Parallel@Illinois Special Seminar Series

Medical Imperative: New Questions to Be Answered by Metabolic Imaging

Wednesday, February 18^th at 4:15 pm CT, B02 CSL

Keith R. Thulborn, M.D., Ph.D., Director of the Center for Magnetic Resonance Research, University of Illinois-Chicago

ABSTRACT: Current clinical MR imaging, the imaging modality of choice for examination of the human brain, is based on the signal arising from the hydrogen nucleus (proton). The intrinsic insensitivity of the MR signal has been compensated by using a signal from the high concentrations of protons in water (80M), by increasing the static magnetic field strength of clinical scanners (currently 3.0 Tesla) and by implementation of other technological advances in signal detection (e.g., parallel imaging). These advances have been driven by the need to produce higher quality images in less time within a business plan that remains viable in a restrictive healthcare reimbursement environment.

There is no doubt that proton MR imaging contributes significantly to patient care and has done so over the last two decades by providing exquisite anatomical images and physiological and functional maps of the human brain. These developments have been rapidly extended to other body regions. However, the impact of much of such imaging has been on advanced stages of disease after the disease is apparent medically. Treatments for the advanced states of cerebrovascular disease, neurodegeneration and brain tumors have proven to be costly and of limited success in returning the patient to premorbid levels of quality of life.

As the study of human genomics and proteomics begins to stratify risk for disease processes prior to clinical manifestation, sensitive methods are required to monitor the earliest stages of disease onset and progression in high-risk populations. The earliest expression of disease is revealed by local changes in cellular metabolism that is the functional expression of the cellular machinery (proteomics) translated from the cellular architectural template (genomics). If medicine is to capitalize on these new risk data, there is a need for monitoring the multiplicity and complexity of genetic expression at the site of disease onset. Thus, imaging must move towards metabolic biomarkers of the integrity of the cellular machinery (metabolomics) while maintaining its anatomic, physiological and functional underpinnings.

Magnetic resonance signals are produced by a number of low atomic weight nuclei that make up biologically important metabolites although at reduced sensitivity compared to that of protons. These include sodium-23, oxygen-17, phosphorus-31, carbon-13 and nitrogen-15. Intrinsic sensitivity and concentrations are reduced compared to protons in water and fat. However, strategies for improving MR sensitivity are already available to move towards sensitive MR imaging biomarkers of cellular function. Potential biomarker signals arise from sodium-23, phosphorus-31, carbon-13, oxygen-17 and nitrogen-15. These signals have been used to monitor central metabolic pathways using MR spectroscopic techniques but have been largely qualitative and used in the late stages of disease. These techniques have been confined to a few sophisticated research groups because of the sophistication of the engineering and the complexity of the interpretation.

The early expectations for sodium imaging in the 1980s (Hilal and colleagues) went unmet until improved technologies became available in the 1990s. The combination of an efficient acquisition scheme for quantitative sodium imaging (Boada et al., 1993) and the enhanced sensitivity of the clinical 3.0 Tesla scanner introduced in the mid 1990s realized the clinical use of sodium imaging (Thulborn et al., 1996). Again, the widespread implementation of sodium imaging on clinical scanners has been restricted to “early adopters” due to the lack of multi-nuclear capabilities on most clinical scanners, the absence of a reimbursement code to sustain a viable business plan and few radiologists willing to interpret quantitative metabolic information.

Other nuclei such as carbon-13 and phosphorus-31 have remained in the research domain despite speculation about potential clinical applications.

As the costs of healthcare continue to escalate while remaining oriented towards treating advanced and end-stage disease, the NIH roadmap has been refocused to encourage medical interventions earlier in the disease process which are less expensive for society while have a greater impact on the length of the quality of life of the individual patient. This strategy of personalized healthcare leverages the advances of genomics and proteomics but still requires sensitive methods for monitoring of the localized cellular consequences of the risk factors and the interventions themselves. Anatomically-based proton MR imaging of water and fat is unlikely to provide a sensitive solution. Metabolic MR imaging based on accessible non-proton signals offers a potential solution. Just as the introduction of the 3.0 Tesla clinical scanner supported the expansion of clinical blood oxygenation level dependent (BOLD) functional MRI (fMRI), the development of ultra-high field magnets at 7.0 and 9.4 Tesla for human imaging ha s increased sensitivity so that non-proton images (e.g., sodium-23, oxygen-17) can be acquired at adequate resolution (e.g., ~5 x 5 x 5 mm3) in acceptable time times (e.g., <10 minutes).

The use of metabolic MR imaging as the sentinel test for early changes in central pathways of cellular metabolism will first be discussed using sodium MR imaging as the prototype signal. It has been used at 1.5 and 3.0 Tesla as a measure of tissue viability in stroke. As the maintenance of sodium ion gradients across cell membranes is a fundamental energy-consuming cellular process for cellular integrity, loss of energy production compromises this process and ultimately indicates loss of viability. Tissue sodium concentration (TSC) increases as the normally low intracellular sodium concentration (~10mM) is replaced by extracellular sodium concentration (140mM) that is maintained by the body’s sodium ion homeostatic mechanisms (renal function, endocrine function). Thus, TSC increases as intracellular volume becomes extracellular space with loss of cellular integrity. This mechanism has been used to address the important medical problem of classifying local tissue regions as viable or non-viable. In the setting of stroke when the goal of treatment is to salvage tissue with compromised perfusion, knowing if viable tissue still exists is fundamental to the decision to treat. Sodium MR imaging has been used to address this question directly (Thulborn et al., 1999). The opposite scenario exists in oncology where treatment aims to selectively kill neoplastic tissue. Treatment involves toxic pharmaceutical agents and radiation to which individual tumors may or may not respond. Rather than using population statistics for treating individual patients, a direct measure of regional cell kill would allow personalized healthcare in which treatment could be tailored for each patient and their tumor. This application is now being investigated in human brain tumors during radiation treatment (Thulborn et al., 2005). The human applications of MR signals from other nuclei besides sodium are still in their infancy. Early examples include carbon-13, oxygen-17 and phosphorus-31 will be briefly mentioned.

Metabolic imaging can be enhanced with new signal processing techniques and new hardware. Strategies that have been developed for proton imaging are being capitalized on to enhance spatial and temporal resolution into acceptable ranges for human imaging of non-proton MR signals. The use of a priori information from proton anatomic imaging co-registered with non-proton signals allows enhanced anatomic display of the metabolic information without cost of increased acquisition time. Anatomically constrained image reconstruction will be demonstrated for sodium and oxygen MR imaging. The goal of this lecture is emphasize the need for speed if metabolic imaging is to be used in clinical trials of metabolic imaging.

BIO: Keith Thulborn received his Ph.D. in Biochemistry from University of Melbourne, Australia in 1980, traveled to Oxford University, England, as a post-doctoral researcher from 1979 to 1981 and moved to the USA for medical training, receiving his M.D. degree from Washington University, St. Louis, MO in 1984.

He then moved to Boston for his internship in pediatrics at the Boston Children’s Hospital followed by residency and fellowship training in radiology at Massachusetts General Hospital. He became at attending in Radiology at this hospital and was associated with Harvard University from 1989 to 1993, first as Instructor, then as Assistant Professor and finally as Associate Professor of Radiology. He moved to the University of Pittsburgh Medical Center in 1993 to develop one of the earliest high-field MRI Centers for Functional Neuroimaging and developed the first clinical 3T MRI scanner with GE Medical Systems. He left as a full Professor of Radiology to move to the University of Illinois at Chicago (UIC) in 2000 as tenured Professor of Radiology, Physiology & Biophysics, and Director of the Center for Magnetic Resonance Research at UIC. Dr. Thulborn has made fundamental contributions to functional neuroimaging in both imaging technology and clinical applications. Under his leadership, the Center for MR Research at UIC is now operating the world’s first 9.4T MRI scanner for human metabolic neuroimaging, which promises to provide unprecedented opportunities to decode the human brain function at the biochemical level.