAD and the clinical impact of CAD. Quantitative image analysis is an integral part of CAD schemes, and we have specialized techniques for extracting quantitative information from images, such as computerized assessment of tumor response, breast cancer risk using texture analysis of mammograms, and risk of bone fracture using texture analysis of bone radiographs.

    Multi-modality Image Correlation
    (Pelizzari, Chen, Halpern, Karczmar)
    We are engaged in ongoing investigations of software techniques to correlate multiple volumetric imaging studies of patients (CT, MRI, SPECT, PET), and to apply these techniques to clinical problems in radiation oncology, diagnostic radiology, nuclear medicine, neurology and surgery.

    Applications of Imaging Techniques to Radiation Therapy
    (Pelizzari, Pan, Wiersma) MRI
    Radiation therapy is one of the primary modalities for cancer treatment. We have ongoing research projects in the areas of 3D conformal, intensity modulated and image guided radiotherapy (IMRT, IGRT). The goal of conformal radiation therapy is to deliver a high radiation dose to the tumor volume, while minimizing the radiation exposure of healthy tissues that surround this volume. Therefore, position knowledge of the tumor is a critical step for accurate and effective radiation therapy. We are presently investigating both external marker and dose-efficient internal X-ray image guidance methods for detecting and correcting patient positioning errors in high precision radiation therapy treatments. By introducing state-of-the art image acquisition and reconstruction methods into radiotherapy image guidance, we are developing techniques that can lead to improved treatment accuracy while simultaneously reducing unnecessary imaging X-ray dose to the patient.

    Functional Imaging for Therapy Guidance and Assessment
    (Chen, Pelizzari, Karczmar, Al-Hallaq, Halpern, Yenice, Aydogan, O’Brien-Penney)
    Functional image information can be utilized both for optimizing the design of therapies, and for evaluating their effects. We have numerous research projects involving nucleaer medicine, vascular and functional MRI, EPR oxygen imaging and emerging molecular imaging techniques aimed at improving the targeting of therapies, and the use of functional image information to adapt therapeutic regimens to the pattern of response of the patient.

    Computer Applications in Radiation Therapy
    (Pelizzari)
    Three dimensional treatment plan calculations and graphic display, use of advanced imaging techniques (CT, MRI, PET) in radiation therapy planning and assessment of efficacy, techniques and applications of multi-modality image correlation, quantitative evaluation and optimization of treatment plans.

    Development of Advanced Techniques for Therapy Delivery
    (Al-Hallaq, Aydogan, Yenice, Pelizzari, Reft)
    We are actively engaged in developing advanced computer simulation and evaluation methods for intensity-modulated and image-guided radiotherapy; development and implementation of techniques for stereotactic body radiotherapy / radiosurgery; researching strategies for minimization of dosimetric uncertainty due to patient motion; developing advanced methods of total body irradiation and intensity-modulated total marrow irradiation; and investigating radiobiological effects of treatment delivery time and field sequencing. We are pursuing the development of novel techniques for treatment delivery for breast, metastasis and gynecological cancer radiotherapy.

    Dosimetry
    (Reft)
    In radiation therapy we need to make accurate measurements in various types of radiation fields including those with mixed neutrons and photons as well as pure photons and electrons. For these measurements we utilize a number of different types of detectors such as ionization chambers, proportional counters, thermoluminescent dosimeters, optically stimulated luminescent dosimeters, silicon diodes, diamond detectors, Geiger tubes and neutron activation analysis. We also perform Monte Carlo and transport calculations to verify our measurements as well as to extend the calculations to areas where measurements are not practical. All of the measurements and calculations are directed toward various applications in radiotherapy, radiation protection, and radiobiology.

    Dose Computation and Verification
    (Reft, Pelizzari)
    Development of algorithms for the computation of absorbed dose distributions for radiation-therapy treatments, taking into account the three-dimensional nature of the problem and the patient's anatomy, using as input computed tomography and x-ray beam attenuation and scatter data; verification of the results of the computations by direct measurements in anthropomorphic phantoms; experimentation with interactive methods of displaying the results in order to optimize information transfer to the physicians.