Accurate measurement of the dynamic response of a scanning electronic portal imaging device

An important condition for the safe introduction of dynamic intensity modul ated radiotherapy (IMRT) using a multileaf collimator (MLC) is the ability to verify the leaf trajectories. Ln order to verify IMRT using an electroni c portal imaging device (EPID), the EPID response should be accurate and fa st. Noninstantaneous dynamic response causes motion blurring. The aim of th is study is to develop a measurement method to determine the magnitude of t he geometrical error as a result of motion blurring for imagers with scanni ng readout. The response of a liquid-filled ionization chamber EPID, as an example of a scanning imager, on a moving beam is compared with the respons e of a diode placed at the surface of the EPID. The signals are compared un der the assumption that all EPID rows measure the same dose rate when a str aight moving field edge is imaged. The measurements are performed at severa l levels of attenuation to investigate the influence of dose rate on the re sponse of the detector. The accuracy of the measurement method is better th an 0.25 mm. We found that the liquid-filled ionization chamber EPID does no t suffer from significant motion blurring under clinical circumstances. Usi ng a maximum gradient edge detector to determine the field edge in an image obtained by a liquid-filled ionization chamber EPID, errors smaller than I mm are found at a dose rate of 105 MU/min and a field edge speed of 1.1 cm /s. The errors reduce at higher dose rates. The presented method is capable of quantifying the geometrical errors in determining the position of the e dge of a moving field with subpixel accuracy. The errors in field edge posi tion determined by a liquid-filled ionization chamber EPID are negligible i n clinical practice. Consequently, these EPIDs are suitable for geometric I MRT verification, as far as dynamic response is concerned. (C) 2001 America n Association of Physicists in Medicine. [DOI: 10.1118/1.1350582].