SU-E-T-104: Utilizing Statistical Process Control and EPID for Routine QA of Medical Linear Accelerators

Abstract

Purpose: To investigate the suitability of an aSi EPID as a dosimetric quality assurance device. Action thresholds are obtained using statistical process control (SPC) and introduced known errors detected by the EPID are related to measurements with other quality assurance devices. Methods: Measurements were performed using a Varian Clinac 2300iX linear accelerator (Varian, Inc., Palo Alto, CA, USA), equipped with an aSi EPID (Varian aSi1000). The suitability of the EPID as a device for daily monitoring of the beam output was investigated for 6MV and 15MV by comparing measurements with the EPID and PTW Linacheck (PTW GmBh, Freiburg, Germany). The ability of an aSi EPID to verify the beam quality and beam profile parameters was investigated by introducing known errors. A MATLAB© script was developed for analysis of the acquired EPIDimages. Using statistical process control (SPC), the data was validated and action thresholds were obtained. Results: The daily measurements with the EPID resulted in good agreement with the PTW Linacheck in detecting output deviations. Obtaining actions thresholds with the use of SPC analysis increased the probability to detect systematic errors in comparison to the use of the mean and standard deviation as primary tools to evaluate the data. Introduced known errors in beam quality and profile parameters could be readily detected by the aSi EPID and was traceable to water measurements and data collected during commissioning of the linac. For example, a deviation of 1.5% in the TPR measured in water corresponded to a 0.9% change in hump measured with EPID. Conclusions: An aSi EPID has proven to be a useful device for daily constancy control of linac output. With the use of SPC one can readily detect systematic errors. Simultaneously it can be utilized as a device to control the constancy of beam quality and beam profile parameters.