SU‐E‐T‐05: The Effect of Leaf Corner Details in Sliding Windows Distributions, Implementation of Corrections in a Commercial Planning System

Abstract

Purpose: The dosimetric effect of leaf tip corner details in sliding window dose distributions has been studied, a solution to account for this corner leakage has been designed and implemented. Methods: The dosimetric effect of leaf tip corner deviations from perfect 90 degree borders between the front face and leaf side, was analyzed for sliding window IMRT beams. The study was carried out using film, EPID and diode arrays on Elekta Linacs fitted with MLCi and MLCi2 MLC models. Although the effect is general and does not depend on a particular MLC model, the magnitude and distribution is dependent on MLC model. The study consisted of simple fields designed to de‐couple the combined effects of the leaf properties (transmission, interleaf leakage, tongue and grove, corner leakage), such that corner leakage could be quantified. An enhancement to the transmission filter of a commercial Monte Carlo treatment planning system (Monaco, Elekta AB) was implemented to account for this effect, allowing the user to model the corner leakage for individual leaf tip corners. Results: The analysis of test beams indicates that corner leakage will introduce additional fluence in sliding window IMRT distributions. Of the two mlc models studied, this effect is more important for MLCi, but is also present for MLCi2. Generally, corner leakage will be present in every mlc in which the corner is not perfectly cut at 90 degrees. In this light, small manufacturing differences give rise to small but visible effects on the dose distribution. Correct modeling of these small effects is increasingly important as users seek improved patient specific QA results, in particular using local gamma. Conclusion: In this work we study the effect of mlc leaf corner leakage in sliding window distributions, presenting a software implementation to model this effect in a Monte Carlo‐based Treatment Planning system. This work was funded by Elekta Inc.