end ot December 2000 (each patient had made measurement once or two or three times). Semiconductor detectors (with OPD 510 by Scanditronix) were used during in-vivo dosimetry. Doses were calculated and measured in (1) the centre ot the irradiation tield; (2) supraclavicular region; (3) mediastinum; (4) lower edge ot the tield and (5) necko Patients were irradiated at various accelerators, most ot them at Neptun with photons 9 MeV.
Results: Ali patients were divided into three groups. The criterion of inclusion was the per cent difference between calculated and measured doses average for all dosimetrical points. The ranges for the groups were: 0-5%, 5-10% and over 10 %. The mean per cent differences in the first group of 43 patients was 3.1 %, in second of 27 patients -6.3%, and in third of 6 patients -17.6% respectively. There was no elear reason, beside an aeeidental error why for the eertain patient difference was much larger than for the another. Mean difference tor all groups was equal to 5.3%.
In the table mean per cent differences between doses calculated and measured and their standard deviations (SD) in the whole group ot patients are shown tor central axis, mediastinum and supraclavicular region. central axis mediastinum supraclavicular region Mean diff. SD Mean diff. SD Mean diff. SD 1.5% 4.1% -0.3% 4.6% 2.0% 5.7%
Conclusion: Mean difference in the whole group of patients showsgood agreement between pre-calculated and measured doses, especially for three clinically important regions (table ). It is accompanied by low standard deviation which is an indicator ot smali deviations between doses inside the whole group.