UC Berkeley

In this dissertation, I investigate various applications of prepolarized magnetic resonance imaging (MRI) at ultralow fields, typically at 132 µT, detected with a superconducting quantum interference device (SQUID). One of the major advantages of working at ultralow fields is enhanced longitudinal-relaxation-time (T1)-weighted contrast. I measure T1 of healthy and cancerous prostate tissue specimens--within a few hours of their surgical removal--from approximately 50 patients. The measurements involve a field-cycling imaging technique in which I prepolarize protons in fields up to 150 mT. After this field turns off, the image of each pair of samples is encoded using magnetic field gradients, and the proton nuclear magnetic resonance signal is measured using a SQUID inductively coupled to an untuned, second-derivative gradiometer. The observed T1 contrast is significantly greater than that at (say) 1.5 T, suggesting that one may be able to distinguish tumors from healthy tissue without a contrast agent: on average I find that T1 of 100% tumor is 66% that of 100% normal prostate tissue. To make this imaging system suitable for in vivo imaging of human prostates, I integrate a 200-A, 150-mT prepolarizing coil that will adequately polarize the human prostate. Assuming a prepolarizing field of 150 mT at the prostate, and a noise of 0.2 fT Hz-1/2, we can acquire a T1 weighted contrast image of the prostate with resolution 2×2×3 mm3 in 22 minutes with a contrast-to-noise ratio of 4. Measurements of preliminary standard phantoms designed by NIST in Boulder, CO for measuring T1, proton density, and resolution are discussed. I also present calculations for a method of tuning the input coil of the SQUID at low frequencies (~10 kHz) while adding minimal noise. This tuning would be useful to block very low frequency, high amplitude drifts of the ambient second-order gradient magnetic field while preserving the high balance of the gradiometer.