Ionization chamber response in intensity-modulated radiotherapy beams

The response of ionization chambers in IMRT fields has been investigated. Differences between measured and calculated values of average chamber dose associated with chamber position, volume averaging and low monitor unit effects were assessed with regards to patient specific IMRT quality assurance (QA). Calculations were performed with a commercially available treatment planning system (TPS).; Inaccuracy of chamber positioning during QA was shown to adversely affect measurements due to steep dose gradients in the vicinity of penumbras generated by MLC leaf-tips. The measurement error was inversely related to chamber size. The Exradin A12 chamber showed errors of 4.1% and 9.6% in its short (0.61 cm) and long dimensions (2.2 cm) while the Exradin T14 Microchamber showed an error of 15.4% in its useful dimension (0.1 cm). A QA phantom was designed to improve positional accuracy and reproducibility. Setup reproducibility results showed a standard deviation of 0.93 and 1.1 mm in the longitudinal and lateral directions through the use of this phantom.; The effect of leaf shadowing of a measurement chamber during IMRT QA was correlated to the error in point dose results. The error between calculation and measurement increased with a higher degree of chamber shadowing, which indicated that point dose results could be improved by placing the chamber in areas where the effect of shadowing was minimized.; A linear systems approach was utilized to determine the chamber response function of three chambers for subsequent implementation in a TPS. Improvement in dose calculation accuracy was expected. The response functions were used in a TPS to design special chambers that had the same averaging properties as exhibited during water phantom measurements. Results were improved over using a chamber represented by a perfect cylinder, but spatial resolution limitations prevented optimal results. Another chamber design that was based on a heuristic approach yielded desired results with static beam penumbras but was not optimal when used with IMRT fields.; Low numbers of monitor units delivered during each subfield of an IMRT treatment can force the linac control system to skip segments which introduces dosimetric error. For individual IMRT fields this effect accounted for an error of up to ±4%. It was found that most of this error possesses a degree of randomness that is expected to cancel over the delivery of all fields in a clinical situation.