Encrypted login | home

Program Information

Inaccurate Source to Electronic Portal Image Distance May Cause MLC Leaf Position Test Failure

H Chen

H Chen*, K Yang , K Huang , A Sarkar , J McGlade , D Jacob , F Mourtada , Christiana Care Hospital, Newark, DE


SU-I-GPD-T-477 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: Due to the heavy weight of EPID panel, we found the EPID Source-Image Distance (SID) changes with gantry angle. When inaccurate SID is used, it may falsely cause the MLC leaf position test failure.

Methods: A simple tool consists of some BB dots with known distances between them was built to verify the SID information. This tool was placed at isocenter plane and imaged at 0 degree gantry angle. By measuring the distances between BB dots on the acquired EPID images, we were able to determine the actual SID of EPID. Using the same method, the graticule tray supplied by the machine vendor was used to measure the magnification factors for four cardinal angles.

Results: Two Elekta machines (Synergy and Infinity) in our department were checked. The BB dots measured SID is 159.8 cm and 159.7 cm for the Synergy and Infinity respectively. The measurements with graticule tray at gantry 0 degree are in good agreement (< 1 mm) with the BB dots results. The graticule results show that the SID decreases with different amount at different gantry angle. The reduction was ~ 0.9 cm at 90 or 270 degree and ~ 1.6 cm at 180 degree gantry angle for both machines. For our MLC QA, we measure the leaf positions at 2.5, 7.5 and 12.5 cm from isocenter on both sides. Just considering the scaling effect, the 1.6 cm error in SID will introduce 0.2, 0.6 and 1.2 mm error at these positions, which explains why our MLC tests used to fail at 180 degree gantry angle.

Conclusion: EPID is an excellent tool for machine/patient QA, but it’s important to make sure the correct SID information is used to scale the images. We provided a simple and accurate method to measure SID at any gantry angle.

Contact Email: