Encrypted login | home

Program Information

A Feasibility Study of Quick MLC Position and Verification Using a Custom MATLAB GUI

no image available
N Thorne

N Thorne*, B Morris , C Yang , University of Mississippi Medical Center, Jackson, MS


SU-I-GPD-T-466 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To develop an efficient MLC position and collimator angle verification QA using a custom MATLAB GUI.

Methods: The Hancock MLC position test is a procedure to identify the leaf position of an MLC system. Midline of a field is determined from combining two EPID images of top half field with one leaf pair open beyond the isocenter taken at 0º collimator and bottom half field with one leaf pair open beyond the isocenter at 180º collimator. Individual leaf positions may then be measured from this determined isocenter and the pixel readings of the EPID panel. Based on the original Hancock MLC position test idea, a customized Matlab code was developed. The code detects the Full Width at Half Maximum for each pixel row along each MLC bank and uses an edge detection algorithm to identify the position of each leaf. The code was tested by running it before and after Elekta’s MLC position verification software, Autocal, to assess position of the MLC’s. The system was also able to extend beyond the traditional Hancock MLC test to assess misalignment in collimator angle. The degree of rotation was calculated by fitting the tangent of the rotation angle as the slope of the drift in each bank horizontal position over the vertical relative to the pixel orientations. A standalone Matlab GUI was also constructed allowing easy implementation.

Results: The GUI worked reasonably well to assess MLC position and can detect changes as small as 0.5 mm in MLC position change. Collimator angle difference as small as 0.1 degrees can be detected with this measurement.

Conclusion: The method developed may serve as a powerful verification tool on Autocal software used in evaluation of MLC leaf position as well as a weekly QA test for MLC performance.

Contact Email: