Abstract
BACKGROUND: The Elekta Unity MR-Linac system integrates magnetic resonance imaging (MRI) with a linear accelerator (Linac) for adaptive radiation therapy. Traditional quality assurance (QA) methods for multi-leaf collimators (MLCs) face challenges in this system due to the magnetic field and limited field size of electronic portal imaging devices (EPID). PURPOSE: This study aims to develop a 'virtual picket fence' test using machine log files to evaluate MLC performance in the Elekta Unity MR-Linac system, providing a more efficient and comprehensive QA method that overcomes the limitations of traditional approaches. METHODS: A picket fence test plan with 11 segments was delivered on the Elekta Unity system. Maximum absolute error and root mean square (RMS) error for each leaf were calculated by comparing log file data with nominal values. A deliberate 1 mm error was introduced in selected MLCs to test the sensitivity of the virtual test. The results from the log file were further compared with measurements from radiochromic films. RESULTS: The maximum deviation between log file data and nominal values was within 1 mm for all leaves. The virtual picket fence test successfully identified MLCs with deviations beyond the 0.5 mm warning threshold in the error-introduced test. Comparisons with film-based measurements showed good agreement, with deviations between film and log file data also within 1 mm. CONCLUSIONS: The virtual picket fence test provides an efficient and comprehensive method for MLC QA in the Elekta Unity MR-Linac system. This method can be integrated into weekly QA workflows alongside traditional film-based methods for thorough quality control.