A modified formalism for electron beam reference dosimetry to improve the accuracy of linac output calibration

PURPOSE: To present and demonstrate the accuracy of a modified formalism for electron beam reference dosimetry using updated Monte Carlo calculated beam quality conversion factors. METHODS: The proposed, simplified formalism allows the use of cylindrical ionization chambers in all electron beams (even those with low beam energies) and does not require a measured gradient correction factor. Data from a previous publication are used for beam quality conversion factors. The formalism is tested and compared to the present formalism in the AAPM TG-51 protocol with measurements made in Elekta Precise electron beams with energies between 4 MeV and 22 MeV and with fields shaped with a 10 x 10 cm(2) clinical applicator as well as a 20 x 20 cm(2) clinical applicator for the 18 MeV and 22 MeV beams. A set of six ionization chambers are used for measurements (two cylindical reference-class chambers, two scanning-type chambers and two parallel-plate chambers). Dose per monitor unit is derived using the data and formalism provided in the TG-51 protocol and with the proposed formalism and data and compared to that obtained using ionization chambers calibrated directly against primary standards for absorbed dose in electron beams. RESULTS: The standard deviation of results using different chambers when TG-51 is followed strictly is on the order of 0.4% when parallel-plate chambers are cross-calibrated against cylindrical chambers. However, if parallel-plate chambers are directly calibrated in a cobalt-60 beam, the difference between results for these chambers is up to 2.2%. Using the proposed formalism and either directly calibrated or cross-calibrated parallel-plate chambers gives a standard deviation using different chambers of 0.4%. The difference between results that use TG-51 and the primary standard measurements are on the order of 0.6% with a maximum difference in the 4 MeV beam of 2.8%. Comparing the results obtained with the proposed formalism and the primary standard measurements are on the order of 0.4% with a maximum difference of 1.0% in the 4 MeV beam. CONCLUSIONS: The proposed formalism and the use of updated data for beam quality conversion factors improves the consistency of results obtained with different chamber types and improves the accuracy of reference dosimetry measurements. Moreover, it is simpler than the present formalism and will be straightforward to implement clinically.