Purpose: To test the accuracy of a modified formalism for electron beam reference dosimetry that adopts Monte Carlo (MC) calculated beam quality conversion factors, k(Q), that implicitly include gradient effects.
Methods: Four cylindrical (NE2571, PTW30013, IBA CC13, Exradin A1SL) and two parallel-plate (Scanditronix NACP-02, PTW Roos) ion chambers are used to calibrate electron beams from an Elekta Precise linac with 4, 8, 12, 18 and 22 MeV nominal energies. Different methods are used for the calibrations: (1) using the TG-51 formalism and data therein and cross-calibrating plane-parallel chambers against cylindrical chambers in high-energy electron beams, (2) using MC calculated k(Q) factors that include gradient effects by definition [data from Muir and Rogers, Med. Phys., 2014, p. 111701] with chambers calibrated directly in cobalt-60, and (3) using ion chambers directly calibrated against primary standards in linac electron beams, considered the gold-standard in this work for comparison of methods. The potential for using parallel-plate chambers directly in cobalt-60 rather than using cross-calibration is also investigated.
Results: When results for absorbed dose to water per monitor unit are obtained following TG-51, differences are between 0.5-1.3 % with a maximum difference of 2.7 % in the 4 MeV beam compared to results that use ion chambers calibrated directly against primary standards. When using MC calculated k(Q) factors and chambers calibrated directly in cobalt-60, differences are less than 0.75 %, on the same comparison basis, indicating that this approach is more accurate. The methods are internally consistent within 0.2-0.5 % when comparing results obtained using different chamber types.
Conclusion: MC calculated k(Q) factors and chambers calibrated directly in cobalt-60 provide more accurate linac calibration results than those obtained following the TG-51 formalism directly.