Purpose:To assess the sensitivity of gamma index and other IMRT QA metrics to couch and collimator rotations.Methods:Two brain IMRT plans with couch and/or collimator rotations in one or more of the fields were evaluated using the IBA MatriXX ion chamber array and its associated software (OmniPro‐I'mRT). The plans were subjected to routine QA by 1) Creating a composite planar dose in the treatment planning system (TPS) with the couch/collimator rotations and 2) Creating the planar dose after “zeroing” the rotations. Plan deliveries to MatriXX were performed with all rotations set to zero on a Varian 21ex linear accelerator. This in effect created TPS‐created planar doses with an induced rotation error. Point dose measurements for the delivered plans were also performed in a solid water phantom.Results:The IMRT QA of the plans with couch and collimator rotations showed clear discrepancies in the planar dose and 2D dose profile overlays. The gamma analysis, however, did pass with the criteria of 3%/3mm (for 95% of the points), albeit with a lower percentage pass rate, when one or two of the fields had a rotation. Similar results were obtained with tighter criteria of 2%/2mm. Other QA metrics such as percentage difference or distance‐to‐agreement (DTA) histograms produced similar results. The point dose measurements did not obviously indicate the error due to location of dose measurement (on the central axis) and the size of the ion chamber used (0.6 cc).Conclusion:Relying on Gamma analysis, percentage difference, or DTA to determine the passing of an IMRT QA may miss critical errors in the plan delivery due to couch/collimator rotations. A combination of analyses for composite QA plans, or per‐beam analysis, would detect these errors.
Read full abstract