A method to check the accuracy of dose computation using quality index: application to scatter contribution in high energy photon beams.

Computerized dose calculation verification is a relevant component of radiotherapy treatment planning quality assurance. The usual procedure is to compare measurements to computations for several standard situations. As cases become more complex, special test phantoms and beam arrangements must be used, and an experimental procedure must be carefully established. In this paper we follow a new methodology to prepare a set of reference data that may be used to verify the accuracy of dose calculations involving changes in the scatter component of photon beams. The advantage of this methodology is that local measurements are not required. A quantitative evaluation of dose modifications was performed by means of correction factors (CF). For this purpose, three geometrical configurations were designed (asymmetric, symmetric, and reference) where the primary component was kept constant and the scatter component was varied by changing the height (h) of lateral columns. Measurements were performed in polystyrene phantoms for seven photon beam energies. CF were derived as the ratio of the absolute dose measured at the point of interest to the absolute dose for the reference configuration, for the asymmetric and symmetric configurations, respectively. They were expressed as a function of beam quality (QI). We have verified that, for all configurations studied, CF decrease with QI. For h = 15 cm, CF remain practically constant, whatever machine technology is used [the mean values of CF for the asymmetric and symmetric cases are CFa= 1.028 (0.2% 1 s.d.) and CFs= 1.058 (0.4% 1 s.d.)]. We have developed a test protocol and we have chosen those configurations corresponding to h = 15 cm because they both present greater values of the CF and lower standard deviations. The direct application of the method is straightforward. The user can reproduce on his local TPS the three experimental configurations described in the test protocol, and then compute CF which can be compared to our reference data set for any beam quality.