Measurements of characteristics of time pattern in dose delivery in step-and-shoot IMRT.

Background And Purpose: Although intensity-modulated radiotherapy (IMRT) has already shown its clinical benefit, there are some issues which are not yet fully understood. Among these is the question whether the protracted dose delivery due to the lowered dose rate has any radiobiological consequences. To investigate this question, an exact characterization of dose rate profiles in typical clinical plans is needed. Furthermore, such a characterization may lead to an increased knowledge how to improve IMRT technically.

Material And Methods: A new IMRT phantom which allows precise measurement of up to nine points of interest simultaneously with pin-point ionization chambers was developed. To examine dose rates, a new software tool (GRAYHOUND) was developed which can measure doses in short time intervals of up to 0.5 s. 250 points in four clinical IMRT plans were examined. A set of parameters was defined to describe the dose rate profiles including the effective fraction time (eft, which is the percentage of the fraction time in which any dose is delivered to a specific point), and a quotient of the percentage of dose delivered in high dose pulses (> 0.01 Gy/s) divided by the percentage of fraction time needed to deliver this dose (d(HD)/t(HD)).

Results: These quotients are excellent markers for the inhomogeneity of dose rate delivery in IMRT. In both parameters a wide variance in points of the same plan and between different plans was found. For example, eft ranged between 11.6% and 37.3% in high dose points and the time in which high dose rates are delivered to a single high dose point ranged between 3.6% and 10.1% of total fraction time.

Conclusions: These data show a great inhomogeneity of dose rates not only between different plans but also between different points in the same plan. Biological investigations are needed to quantify the relevance of these inhomogeneities. The parameters which are introduced in this work may be suitable to compare different optimization algorithms in IMRT.