SU‐E‐T‐269: Study IMRT Dosimetry Uncertainty Factors by Simulation

Abstract

Purpose: To analyze IMRT dosimetry uncertainty by simulation, and identify the key error from IMRT verification. Method: IMRT uncertainty factors are convoluted in the routine verification results, and can be from treatment plan, LINAC, and QA tools. With similar dosimetry coverage, several controlled IMRT plans can be generated by varying optimization parameters and algorithms. Some of them are more sensitive to certain perturbation and may fail to pass routine IMRT QA. First we modify the plan parameters to simulate the uncertainties. For example, we change MLC leaf and isocenter positions to simulate leaf and setup errors, and we change machine with different output factors to simulate small field measurement error. Then we import the plan back for dose calculation and comparison. Results: In the high dose region with low gradient, the mean dose is quite robust to positioning perturbation. In prostate case, the mean dose changes are within 2%, even with ±5mm random MLC leaf, or setup error. For HN cases, the mean dose changes are within 3% for most of perturbed plan with same uncertainties. However, large variations for the point dose from each beam are observed. The random MLC leaf error results the distribution broadened in the analysis of planar dose error distribution. As for the small field output factor error, it can cause a typical prostate IMRT case fail to pass standard QA criteria dependent on the plan. The small field error can behave randomly in the distribution of percentage dose error. In the low dose region with low gradient, simulation results indicated systematical MLC leaf position error is important, apart for MLC modeling parameters. Conclusion: IMRT verification is a part of on‐going efforts to improve the IMRT process. By identifying the key factors from QA data to optimizes the IMRT treatment.