Prospective Assessment of Image Registration for the Diagnosis of Prostate Cancer

Each biopsy session would be preceded by mpMRI, which would be delineated and assigned a degree of suspicion by a radiologist (see above).The PI-RADS scoring system will be used to assign a degree of suspicion to regions of interest within the prostate. A second reader will independently score the RSI on a Likert scale, blinded to the other MRI data. The regions of interest will be delineated using software developed by Eigen in collaboration with a study co-author (D.M.), now commercially available and in use by the UCLA team for the past 2 years (ProFuse, Eigen).The RSI data will be integrated with the standard mpMRI data and any change in scoring or presence of additional lesions, determined by RSI, will be quantified. For men with a MR-visible target of PI-RADS score 3 or more, irrespective of RSI score, the biopsy session would then proceed in an ordered routine, as follows:

1. Conventional ultrasound-guided 12-core systematic biopsy would be performed first. This portion will be performed without operator knowledge of the MRI report, i.e., the urologist will be blinded to possible tumor location and use the method in standard practice throughout the U.S. for many years.
2. Next a targeted biopsy would be performed using visual guidance (cognitive fusion), under the supervision of a radiologist specializing in prostate MRI. The radiologist will be in the biopsy suite and help the urologist direct needle at location of region of interest in the prostate seen on MRI. Three directed biopsy cores will be obtained.
3. Third, a targeted biopsy using Artemis device fusion of MRI and ultrasound images would be performed. The prostate will be scanned and the MRI region of interest (target) brought into the 3D model via device fusion. Targeted biopsy will be performed by taking three cores of tissue from the target area, visualized as a 3D region in the fusion device.

Biopsy sites to be dictated by geometric guides (12 point pattern vs visual direction of radiologist vs fusion target), not chosen arbitrarily.

The above biopsy schema will not require any more procedure time or samples taken than fusion biopsy as performed under IRB approval at our institution for the last five years. Additional cores will be required for the visual biopsy method, but no biopsy cores will be obtained from secondary targets. Most patients exhibit secondary targets. An analysis of data from past 2 years demonstrated that the chance of a secondary target showing significant cancer, not present in a primary target, is less than 1%. Therefore, secondary targets will not be sampled, as cores are instead taken from primary targets using the two methods. In 200 men undergoing initial biopsy, an average of 17 +/- 3 S.D. cores/patient has been obtained. In the present proposal, 18 cores will be taken. Thus, the number of cores/patient in this trial will not substantially exceed the number that has been routinely taken in our practice in the past.

A sampling method of three directed cores per target was chosen as a compromise between what is clinically feasible and a statistical ideal of taking additional cores for significant cancer detection in lower grade targets.

The cognitive biopsy will require approximately 90 seconds of additional time, but this added time will be more than compensated by the reduced time obtained from excluding secondary targets. The overall biopsy schema should require no more (and probably less) than the 15-20 minutes/procedure as in the past. Biopsies will be performed by an experienced team, which has been working together since 2009.