Noninvasive Real-Time Prostate Tracking Using A Transperineal Ultrasound Approach

Interfraction image guidance in the management of prostate cancer reduces uncertainty and has become standard of care. More recently, real-time fiducial/transponder tracking to further reduce uncertainty from intrafraction target motion has become available. Currently available techniques for prostate tracking require the placement of markers. We describe a new transperineal ultrasound system that offers a noninvasive means of correcting for interfraction and intrafraction motion. This system uses a rigid registration algorithm to calculate changes between a current image and the reference image allowing the calculation of translations and rotations every 0.5 sec. Traditional trans-abdominal ultrasound-based techniques are used for interfraction adjustments; this approach is affected by user variability (pressure and approach), anatomy, bowel gas, and bladder filling impacting image quality and accuracy. In addition, trans-abdominal ultrasound is not amenable to real-time tracking. The transperineal ultrasound imaging approach significantly reduces these issues. A programmable 3D motion phantom consisted of a tank filled with a water/Zerdine mixture, and motorized control of a target along thee translational axes in the liquid. As the submerged target could only be tracked by the Clarity 4D system, Calypso beacons as well as optical markers were affixed to an external stem which moved in tandem with the target within approximately 1 mm. A 3D sine wave as well as 6 motion profiles taken from published studies were programmed and transferred to the phantom. A large unidirectional excursion was performed prior to each sequence to synchronize the data sets. Data from all 4 sources (Calypso transponders, optical imaging, 4D-US and the programmed motion) were compared in 3 directions (ant/post, Sup/Inf, L/R). Mean, Standard deviation, 25%, 50% and 75% percentiles of the directional differences as well as the min, mean, 50%, 90%, 95% percentiles and max value for the absolute directional differences were calculated for each profile. Phantom motion tracked via the optical and the Calypso systems showed the 95% bond of the maximum distance variation to be less than 0.6 mm. There was no consistent directional error. For the Clarity system, 95% of the maximum distance variation was within 1.3 mm of programmed motion, the majority were less than 1 mm. No directional preference was identified. Part of this variation may be attributed to the 1 mm uncertainty between the internal target and the external stem. Phantom studies show that a novel 4D transperineal ultrasound system accurately and reproducibly tracks prostate phantom motion. Transperineal ultrasound tracking of the prostate offers a noninvasive alternative to currently marker-based systems.