Error Detection with a 3D Patient Specific QA System
R Rivest*, J Alpuche Aviles, K Nakonechny, J Beck, B McCurdy, CancerCare Manitoba, Winnipeg, MBSU-E-T-357 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall
Purpose: To determine if our in-house patient specific QA process, which utilizes the COMPASS (v2.1, IBA-Dosimetry) system, can detect delivery errors introduced in clinical IMRT plans.
Methods: The COMPASS system consists of a 2D chamber array mounted in the gantry, an independent gantry angle sensor and associated software. Patient specific QA at our clinic involves the creation of two QA structures, one consisting of all clinical PTVs, the other generated from the 20% isodose line. The gamma agreement index (GAI) within the two QA structures is used to compare 3D COMPASS reconstructed dose with planned dose distributions. Ten clinical DMLC-IMRT plans (5 prostate, 5 head & neck) were included in this study. Plans were modified by systematically shifting leaf pair positions so that that all nonzero leaf gaps were increased by 2.0 mm and 1.0 mm for the two sites, respectively. Verification plans were created for both the original and modified plans in order to quantify the dosimetric effect of the introduced errors via in-phantom ion chamber measurements. Finally, plans were delivered to COMPASS and 3D dose distributions were reconstructed on both the phantom and patient CT datasets. Reconstructed doses on the patient CT were analyzed as per our QA procedure.
Results: On average, measured ion chamber values for the modified plans deviated from the original planning doses by 3.1 ± 1.2% (Range: 1.0%-5.3%). For each of the six modified plans in which this metric exceeded 3%, GAI values for at least one of the QA structures were below 95%, which would be flagged by our QA process. Only one of the other modified and none of the original plans had GAI values <95%.
Conclusions: Gamma analysis of patient specific QA with COMPASS can effectively detect dosimetrically significant delivery errors introduced by systematically increasing leaf gaps in clinical IMRT plans.