Physics Medical Imaging Seminars - 2007


Wednesday, 31 January 2007, 4:30pm

Digital x-ray 101

John Sabol, GE ASL


Wednesday, 7 February 2007, 4:30pm

CT 101

Jiang Hsieh, GE ASL


Wednesday, 14 February 2007, 4:30pm

PET 101

Scott Wollenweber, GE PET


Wednesday, 21 February 2007, 4:30pm

Tomographic Reconstruction 101

Sarah Patch, UW-Milwaukee, Physics


Wednesday, 28 February 2007, 4:30pm

MRI 101

Kathleen Schmainda, Medical College of Wisconsin


Wednesday, 7 March 2007, 4:30pm

Ultrasound 101

Anne Hall, GE US


Wednesday, 14 March 2007, 4:30pm

Volumetric CT - Images from GE's first 64-slice Scanner

Dr. Dennis Foley, MCW - Radiology


Monday, 26 March 2007, 1:30pm

Combining Light and Sound: Can Ultrasound Become the Preferred Modality for Functional and Molecular Imaging?

Dr. Shai Ashkenazi, Univ. of Michigan, Biomedical Ultrasonics Lab

Ultrasound imaging is widely used in medical diagnostics. It provides tissue structure imaging with sub-milimeter resolution at a depth exceeding 10 cm. Higher frequencies increases resolution (<0.1 mm) at the expense of reduced penetration. High resolution end is limited by current transducer technology. In this seminar, I will present how combining optics and ultrasound elevates the imaging in two major aspects: increasing resolution by forming high density transducer arrays and providing functional and molecular sensitivity by interaction with optical contrast agents. Clinical implementation of these techniques will have a major impact on both diagnostics and imaging assisted therapy of cardiovascular diseases and cancer. Optoacoustic transducers are based on high quality factor optical resonators for ultrasound sensing and efficient thermo-elastic materials for converting optical pulses into ultrasound emission. This alternative ultrasound technology enables high density packing of ultrasonic transducer elements in a small area, exhibiting high bandwidth operation for high resolution 3D imaging. The technology is tailored for specific medical applications such as intravascular imaging and biopsy guiding. In the second part of the seminar, photoacoustic imaging will be introduced. The technique combines optical contrast with the high resolution of ultrasound for deep tissue imaging. It relies on sound generation in tissue illuminated by a pulsed laser. Optical absorption followed by heat deposition and rapid thermal expansion creates a volume distributed acoustic source. The acoustic field is then detected by a receiver array which allows reconstructing the initial heat deposition, reflecting the distribution of optical absorption in the tissue. I will demonstrate applying 2D optoacoustic device for 3D photoacoustic imaging. Using optical contrast agents extends the scope of photoacoustic imaging to functional and molecular imaging. We have successfully applied cancer cell targeting nanoparticles as photoacoustic contrast agents. We have also studied dye indicators for imaging of tissue pH level. Future research plan in this field includes developing a range of functional imaging agents such as: intra-cellular calcium dynamics, enzymatic activity and tissue oxygenation.


Monday, 2 April 2007, 1:00pm

Optical Imaging With Ultrasonic Resolution

Dr. Hao F. Zhang, Dept. of Biomedical Engineering , Washington University-St. Louis

Although bio-optical imaging offers high contrast in visualizing both anatomical and functional information of biological tissues in vivo, bio-optical imaging suffers from short penetration depth due to strong optical scattering. Hence, existing high-resolution optical imaging modalities, including confocal microscopy, two-photon microscopy, and optical coherence tomography, are limited only to superficial tissues (~ 1 mm deep). Beyond this depth, high spatial resolution cannot be maintained since photons are multiply scattered. Ultrasonic imaging, on the contrary, provides good image resolution in deep tissues but has strong speckle artifacts as well as poor contrast in early-stage tumors. We have developed the photoacoustic microscopy (PAM) to achieve ultrasonic resolution and optical contrast in deep tissue imaging. When laser pulses illuminate biological tissues, optical absorption induces a rapid thermoelastic expansion, which produces short ultrasound pulses (referred to as photoacoustic waves). PAM detects the photoacoustic waves using a high-frequency focused ultrasonic detector to form volumetric images. Moreover, based on multi-wavelength illumination, distributions of both endogenous pigments and exogenous contrast agents can be estimated to achieve functional imaging and molecular imaging. Photoacoustic imaging promises to open new windows for basic science and diagnostic medicine, and may have important implications for the future of medical discovery. This presentation reports on the volumetric imaging of microvasculature, skin tumors, acute skin burns, hemoglobin oxygen saturation, and hemodynamics. Molecular imaging, functional brain imaging, and development of new PAM systems will also be discussed.


Wednesday, 4 April 2007, 4:30pm

Measuring and Modeling Elasticity Distribution in the Intraocular Lens

Dr. Kyle Hollman, Biomedical Ultrasonics Lab , University of Michigan

Current theory holds that presbyopia (the loss of accommodation with age leading to the need for bifocals or reading glasses) is caused by increased stiffening in the lens as we age. While recent measurements of lens stiffness do show an age related increase, they also indicate that stiffness and age-related change in stiffness are not homogeneous. To understand these effects, this study uses a novel ultrasound method to measure the elasticity distribution in a lens and then mathematically models accommodation. For the measurement technique, acoustic radiation force from ultrasound is used to push on a laser-produced bubble in in-vitro lenses. The position of the bubble is tracked with low-amplitude, high-frequency ultrasound. Knowing force and displacement, an elastic modulus can be calculated at every position a bubble is placed, mapping elasticity in a lens. Values from these measurements and other studies have been incorporated into a finite element model. In a model effects from changes in elasticity distribution can be separated from the average increase in stiffness. These measurements and modeling give us a better understanding of lens biomechanics which could lead to a laser-microsurgery correction for presbyopia.


Wednesday, 11 April 2007, 4:30pm

Optoacoustics and OCT for Biomedical Monitoring and Imaging

Dr. Rinat O. Esenaliev, Professor, Director of Laboratory for Optical Sensing and Monitoring, Center for Biomedical Engineering, Dept. of Neuroscience and Cell Biology, Dept. of Anesthesiology, Univ. of TX Medical Branch, Galveston

Two emerging techniques, optoacoustics and Optical Coherence Tomography (OCT), were recently proposed for high-resolution imaging in tissues. Optoacoustics utilizes absorption contrast in tissues and detection of ultrasound induced by short optical pulses. OCT is based on detection of backscattered low-coherent light and utilizes scattering contrast in tissues. Recently, we proposed to use these imaging modalities for noninvasive, accurate, and continuous monitoring of a variety of physiological parameters including blood oxygenation and hemoglobin concentration (optoacoustics) as well as blood glucose concentration (OCT). Our phantom, in vitro, animal, and clinical studies demonstrated high resolution and contrast of optoacoustic and OCT images and signals. The high resolution and contrast of these techniques allowed for probing and quantitative characterization of specific tissues: blood vessels, tumors, and skin layers. Moreover, optoacoustic and OCT signal parameters were linearly dependent on concentration of the blood analytes that may provide noninvasive monitoring with clinically acceptable accuracy. I will also discuss other potential applications of these techniques for medical diagnostics.


Wednesday, 18 April 2007, 4:30pm

Advanced CT recon

Dr. Guang-Hong Chen, Assistant Professor, UW-Madison, Dept. of Medical Physics; Director of Diagnostic CT Lab


Wednesday, 2 May 2007, 4:30pm

What fMRI can, can't, and might do: a closer look at the limits and potential of fMRI

Dr. Peter A. Bandettini, National Institute of Health; Chief, Section of Functional Imaging Methods and Director, Functional MRI Core Facility

Over fifteen years ago, the first papers appeared on functional magnetic resonance imaging (fMRI). During a relatively short time, the field has grown explosively in the number of people using fMRI, the breadth and depth of their research, and the range of applications. The reasons for this growth include both the ease of implementation, as well as the quality of neuronal and physiologic information obtained. The evolution of fMRI can be viewed as a progression along four highly interacting domains: Interpretation, Methodology, Technology, and Applications. In this lecture, I will give an overview of some of the latest and most exciting advances in these domains and demonstrate how advances in each domain drive and feed off the other domains. The lecture will conclude on thoughts of where the field of fMRI as a whole is going and what the ultimate limits of fMRI are in terms of its clinical and research utility.


Monday, 7 May 2007, 3:00pm

Multi-Dimensional Imaging of Moving Human Morphology

Dr. Jing Deng, Dept. of Medical Physics, University College, London

Our knowledge about 3D anatomy of dynamic body parts is based largely on invasive and non-physiological (cadaveric and intraoperative) examinations, and/or on mental reconstruction of cross-sectional images. As a result, dynamic morphology of even many fundamental actions, such as smiling or kissing, remains in the speculative and theoretical domains. This talk will cover our studies over a decade in developing dynamic 3D (4D) imaging methods and minimally-compressive scanning techniques for removing motion and deformation artifacts while attaining dynamic information. These include [1] how to use M-mode or spectral Doppler for gating fetal cardiac motion, [2] how to use complementary gray-scale and color-Doppler datasets for creating cardiac chamber - intracardiac flow fused 3D pictures, [3] how to use a water-bath and an ECG-simulator for 3D visualization of a pouting oral musculature, and [4] how to devise an acoustic vagina for real-time 4D imaging of erection. Also to be presented are subsequent clinical applications, ranging from detection of complex fetal heart malformations and intracardiac flow disorder, to evaluation of facial disfigurement and cosmetic consequences, and to assessment of erectile dysfunction in chronic and serious diseases (pre- and post-sexual enhancement medications). These results demonstrate that 4D ultrasonography can be further developed into an objective imaging modality for non-invasive observation of functional anatomy of many bodily parts in their spatial totality and temporal reality, providing a powerful tool for physiological, pathological and behavioral studies. The talk will also introduce our theory of "Direct Volume Scan", proposed for assessment of new multidimensional imaging modalities. It can be summarized in four F sentences: [1] the imaging volume must be FULL enough to cover an inseparable entirety of the volume of interest, [2] the temporal resolution must be FAST enough to distinguish dynamic events of interest, [3] the spatial resolution must be FINE enough to resolve the structural details of interest, and [4] the scanning environment must be FREE enough to avoid imposing undue constraints on the functionality of an anatomy.


Wednesday, 5 September 2007, 4:30pm

Shear Wave Crawling Waves and Tissue Elasticity

Clark Wu, GE - US


Wednesday, 12 September 2007, 4:30pm

Positron Emission Tomography - 'The Next Dimension' - Practical Aspects & 4D PET-CT

Scott Wollenweber, GE - PET


Wednesday, 19 September 2007, 4:30pm

RF Electromagnetic Simulations in MRI

Graeme McKinnon, GE - MRI


Wednesday, 26 September 2007, 4:30pm

Volume Flow Ultrasound Imaging (topic)

Anne Hall, GE - US


Thursday, 11 October 2007, 3:00pm

Iterative Methods For Image Formation in MRI

Professor Jeff Fessler, Univ. of Michigan, Dept. of Electrical Engineering & Computer Science

After a brief review of the physics underlying magnetic resonance imaging (MRI), I will describe some of the applications of iterative methods for image formation in MRI, particularly when the physics underlying the signal are modeled more completely than by the usual Fourier transform. A particular focus will be image reconstruction in the presence of magnetic field inhomogeneity, which is important in functional MRI. I will also describe signal processing algorithms that reduce computation time.


Wednesday, 17 October 2007, 4:30pm

Iterative Reconstruction in Clinical PET

Chuck Stearns, GE - PET

PET data represent line integrals of the activity distribution in the patient. As such, PET data is mathematically similar to CT data (line integrals of attenuation coefficients in the patient), and PET data can be reconstructed using filtered backprojection algorithms with the same mathematical underpinnings as CT reconstruction. However, given the poor statistical quality of PET data, the filtered backprojection reconstruction images are prone to streak artifacts and are unacceptably noisy overall. Iterative reconstruction algorithms such as ML-EM can better handle the statistical nature of the PET data and produce higher quality images. This presentation will discuss the evolution of the iterative algorithms used in commercial PET scanners over the last decade, including a focus on the interplay between the iterative reconstruction process and the corrections that must be applied to PET data, and to the accommodations made so that these algorithms can be used in a routine clinical setting.


Wednesday, 24 October 2007, 4:30pm

Photoacoustic Computed Tomography

Minghua Xu, Univ. of Pennsylvania - Dept. of Radiology

Photoacoustic (PA) computed tomography is based on the reconstruction of an internal PA source distribution from measurements acquired by scanning ultrasound detectors over a surface that encloses the source under study. The PA source is produced inside the object by the thermal expansion that results from a small temperature rise, which is caused by absorption of externally applied radiation of pulsed electromagnetic (EM) waves. This technique has great potential for application in the biomedical field because of the advantages of ultrasonic resolution in combination with EM absorption contrast. In general, different measurement geometries need different reconstruction algorithms. The three canonical geometries are planar, cylindrical and spherical surfaces. In these algorithms, the acoustic property of the tissue is often assumed to be homogenous as the speed of sound in soft tissue is relatively constant at 1.5 mm/?s. The unique advantage of photoacoustic imaging is its ability to detect the inhomogeneous EM absorption property of tissues even when the acoustic property is relatively homogeneous, whereas pure acoustic property differentiation should appeal to conventional ultrasound imaging. Spatial resolution is one of the most important parameters in photoacoustic (PA)tomography. In most cases, laser pulses with a short duration are used to excite PA signals and small-aperture detectors with a limited-frequency-bandwidth are positioned around to pick up the outgoing PA signals. Then, algorithms based on point measurements are used in the reconstruction. In this case, two major factors limit the spatial resolution-the finite frequency bandwidth and the finite detector aperture size. In this talk, we introduce a universal back-projection algorithm for the three geometries and a complete theoretical explanation of spatial resolution. Full view of acoustically homogeneous objects is assumed.


Wednesday, 31 October 2007, 4:30pm

Image Reconstruction in Photoacoustic Tomography

Jin Zhang, Ph.D. Candidate, Biomedical Engineering Dept., IL Inst. of Technology

Photoacoustic tomography (PAT) is an emerging hybrid imaging modality that combines the advantages of both optical and ultrasonic imaging principles. It has demonstrated great promises for important biomedical imaging applications from soft-tissue characterization to functional brain imaging. Image reconstruction in PAT is to solve an inverse source problem, where the source represents the optical energy absorption distribution in the object. We reexamine the PAT image reconstruction problem from a Fourier domain perspective by use of established time-harmonic inverse source concepts. We develop and implement various time-domain iterative reconstruction algorithms for PAT. We propose a novel data-space preconditioning method and systematically investigate it for different PAT image reconstruction tasks. We also investigate data sufficiency conditions and data redundancy in PAT and develop exact image reconstruction algorithms from reduced PAT data for various measurement geometry. Amplitude attenuation and wavefront distortion of PAT signals that propagate through inhomogeneous acoustic medium can lead to blurring and distortion in reconstructed PAT images. We formulate the PAT imaging model that accounts for realistic frequency-dependent acoustic attenuation and derive an efficient algorithm for correcting it. We also propose a heuristic iterative algorithm that simultaneously estimates both the acoustic speed and optical absorption distributions from PAT data alone. Assuming the heterogeneous acoustic speed of the weakly scattering object possesses a finite-dimensional parameterization, we develop an optimization strategy to iteratively solve this non-linear inverse problem.


Wednesday, 7 November 2007, 4:30pm

Small Animal SPECT Imaging of the Lung

Anne V. Clough, Professor of Mathematics and Biomedical Engineering, Marquette University

Single-photon emission computed tomography (SPECT) is becoming an increasingly valuable tool in physiology laboratories because of its relatively low cost and technology requirements. We are investigating the pulmonary vasculature of rats using SPECT to provide functional images and micro-CT for delineation of the anatomy. The imaging systems within our laboratory will be described, followed by a presentation of two specific projects. The first is a model of angiogenesis in the lung induced by ligation of one of the pulmonary arteries. This is an important model for evaluating pro- and anti- angiogenic therapies. SPECT and micro-CT are used to monitor the development of the bronchial circulation in the lung following ligation, as a measure of angiogenesis. The second is an investigation of the influence of the redox status of the lung on the pulmonary uptake of radiolabeled HMPAO, a common SPECT radiopharmaceutical. The objective of this project is to discern the ability of HMPAO to serve as an early indicator of lung injury or adaptation.


Wednesday, 14 November 2007, 4:30pm

Target-Specific Imaging of Bacterial Infection and Acute Cell Death

Ming Zhao, Assistant Professor, Depts. of Radiology and Biophysics, Medical College of Wisconsin

Non-invasive imaging provides important spatial and dynamic information in the living system, and constitutes a vital tool in research and clinical diagnosis. At the Molecular and Medical Imaging Research Leb, we are interested in characterizing the physiological processes of health and diseases in a multi-modality imaging approach, with an emphesis in translational application. To enhance this purpose, novel molecular imaging technologies are developed to facilitate new discoveries in the structural and functional roles of biomolecules. In this upcoming seminar, I'd like to share some data on our research projects, which include imaging bacterial infection in deep tissue, In vivo molecular imaging of acute cell death, and how molecular targeting can potentially enhance the efficacy of stem cell-based regenerative therapies.


Wednesday, 5 December 2007, 4:30pm

High-Frequency Ultrasound Imaging

Tim Stiles, UW Medical Physics

Clinical ultrasound scanners generally operate at frequencies between 2 and 10 MHz, providing resolutions on the order of a millimeter. The use of higher frequencies would provide greater resolution at the cost of decreasing the depth of penetration due to increased absorption of sound energy. However, the recent development of high sensitivity transducers have allowed for the construction of research devices that utilize ultrasound in the 20 to 100 MHz range with resolution approaching 10 microns. This talk will provide a brief overview of the physics of medical ultrasound imaging and insight into the development of high frequency ultrasound imaging for biomedical and industrial applications.