Unencrypted login | home

Program Information

BEST IN PHYSICS (JOINT IMAGING-THERAPY): Prototype for An In-Room Supercompounded 3D Volumetric Ultrasound Image Guidance System for Prone Breast SBRT

T Chiu

T Chiu*, Y Zhang , B Hrycushko , B Zhao , R Chopra , S Jiang , X Gu , UT Southwestern Medical Center, Dallas, TX


WE-DE-BRA-2 (Wednesday, August 3, 2016) 10:15 AM - 12:15 PM Room: Ballroom A

Purpose: SBRT requires both substantial positioning precision and accuracy. However, accurate dose delivery is challenging for prone breast SBRT due to breast deformation, lumpectomy cavity shrinkage and ineffective CT/CBCT on soft tissue. We developed a prototype in-room 3D volumetric ultrasound (US) image guidance system for prone breast SBRT.

Methods: The designed system, set beneath a prone breast board, consists of a customized breast cup, rotational scanning water container, and a portable US scanner. The breast CT scan is converted to STereoLithographic model then 3D printed as a mold to construct a patient-specific breast cup. The holder is fabricated with liquid urethane resin and cured to a ~1 mm thick semi-rigid polymerized cup. The US transducer is attached to the container but separated from a water compartment by a urethane resin membrane. The transverse scan is acquired by simultaneously rotating the water container and transducer every 5°. The 3D US image is constructed with multiple transverse images using a supercompounding-technique-based volumetric US reconstruction algorithm. The performance of the prototype US system was evaluated on a gelatin based breast phantom with five embedded reflective markers. Reconstructed image quality (size and geometric accuracy) was compared with MRI images of the same phantom.

Results: In the evaluated plane, the five embedded markers were compared to MRI in terms of size and location. Diameter differences were 1.85, 1.71, 2.37, 1.00 and 1.85 mm between MRI and US. The 1-2 mm variations come from the different edge response the material has on the two imaging modalities. The difference in distance between adjacent marker centers between MRI and US were 0.57, 0.84, 0.75 and 0.80 mm.

Conclusion: The phantom study results show geometric agreement between MRI and US indicating that the proposed image-guidance system can provide precise and accurate patient setup for each SBRT fraction.

Funding Support, Disclosures, and Conflict of Interest: CPRIT Grant R1308

Contact Email: