UCSF

Purpose

By measuring the discrepancy of distance in electromagnetic tracking (EM) versus ultrasound images to calculate an error due to speed of sound (SOS) variances, this work establishes correctional values for inputs into post processing algorithms that will provide a correction based on B-Mode ultrasound images and consequently radiation dose plans in brachytherapy procedures.

Background

The speed of sound (SOS) within the soft tissue category is approximated to be 1540 m/s [1]. This fixed value is used for the SOS despite actual ranges in vivo varying considerably with one study measuring values of 1450 to 1613m/s in the field of view (FOV) [2]. The current state of technology doesn’t include a reasonable method for a patient-specific recalibration. Furthermore, inaccurate assignment of acoustic velocity may result in speed displacement and dispersion artifacts, especially at organ boundaries [2]. Brachytherapy for prostate cancer utilizes a transrectal ultrasound (TRUS) probe , which can provide 300-500 micron spatial resolution images for anatomical reference during needle implantation; however, the local composition and configuration of this soft tissue anatomy may exhibit a range of SOSs within the imaging field [3]. Images acquired with a SOS based on 1540 m/s and used in conjunction with stereotactic external coordinate systems can cause image distortion and displacement errors of several millimeters. In combination with the steep 1/r2 dose falloff around radioactive seeds, this can lead to erroneous dose delivery [4, 5]. Previous studies have outlined accuracy of 3-6 mm, leaving ample space for future robotics to overtake the manual process of seed implantation [6].

Method & Materials

The Aurora EM tracking system (NDI Waterloo, Ontario, Canada) generates a 500mm x 500mm x 500mm local electromagnetic (EM) field which can detect the spatial location and orientation of sensors placed within. Philips (Amsterdam, The Netherlands) has developed a brachytherapy needle with an embedded sensor that produces a signal in the EM coordinate system.

Results

Using custom-built phantoms with SOS ranging between 1430 and 1530 m/s, EM tracking resolution was verified to <1 mm precision while US localization resulted in up to 3 mm displacements from physical measurement. We present a model of correction to mitigate error in needle trajectories due to SOS variances and using these values a correctional system will be demonstrated to show the effects of changes in dosimetry due to patient specific SOS variances in the prostate. Results show SOS measurements to be within 1% of measured using a water tank system.

Conclusion

Previous studies have demonstrated that correct speed through tissue can be corrected, but may involve ionizing radiation or physical reference for these corrections [2,7]. The technique proposed will provide a new input of data, derived from the EM modality to approximate a more accurate SOS and result in an evaluation of accuracy of current clinically used systems.