This proposal describes a research program that will lead to the development and validation of clinically useful techniques for quantitative imaging of tissue perfusion in the brain using magnetic resonance. The ability to quantitatively determine perfusion is widely applicable to diseases of the brain, including disorders of intracranial pressure, traumatic conditions with closed head injuries, malignancy, cerebrovascular disease, and degenerative conditions. Preliminary studies in our laboratories have indicated that the currently accepted random flow model of perfusion may not accurately reflect the actual nature of perfusion in the brain. Furthermore, it is not known how the motion of protons is related to """"""""blood"""""""" perfusion, what the accuracy, sensitivity, or range of validity of available methods is, or how to resolve the conflicting assumptions of these models.
The specific aims of this proposal are centered around the validation of the quantitative separation of perfusion and diffusion components using MRI in mechanical phantoms, an excised animal kidney, and in an animal model (normal perfusion and induced acute ischemia in the brain). The results in the animal model will be correlated with regional microsphere measurements of blood flow and with blood flow and blood volume images determined by positron emission tomography (PET). As part of the development, new methods of presenting perfusion data quantitatively will be implemented, including 2- or 3-dimensional magnitude and direction angle images of coherent perfusion and color coded quantitative perfusion and diffusion images. Techniques for reducing or removing motion artifacts will be investigated as well. The fundamental direction of this proposal is towards understanding the relationship between what is seen in perfusion/diffusion-weighted images and microcirculation in normal and ischemic brain tissue.
