The impact of mathematical models on predicting tumor growth can be enhanced with the addition of imaging data because parameters available from imaging can be obtained noninvasively, in 3D, longitudinally, and specifically for each patient. The predictions made from mathematical models parameterized by imaging data are patient specific and can readily be checked against actual measurements obtained at later time points. We show how a simple mathematical model of tumor growth can be parameterized by values obtained from sequential diffusion-weighted magnetic resonance imaging (DW-MRI). DW-MRI data obtained early in the course of therapy was used to estimate tumor proliferation rates, and these rates where then used to predict tumor cellularity at the conclusion of therapy.

MR imaging-based biomechanics can be used to non-invasively assess function and to elucidate the relationship between mechanics and disease. This talk will describe projects using MR data 1) to characterize the physiologic flow field in the vertebrobasilar system using computational fluid dynamics (CFD); 2) to study cerebral artery hemodynamics in pediatric sickle cell disease patients using CFD for the long term goal of elucidating the association between hemodynamics and stroke risk; and 3) to determine the relationship between structure and function in the lower extremity muscles in healthy subjects using fiber tracking for the purpose of applying this framework in studies with Becker Muscular Dystrophy patients.

While ultrasound (US) is well known for creating real-time 2D maps of anatomy and blood flow, emerging applications have demonstrated many exciting possibilities that will expand the capabilities of ultrasound to include image-guided localized drug and gene delivery, targeted imaging, and generation of local hyperthermia. Successful and safe implementation of these technologies requires development of novel hardware and a multi-disciplinary approach that considers the biological problem being addressed as well as underlying physical mechanisms. In this talk, I will first discuss the development of an open source computed tomography (CT) and US fusion designed to facilitate image-guided application of mild hyperthermia with ultrasound. Our real-time fused CT/US system provides navigational assistance to US for therapeutic intervention, as well as tissue mapping, which is vital for ultrasound thermometry. The CT/US system has accuracy on the order of 1 millimeter and is implemented in an extendable, open-source framework on a C-arm cone-beam CT scanner with a flat panel detector, similar to a new generation of CT scanners available in interventional radiology suites. The second part of the talk will discuss mechanisms for vascular permeability enhancement using micron-sized, lipid-encapsulated contrast agents, called microbubbles. Co-injection of these bubbles with a drug or gene of interest, followed by application of ultrasound has been shown by many researchers to enhance drug delivery and permeability; however, the mechanisms by which this occurs are not fully understood. Here, I will relate images of microbubbles undergoing MHz-scale oscillation to theoretical formulations of microbubble activity to define a parameter space for safe and effective use of microbubbles for drug delivery.

Recent advances in robotics technology have brought to the near horizon some new possibilities with respect to the development of assistive devices for purposes of enhancing the mobility and/or functionality of persons with physical disabilities. This talk will focus on the development of three such assistive devices, which are intended to provide enhanced mobility and/or functionality for persons with lower limb loss, upper limb loss, and with paraplegia, respectively. Specifically, the talk will describe the development of a powered transfemoral prosthesis for lower extremity amputees, the development of a multigrasp hand for upper extremity amputees, and the development of a lower limb exoskeleton for legged mobility assistance in individuals with paraplegia.

Evaluating axon morphology would provide insights into connectivity, maturation, and disease pathology. Conventional diffusion MRI can provide metrics that are related to axon morphology, but cannot measure specific parameters such as mean axon diameter. Q-space imaging (QSI), an advanced diffusion MRI technique, offers potential for indirect estimation of specific axonal architecture metrics by exploiting the regularity of molecular diffusion barriers from axon membranes and myelin sheaths.

QSI theory stipulates that the molecular diffusion during the diffusion gradient must be negligible. Consequently, strong gradient amplitudes not commercially available are needed to accurately study structures as small as axons. By using a custom 5000 G/cm single axis gradient and RF coil set, we have sufficient gradient strength to execute near ideal QSI experiments. With these unique capabilities, we can therefore investigate the true potential of QSI imaging to indirectly assess axon morphology. We examined excised fixed mouse spinal cords with QSI to estimate mean axon diameter, axon diameter distribution, and intracellular volume fraction and compared the results with histology.

Intracranial atherosclerosis is an important cause of TIA and ischemic stroke. With conventional angiography methods, like DSA or MRA, intracranial arterial pathology only becomes visible when it gives rise to luminal narrowing. More direct visualization of the intracranial arterial vessel wall itself could aid in faster diagnosis and treatment in these patients. A technique for visualizing intracranial arterial vessel walls with 7.0 Tesla MRI is presented, including (future) applications of this technique in the research and clinical practice setting, and validation of this technique using ex vivo material. All applications will be illustrated by patient examples. Preliminary results of the Intracranial Vessel wall Imaging (IVI) study will also be presented.

MR imaging and spectroscopy allow the non-invasive measurement of brain function and physiology, but excellent B0 magnetic field homogeneity is required for meaningful results. The correction of static magnetic field imperfections, so-called B0 shimming, is particularly demanding in the brain where air-tissue interfaces create complex and strong distortions.

Patients with ischemic steno-occlusive disease are at high risk of recurrent stroke, but the correct treatment for these patients is currently unclear. Current standard practice for evaluating stroke risk in such patients is focused on angiographic imaging which is used to visualize stenosis severity. To enhance stratification of stroke risk and guide treatment, it would be of equal importance to assess cerebral hemodynamics and vascular compliance. Currently, we have implemented a clinical protocol at VUMC which uses hemodynamic imaging modalities, such as blood oxygenation level-dependent (BOLD) and arterial spin labeling (ASL) imaging, in conjunction with a hypercarbic hyperoxic gas challenge to assess cerebrovascular reactivity. In this talk I will demonstrate how we have successfully used this protocol to assess cerebrovascular reactivity in patients with intracranial stenosis, and how these methods are particularly sensitive to lateralizing disease. Additionally, I will show how hypercarbic hyperoxic BOLD and ASL can be used to longitudinally monitor the success of cerebral reperfusion following cerebral revascularization surgery in a sub-population of patients with an intracranial stenosis of unknown etiology, referred to as Moyamoya disease. Lastly, I will demonstrate how these methods have allowed us to better understand the origins of variations in the BOLD signal.