Molecular Imaging Directed, 3d Ultrasound-Guided, Biopsy System

NIH/NCI R01CA156775 (PI: Fei)

Prostate cancer affects 1 in 6 men in the USA. Every man over the age of 45 is at risk for prostate cancer. Systematic transrectal ultrasound (TRUS)-guided biopsy is the standard method for a definitive diagnosis of prostate cancer. More than 1.2 million prostate biopsies are performed annually and the medical cost is more than two billion dollars each year. However, this technique has a significant sampling error and is characterized by low sensitivity (39-52%). This "blind" biopsy approach can miss 30% of prostate cancers. As a negative biopsy does not preclude the possibility of a missed cancer, both the physicians and patients face challenges in making treatment decisions. Due to the increasing number of younger men with potentially early and curable prostate cancer, this problem must be addressed in order to improve cancer detection rates. At our NIH/NCI-supported Emory Molecular and Translational Imaging Center, positron emission tomography (PET) with the new molecular imaging tracer FACBC has shown very promising results for prostate cancer detection in human patients. We hypothesize that FACBC PET molecular images can be incorporated into ultrasound-guided biopsy for improved cancer detection. The proposed research is to develop a molecular image-directed, 3D ultrasound-guided system for targeted biopsy of the prostate. Specific Aim 1: To modify a real-time, mechanically assisted, 3D ultrasound-guided device. Compared to conventional 2D image guidance, 3D images of the prostate will be used to guide the biopsy. Specific Aim 2: To develop fast deformable and statistical appearance model based segmentation methods for 3D ultrasound images of the prostate. Statistical shape models will be developed from our database and will be used to guide automatic segmentation of the prostate. Specific Aim 3: To combine FACBC molecular images with 3D ultrasound for targeted biopsy. New deformable image registration methods based joint saliency map and fuzzy point correspondence will be developed in order to solve major limitations of mutual information based image registration. Specific Aim 4: To test the accuracy of the integrated biopsy system in phantoms and animals. The complete biopsy system will also be tested in a small number of human patients. Our FACBC patient studies are supported by the NIH-sponsored imaging center program (P50CA128301) and by the Phase II Clinical Trial (NCT00562315). If completely developed, the multimodality molecular image-guided system will be able to be used not only for biopsy but also for brachytherapy, radiofrequency thermal ablation, cryotherapy, and photodynamic therapy.

Figure 1

Three-dimensional visualization of the segmented prostate (red) as compared to the manual gold standard (gold color) in four different views (a-d) of the same human prostate.  (Akbari H, Fei BW (corresponding author). 3D ultrasound image segmentation using wavelet support vector machines, Medical Physics  2012;39:2972-2984. Featured by the Journal Cover)