Impact of Multileaf Collimator Cleaning on Enhanced Leaf Model Configuration In Eclipse v18
Abstract
Purpose
Eclipse v18 introduces an enhanced leaf model (ELM) for multileaf collimator (MLC) configuration. The model requires dosimetric measurements acquired with MLC-shaped fields and is therefore sensitive to leaf condition. Although leaf cleaning is commonly recommended prior to ELM implementation, it demands substantial resources and lacks evidence of clinical gain, which has limited its clinical adoption. This study quantifies the impact of leaf cleaning on ELM configuration and IMRT plan quality.
Methods
Before leaf cleaning, open, closed, and sweep-gap fields were delivered on a TrueBeam linac equipped with high-definition MLCs. Dose was measured using a Farmer ionization chamber, with each measurement repeated five times and averaged for dosimetric leaf gap (DLG) modeling in ELM. Twenty clinical IMRT plans were optimized using the updated MLC model. PSQA was performed with portal dosimetry, and gamma pass rates were evaluated. The same measurements, modeling, and PSQA process were repeated after leaf cleaning. Differences in pre- and post-cleaning gamma pass rates were assessed using paired t-tests.
Results
After leaf cleaning, mean DLG values decreased by -0.61 mm across all energies. The largest reduction occurred for 10 MV (-0.99 mm), and the smallest for 6 MV FFF (-0.15 mm). Correspondingly, the mean gamma pass rate increased from 98.7% to 99.5% (± 1.3%) using 3%/2mm criteria (p < 0.001). Among all fields, 54.1% improved with a mean increase of +1.7% (± 1.1%), 29.5% showed a mean decrease of -0.4% (± 0.5%), and 16.4% were unchanged. Site-specific improvements were largest for head and neck (+2.0%, p = 0.01), followed by lung (+1.0%, p = 0.03).
Conclusion
MLC leaf cleaning resulted in reduced DLG values and increased gamma pass rates for PSQA, with notable site-specific gains. Our findings demonstrate a measurable clinical benefit to support the practice of leaf cleaning prior to ELM implementation.