UCSD Radiology

Robert Mattrey, M.D., rmattrey@ucsd.edu
Robert F. Mattrey, MD is the Vice Chairman, Director of Research and a member of the Body Imaging Division. His research is in developing and translating contrast media to the clinic. He directs the In-vivo Cellular and Molecular Imaging Center at UCSD and is the leader of the ultrasound contrast media laboratory. His major research effort is developing ultrasound molecular imaging tools for diagnosis and therapy. He works closely with the Chairman and Clinical Program Director to integrate the research training into the clinical program.

Julie Bykowski, M.D., jbykowski@ucsd.edu
Mentor: Eric Wong, PhD, MD
Lab: Center for fMRI
Arterial spin labeling (ASL) has developed into a powerful tool to measure cerebral blood flow (CBF) using MRI. In patients with collateral or slow flow, factors such as the dynamics of blood delivery and clearance, relaxation of the tagged blood, and the exchange of blood water with tissue water are of critical importance. My research has focused on adapting a velocity-selective arterial spin labeling (VSASL) method developed by Dr. Eric Wong, MD, PhD for potential clinical applications.

Edmund Wong, M.D., PhD, ehwong@ucsd.edu
Mentors: Roger Tsien, PhD, Department of Pharmacology and Howard Hughes Medical Institute, and David Vera, PhD, Department of Radiology
Purpose: To develop an amplifying mechanism using proteases in the tumor microenvironment to accumulate imaging and therapeutic agents

Materials and Methods: We have synthesized and tested novel peptide-based imaging agents that are substrates for matrix metalloproteinases (MMPs), a protease that is highly expressed in the tumor microenvironment. These agents concatenate optical (Cy5), magnetic (Gd-DOTA), or radioactive (99mTc(CO)3) contrast agents, a polyarginine cell-penetrating peptide (CPP), an MMP-cleavable linker, a polyglutamate inhibitory sequence, and a macromolecular carrier to modulate pharmacokinetics. Protease-mediated cleavage of the linker separates the cargo plus polycationic CPP from the inhibitory polyanion, allowing local trapping onto and into cells. These agents have given promising results in isolated cells, 3-D cultures of MDA-MB-231 tumor cells, and slices from resected human squamous cell carcinomas. We have begun tests on two murine models, nude mice bearing xenografts of HT1080 human fibrosarcoma cells and MMTV-PyMT transgenic mice that develop mammary tumors. After IV injection of the probes, we obtain serial images with far-red epifluorescence, T1-weighted MRI, or planar scintigraphy, and harvest post-mortem tissues to quantify cargo uptake.

Results: Optical imaging of Cy5-labeled probes with albumin or dextran carriers gave up to 4:1 contrast for primary tumors over neighboring normal regions in MMTV-PyMT mice, and even higher contrast in metastatic lymph nodes, both peaking around 24-48 hr after injection. MRI of Gd-DOTA-labeled probes also highlighted metastatic lymph nodes. However, 99mTc-labeled agents conjugated to albumin, dextran, or PAMAM dendrimers deposited radioactivity to a much greater extent in the liver (12-18% ID/g), kidney, and spleen than in the xenografted tumors (3% ID/g for the albumin-conjugate).

Conclusion: This approach can image protease activity in live animals, but refinement is needed to improve contrast and reduce uptake into liver, spleen, and kidney.

David Karow, M.D., PhD, dkarow@ucsd.edu
Mentor: Anders Dale, PhD
Multimodal Imaging Laboratory (MMIL)
Dr. Karow graduated from a combined MD/PhD (MSTP) program from the University of Michigan, Ann Arbor. His doctoral work was performed in Ann Arbor and at UC, Berkeley where he explored protein structure/function as well as animal behavior under the mentorship of Dr. Michael Marletta, now Chair of Chemistry at UC, Berkeley. Dr. Karow is currently working most closely with Dr. Dale where he is exploring the role of PET as a biomarker for the detection and monitoring of Alzheimer’s disease. This is a collaborative effort using the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. He works closely with Drs. McEvoy, Fennema-Notestine and Hagler. He is also collaborating with Dr. Saxena in the Department of Psychiatry in order to study human behavior with PET.

Takeshi Yokoo, M.D., PhD, tyokoo@ucsd.edu
Mentors: Graeme Bydder, MD and Claude Sirlin, MD
Lab: 3T MR Research Laboratory
Dr. Yokoo, has taken on several research projects under faculty mentorship of Drs. Graeme Bydder and Claude Sirlin. Dr. Yokoo's research focuses on (1) MR imaging of chronic liver disease and (2) mathematical theory of MR sequence optimization.

1 – MR IMAGING OF CHRONIC LIVER DISEASE
Many chronic liver disease are associated with histolopathological changes characterized by (a) fatty infiltration - steatosis, (b) collagen deposition - fibrosis, (c) iron deposition - hemosiderosis, (d) inflammation - hepatitis. Dr. Yokoo and Dr. Sirlin have active human research projects addressing (a-c).

1a - STEATOSIS: In an ongoing prospective study involving subjects with fatty liver disease, Dr. Yokoo and his collegues are studying the diagnostic and staging accuracy of various MR fat imaging techniques as state-of-the-art MR spectroscopy as the reference. If comparable to spectroscopy in accuracy, these imaging techniques would provide rapid, accurate, and widely accessible means of assessing of fatty liver disease.

1b - FIBROSIS: Dr. Yokoo is also conducting a research study to develop an objective image analysis technique to quantify early fibrosis in patients infected with hepatitis C virus. He utilizes a recently developed MR imaging technique, combined-contrast enhanced MR imaging, which uses IV gadolinium and iron oxide to directly visualize fibrosis on MR images. As the fibrotic liver exhibit distinctive texture patterns, a quantitative texture analysis technique would be able to predict histological fibrosis grade. He has received RSNA Research Resident Award in 2007 to partially fund this research project.

1c - HEMOSIDEROSIS: Dr. Yokoo is currently collecting preliminary data on techniques of iron quantification using susceptibility-weighted MR imaging. Iron-overloaded liver (hemosiderosis) is known to demonstrate T2* shortening. Using IV iron oxide infusion (extrinsic iron overload) as a disease model for hemosiderosis (intrinsic iron overload) he attempts to demonstrate that the T2* value is an reliable surrogate measure of the amount of hepatic iron deposition.

2 - MATHEMATICAL THEORY OF MR SEQUENCE OPTIMIZATION: Dr. Yokoo and Dr. Bydder are developing a novel and simple framework of MR physical theory that can be easily adopted by clinicians and applied rapidly to clinical MR pulse sequences. From original Bloch equations, a mathematically rigorous definition of "weighting" and contrast optimization are derived. From this, simple rules relating imaging parameters (TR, TE, flip angle, etc.) to image contrast emerge. This theory is applicable to a variety of pulse sequences (spin echo, gradient echo, inversion recovery, diffusion, etc.). Dr. Yokoo is conducting a human study to demonstrate his theories using new GE 3T research scanner.

Matthew V. Cronin, M.D., mcronin@ucsd.edu
Mentors: Dr. Susan Hopkins, Dr. Kim Prisk and Dr. Eric Wong
Lab: Center for Functional MRI (CFMRI)
Purpose: To improve current understanding of normal lung physiology by quantifying pulmonary perfusion during stress conditions (such as exercise). This will help form a basis for interpretation of regional perfusion abnormalities in diseases that involve alteration in pulmonary blood flow.

To date, several imaging modalities have been applied to assess lung perfusion and these include contrast enhanced CT, MRI and microsphere studies in animals. Arterial Spin Labeling (ASL) has also been utilized in lung perfusion and is a particularly attractive technique because there is no ionizing radiation and studies can be conducted longitudinally since exogenous contrast agents are not used. The most widely utilized ASL technique is a variation of FAIR (Flow-sensitive alternating inversion recovery) called FAIRER. This is essentially a FAIR ASL sequence that uses radiofrequency (RF) inversion pulses to spatially "tag" spins in the blood. While this technique has had success in initial implementations, there are several sources of error that currently exist. First, because FAIRER techniques tag blood based on spatial distribution, there is inevitably a gap between the blood that is tagged and the end point of perfusion at the imaging slice. Second, slowly flowing venous components may potentially contribute to the ASL signal if the time allowed for inflow (commonly called TI) is short. Lastly, arterial and arteriolar vascular signal, representing blood volume, is clearly present in standard FAIRER images. This confounds parenchymal perfusion quantification and is currently a source of error in ASL lung perfusion.

It is hypothesized that these errors can be minimized by a spatially independent tagging arrangement called Velocity Selective Arterial Spin Labeling (VSASL). In this technique, unidirectional gradients are applied in tandem with RF pulses. The resulting module effectively saturates spins moving faster than a prescribed cutoff velocity. This has the benefit of largely reducing the transit gap that can be introduced in spatially selective ASL modules while simultaneously eliminating blood volume signal in large vessels.

The ultimate goal of this study will be to utilize ASL to quantitatively measure regional pulmonary perfusion at varying levels of cardiac output during exercise stress. Because many pathologic pulmonary processes such as COPD, pulmonary embolism and primary pulmonary hypertension involve alterations in pulmonary blood flow, the ability to interpret regional perfusion abnormalities in these populations will depend on a clear understanding of perfusion dynamics in the normal population.

Isabel Newton, M.D., PhD, inewton@ucsd.edu
Mentors: Catriona Jamieson, MD, PhD and Robert F. Mattrey, MD

In vivo optical imaging of Wnt/b-catenin-activated leukemia stem cells

Most radiologic modalities reveal anatomic information on the tissue and organ level, although most important disease processes occur or at least begin on the cellular level. As such, an important area of research concentrates on developing or adapting imaging modalities for the purpose of visualizing cells and cellular processes. These studies have encountered a variety of obstacles, including the development of effective cellular labels (such as antibodies, microbubbles, or iron oxide particles), signal detection and background reduction, and safety concerns such as toxicity and radiation exposure. Although bioluminescence imaging has limited applications in humans, its high sensitivity makes it an excellent research tool for exploring and optimizing cellular imaging techniques. Likewise, cancer stem cells are an excellent model for developing cellular imaging techniques because of their unique cellular and behavioral properties and their obvious clinical relevance.

My research tests the hypothesis that bioluminescence imaging can be used to monitor the properties and homing of leukemia stem cells (LSC) on the cellular level. We use in vivo bioluminescence imaging to demonstrate whether LSC, a subpopulation of advanced phase chronic myeloid leukemia (CML) cells that behave like stem cells, contain activated ?-catenin, a protein critical for self-renewal in normal hematopoietic stem cells. Prior studies in our laboratory have shown in vitro that the CML cells that have acquired the stem cell ability to self-renew (the LSC) do so through activation of the Wnt/b-catenin signal transduction pathway and carry the immunophenotype of granulocyte-macrophage progenitors (GMP). Therefore, we are also testing the hypothesis that in vivo inhibition of the Wnt/b-catenin signaling pathway will result in decreased b-catenin-mediated gene transcription in the LSC and reduced tumor growth and spread. These studies will reveal whether bioluminescence imaging, which is used now to image whole tumors, can also be used to image LCSs, a subset of tumor cells which have greater relevance for cancer progression. They also will test whether those observations we have made in vitro also hold true in vivo and extend our prior findings that the CML cells critical for disease progression are not the abundant terminally differentiated blast cells but the LSC, which are enriched within the CML GMP population.

We hope that these studies will not only further our understanding of cancer progression but will also lead to the development of cellular imaging modalities that eventually could be applied to humans for the purpose of diagnosis and treatment of devastating cellular diseases such as cancer.