| Magnetic resonance imaging (MRI) has enjoyed great success in diagnostic imaging, due to its excellent soft tissue contrast. Part of this soft tissue contrast is attributed to the transverse relaxation time (T 2). Tissues with a short T2 appear darker in traditional MRI images than tissues with a longer T 2. In this thesis I will present methods that make MRI much more selective, with much greater tissue specificity. This allows direct visualization of tissues with a very short T2 such as myelin.; Myelin is the lipid sheath that surrounds neurons in the white matter of the brain, where the high speed, long range communications occur. Diseases such as multiple sclerosis (MS) are due to degradation of this myelin sheath. The ability to directly visualize myelin would improve the diagnosis of demyelinating diseases, and provide a tool for following the disease progression and response to therapy.; Unfortunately only about 10% of the signal is associated with myelin. The remaining 90% of the signal has a longer T2 relaxation time, and is not associated with myelin. My work has focused on developing techniques that image specific ranges of T 2 relaxation times. By approaching this as a filter design problem we have been able to optimize both the acquisition parameters as well as the filter weights. The complete acquisition can be performed in a clinically acceptable time frame of 11 min. In addition, the optimization has been extended to provide a tradeoff analysis between acquisition speed, image quality and accuracy.; This basic framework has a broad array of additional applications. Two will be presented in this thesis. First, we will present a method for designing T2 selective pre-pulses to directly produce very selective T2 contrast. Second, is the efficient processing of dynamic contrast enhanced studies for the detection of breast cancer. |
