P296] Use of statistical interference for tolerance determination in dose measurements

Abstract

Purpose The most important dosimetry quantity that is determined at radiotherapy centres is absorbed dose to water. Fixed tolerances for absorbed doses under reference conditions measured with ionisation chamber for photon resp. electron beams are usually 2% resp. 3%. The aim was to develop a new approach to the evaluation of agreement between measured and reported values based on statistical interference rather than to use fixed tolerance levels. Methods The approach takes partial uncertainties of compared doses into consideration. Three statistical models were proposed describing three different situations to assess the critical values for rejecting the zero hypothesis on agreement: 1) Two measurements performed in time with the same dosimetry chain, 2) Two measurements performed with two different dosimetry chains which were calibrated at the same calibration laboratory, 3) Two measurements performed with two different dosimetry chains at two different calibration laboratories (dose calculations are completely independent). For electron dosimetry, the uncertainty of PDD was comprehensively evaluated and its influence on recalculated dose values from reference depth to the depth of maximum dose was taken into account. In the contribution, uncertainties from TRS 398 were taken as an example. Results Critical values for photon and electron beams case 1 resp. 2 were 2.5% resp. 3.1% (k = 2). When uncertainty of 0.5% (k = 1) for electron PDD at zref was taken into account, critical values for electron beams case 1 resp. 2 were 2.9% resp. 3.4% (k = 2). The sensitivity of critical values to actual partial uncertainties will be shown. Conclusions The models can be adopted to implement statistical interference into the decision on agreement for all quantities, where partial uncertainties are known. This work was supported by the Ministry of Industry and Trade of the Czech Republic, project No. FV20411.