Dosimetric Validation of an Enhanced Leaf Model In Single‑Isocenter Multi‑Target Stereotactic Radiosurgery
Abstract
Purpose
To evaluate the accuracy of the Enhanced Leaf Model (ELM) implemented in Eclipse v18.0 for single-isocenter multiple-target (SIMT) stereotactic radiosurgery (SRS), with emphasis on updated MLC modeling and dosimetric precision.
Methods
A clinical ELM was implemented in Eclipse v18.0 for TrueBeam linac equipped with Millennium 120 MLC. Model validation was performed using several MLC test plans, including a 3 mm narrow MLC slit (nMLC) and dynamic MLC (dMLC) field with 3 mm sweeping gaps. GafChromic EBT4 film was used to compare the profiles with the MLC plans. Several SIMT plans originally calculated with a dosimetric leaf gap (DLG)-based MLC model in Eclipse v15.6 were recalculated using the ELM in v18.0. Dose calculations were performed with AcurosXB for 6MV FFF. Patient-specific QA was performed with a myQA SRS device.
Results
For 6MV FFF, MLC plan profiles calculated with the ELM were comparable to those with the DLG-based model. Overall agreement between calculated and measured MLC profiles was improved with ELM, presenting well matched low-dose penumbra region in the nMLC profile and shoulder region in dMLC profile, particularly in regions sensitive to MLC motion and leaf transmission associated with the updated leaf-tip modeling in ELM. In SIMT plans recalculated with the ELM, small reductions up to -4% in target coverage were observed for selected off-axis targets, which resulted from coverage shifts of up to approximately 0.2 mm, while overall dose differences remained minimal.
Conclusion
ELM provides improved dose modeling for MLC-defined fields. For very small targets, particularly located at off-axis, the combined impact of geometric uncertainties and subtle changes in MLC shoulder modeling may influence target coverage. These findings provide practical guidance for ELM implementation and highlight the importance of defining target size limits and selecting appropriate margins during SIMT SRS planning when using conventional DLG-based MLC model.