UCLA

Methods to diagnosis prostate cancer, a disease affecting approximately 240,000 men in the U.S. annually, have remained largely unchanged in the last several decades. An increased level of prostate specific antigen (PSA) is the usual initiating event followed by an ultrasound-guided biopsy. Such biopsies are performed in a systematic, but blind manner, and tumor discovery is often fortuitous. Furthermore, such biopsies often cannot differentiate between serious, potentially lethal forms of prostate cancer and insignificant, indolent forms. This inadequate method of diagnosis has led to over-treatment of indolent disease, a major concern due to the quality-of-life issues of impotence and incontinence associated with curative treatment.

Targeted biopsy utilizing multi-parametric magnetic resonance (MR) imaging may comprise an important advance in prostate cancer diagnosis. MR-guided biopsies, while effective, suffer from high cost, limited availability, and long procedure times. MR-Ultrasound (MR-US) fusion, marrying the predictive accuracy of MR and the real-time capabilities of ultrasound, offers an alternative that can be performed in most outpatient settings, while potentially retaining the cancer detection accuracy of MR-guided biopsy.

This thesis presents comprehensive research studies that validate targeted biopsy using MR-US fusion. We found that the use of image fusion in targeted prostate biopsy yielded an improved cancer detection rate in a low-risk population. Further, we discover that fusion is appropriate for men with prior negative biopsies and elevated levels of prostate specific antigen (PSA), some of whom may be screened using MRI. In men undergoing repeat biopsy to rule out cancer, we observed a cancer detection rate of almost four times that usually reported (37% vs. 10%).

We also discover that significant components to errors in targeting are volume accuracy and registration between MR and TRUS. To this end, this thesis presents a novel method of real-time 3D prostate imaging suitable for image fusion, transurethral ultrasound (TUUS). A number of engineering challenges have been addressed to bring this concept to realization: a catheter-based transducer theoretically capable of volumetric imaging of the prostate was fabricated and evaluated; reconfigurable hardware was designed to provide flexibility in imaging techniques; and image reconstruction techniques were developed and implemented for MR-US fusion.