SU-F-T-290: Modeling MLC Penumbra Within the Eclipse TPS - A Major Mode of IMRT QA Failure

Authors


Abstract

Purpose:

To illustrate how inaccurate modeling of the MLC penumbra within the Eclipse TPS can lead to dosimetric differences and decreased IMRT QA passing rates.

Methods:

Gafchromic film was used to compare 6 MV measured profiles from 2×2 cm2 MLC-defined fields to profiles calculated by the Eclipse TPS (v13.6) under two different scenarios. The first scenario involved calculations using the original MLC opening (original Eclipse). In the second, we selectively broadened the penumbra by summing two appropriately weighted fields: One with 95% of MUs delivered using the original MLC opening, and a second with the remaining MUs delivered with all MLCs opened an additional 5 mm (modified Eclipse). MATLAB code was written to generate MLC patterns and MU weighting in this manner and used to calculate modified Eclipse doses for 25 different clinical IMRT and VMAT plans. Patient specific QA measurements were made with film and compared to both original and modified Eclipse calculations.

Results:

For the 2×2 cm2 fields, agreement was observed between measured and original Eclipse profiles everywhere except in the penumbral region 5–10 mm from the end of the MLC where doses 10–30% higher than expected were measured. Near perfect agreement throughout the entire profile was observed when comparing measurement to the modified Eclipse calculation with the broadened penumbra. For the 25 clinical treatment plans, the average 2%/0 mm gamma passing rate using the original Eclipse calculations was 82.3%, while for the modified Eclipse calculations it improved to 90.1%.

Conclusion:

The approximation used by the Eclipse TPS to model the rounded end of the MLC is a failure mode that could lead to poor quality and/or failing QA for IMRT and VMAT plans. Improved modeling of the penumbra in the region 5–10 mm from the end of the MLC is required to completely correct this discrepancy.

Ancillary