Comparison of measured and calculated radiotherapy doses in the chest region of an inhomogeneous humanoid phantom.

Errors in dose calculation by treatment planning computers are known to arise when calculation algorithms do not account for electron disequilibrium near interfaces between tissues of different density. The accuracy of a treatment planning system (Plato, Nucletron International BV) was investigated for two treatments in the chest region: tangential 6 MV photons to the chest wall and opposed AP-PA 18 MV photon fields to the mediastinum. Thermo-luminescent dosimeters were used to measure dose at 40 sites in the chest of a humanoid phantom (Rando, Alderson Associates). Measurements were compared with point doses calculated using two different versions of the Plato external beam calculation software: RTS 1.8 and the newer RTS 2.2. Measured and calculated doses differed by 3% or more at more than one quarter of all sites. The greatest discrepancies occurred for points located in lung, which were generally overestimated. The maximum discrepancies for the 6 MV tangential breast irradiation were 8.5% for RTS 1.8 and 3.5% for RTS 2.2. For the 18 MV opposed field irradiation, the maximum discrepancies were 11.4% and 8.1% respectively. RTS 2.2 was more accurate than RTS 1.8, with smaller mean and maximum discrepancies.