The purpose of this study was to determine the dosimetric accuracy of
the treatment of parotid gland tumours using 8 MV X-ray beams. These t
umours are generally situated near the patient's skin. Entrance in viv
o dose measurements with diodes were obtained for 20 patients during 5
sessions per patient, in the anterior-oblique and posterior-oblique w
edged fields, on the central beam axis as well as in points situated i
n a cranial plane 2 or 3 cm off-axis. Phantom measurements were perfor
med in order to determine the actual position of the 95% isodose surfa
ce. The measurements were compared with calculations performed with ou
r three-dimensional treatment planning system. The reproducibility of
the diode measurement of patients was found to be 1.4% (1 SD). The tot
al accuracy in the entrance dose determination for the average of 2 me
asurements was 1.8% (1 SD). The central axis entrance dose for the ant
erior field was on average 1.5% +/- 3.2% (1 SD) higher than the calcul
ated value. For the posterior field, the difference was 0.9% +/- 3.1%
(1 SD). The deviations for the off-axis points were somewhat smaller,
mainly due to overestimation of the block effect in the calculations.
The value of the dose at the isocentre, obtained by extrapolation of t
he measured entrance dose values, differed 0.3% +/- 2.1% from the calc
ulations. The accuracy in dose determination at the isocentre was 2% (
1 SD). After correction for the difference in prescribed and actual so
urce-to-skin distance, the results showed good agreement with phantom
measurements on a polystyrene phantom without inhomogeneities, perform
ed both with diodes and an ionization chamber. The measured 95% isodos
e surface was situated 10.6 mm under the phantom surface, which meant
a shift of 0.5 mm towards the surface compared with calculations. A 1-
15 mm shift towards the block edges was observed. These results formed
the basis of a disometric quality assurance protocol. The entrance do
se on the central beam axis is measured twice at the beginning of the
treatment. The total dose at the isocentre is calculated using the per
centage depth dose values. If the difference between the average measu
red and calculated dose at the isocentre is larger than 4% (2 SD of th
e measurement accuracy at the isocentre), the reason for the discrepan
cy will be investigated and, if necessary, the number of monitor units
will be adjusted.