Encrypted login | home

Program Information

Dose Distribution Differences Caused by System Related Geometric Distortion in MRI-Guided Radiation Treatment System

J Wang

J Wang1*, J Yang2 , S Marshall3 , Z Wen4 , L Court5 , G Ibbott6 , (1) MD Anderson Cancer Center, Houston, TX, (2) MD Anderson Cancer Center, Houston, TX, (3) Monaco, Elekta AB, Tampa, FL, (4) MD Anderson Cancer Center, Houston, TX, (5) UT MD Anderson Cancer Center, Houston, TX, (6) UT MD Anderson Cancer Center, Houston, TX


SU-E-J-205 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: MRI has superb soft tissue contrast but is also known for geometric distortions. The concerns and uncertainty about MRI’s geometric distortion have contributed to the hesitation of using only MRI for simulation in radiation therapy. There are two major categories of geometric distortion in MRI; system related and patient related. In this presentation, we studied the impact of system-related geometric distortion on dose distribution in a digital body phantom under an MR-Linac environment.

Methods: Residual geometric distortion (after built-in geometric correction) was modeled based on phantom measurements of the system-related geometric distortions of a MRI scanner of a combined MR guided Radiation Therapy (MRgRT) system. A digital oval shaped phantom (40x25 cm) as well as one ellipsoid shaped tumor volume was created to simulate a simplified human body. The simulated tumor volume was positioned at several locations between the isocenter and the body surface. CT numbers in HUs that approximate soft tissue and tumor were assigned to the respective regions in the digital phantom. To study the effect of geometric distortion caused by system imperfections, an IMRT plan was optimized with the distorted image set with the B field. Dose distributions were re-calculated on the undistorted image set with the B field (as in MR-Linac).

Results: The maximum discrepancies in both body contour and tumor boundary was less than 2 mm, which leads to small dose distribution change. For the target in the center, coverage was reduced from 98.8% (with distortion) to 98.2%; for the other peripheral target coverage was reduced from 98.4% to 95.9%.

Conclusion: System related geometric distortions over the 40x25 area were within 2mm and the resulted dosimetric effects were minor for the two tumor locations in the phantom. Patient study will be needed for further investigation.

Funding Support, Disclosures, and Conflict of Interest: The authors received a corporate research grant from Elekta

Contact Email: