Linac reported steering error insensitive to 6 MV FFF transverse beam position deviations

Clinically significant beam position deviations were observed for a 6 MV FFF beam during patient specific QA on an Elekta linear accelerator. There was no significant reported transverse steering error from the machine ion chamber, and routine linac QA practices, including cardinal angle Winston-Lutz test, did not identify the deviations. Subsequent investigation using an electronic portal imaging device (EPID) revealed clinically significant beam position deviations for small steering errors. This prompted investigation into further impact and possible solutions. Testing set-points were established by adjusting transverse steering current to achieve introduced 2 T steering errors. Tests at each set-point included scanning water tank profiles and EPID images. A proposed method for adjusting the 2 T error sensitivity was tested via adjusting the 2 T loop parameter such that a reported error corresponds to specific beam position deviations. The testing set-points resulted in positional deviations of greater than 3 mm for reported errors of less than 1. A new method for improving 2 T error sensitivity was implemented. This work has shown that existing vendor protocol for establishing beam steering error for 6 MV FFF beams can lead to beam position deviations without machine interlocks or significant reported steering errors. Thus, an alternative method of establishing steering error sensitivity based on positional deviations is presented.