Unencrypted login | home

Program Information

Assessment of Phantom Positioning Accuracy in IMRT Quality Assurance: Insert Design and Implementation


J Ren

J Ren*, A McNiven, D Letourneau, Princess Margaret Hospital, Toronto, ON

SU-E-T-161 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

Purpose:
A multi-centre program for intensity modulated radiation therapy (IMRT) quality assurance (QA) has previously been established. In this work, a phantom insert has been designed, in conjunction with an image analysis algorithm, to determine phantom positioning accuracy as part of program development.

Methods:
An insert was designed for the ArcCheck phantom that includes four BBs (located outside the sensitive area of the phantom) that are visualized using portal images (PI). A procedure has been developed to acquire PI at 8 gantry angles (2 field sizes per angle) for off-line analysis. The image analysis method includes three steps: 1) cross correlation analysis is used to automatically detect the BB position with respect to the radiation isocentre in each PI; 2) 2D BB coordinates are reconstructed to 3D using the projected images; 3) the phantom positioning accuracy is obtained by a 3D rigid body transformation analysis of the BB coordinates. To evaluate the reliability of the image analysis method, the PI procedure was completed with the phantom aligned to the lasers (reference position) and with known shifts in phantom position (translational and rotational).

Results:
The measured change in phantom position, was compared to the known phantom shifts. Good agreement was found between the measured shift and known shift (within 0.2 mm or 0.2°). For the most complicated shift applied (5° yaw, 3 mm lateral translation and 4 mm vertical translation) the agreement was within 0.1° and 0.1 mm for rotations and translations respectively.

Conclusion:
A phantom insert has been designed to determine the accuracy of phantom positioning with respect to the radiation isocenter. Initial work illustrates that the analysis method can accurately characterize known positioning errors, which will help to isolate the impact of phantom misalignment in the context of IMRT QA.

Funding Support, Disclosures, and Conflict of Interest: Supported in part by Cancer Care Ontario

Contact Email: