Analysis of delivery and recalculation of dose using DICOM treatment records

A procedure has been proposed for patient-specific QA that, instead of comparing a measurement to the planned dose, compares the dose calculated by an independent system to the dose calculated by the planning system. It is still prudent, however, to check the accuracy of the beam delivery. For this purpose, the DICOM records from the first treatment fraction can be compared to the planned treatment using an in-house developed planning system. Totally 1,398 patient portals were subjected to the new QA procedure. The dose distribution for each portal was first recalculated on a water phantom by two treatment planning systems and the dose distributions were compared. When agreement was observed, the patient was allowed to start treatment without a measurement. The record from the first day was imported into an in-house planning system which was used to evaluate the delivery for errors and calculate the delivered dose distribution and compare it to the planned dose distribution. A total of 266 portals passed a strict comparison between the clinical and QA dose calculations and directly used for treatment without measurements. For those portals, the comparison of the delivery records to the plan showed that 99% of spot positions deviated less than 0.2 ​mm and 99.7% of spot metersets deviated by less than 0.3%. On the other hand, 64 portals showed spot size deviations greater than the tolerance of ± 15% with some as large as ±25%. For 32 portals in which the record was used to calculate the delivered dose distribution, the Gamma passing rates between the planned and delivered distributions were always above 95% using a 2% dose difference and 2 ​mm distance-to-agreement criteria. The new QA process has been implemented slowly with strict constraints. The amount of beam time required has been reduced while maintaining safety.