Measured and Monte Carlo calculated K(Q) factors: accuracy and comparison

The journal Medical Physics recently published two papers that determine beam quality conversion factors, k(Q), for large sets of ion chambers. In the first paper [McEwen Med. Phys. 37, 2179-2193 (2010)], k(Q) was determined experimentally, while the second paper [Muir and Rogers Med. Phys. 37, 5939-5950 (2010)] provides k(Q) factors calculated using Monte Carlo simulations. This work investigates a variety of additional consistency checks to verify the accuracy of the k(Q) factors determined in each publication and a comparison of the two data sets. Uncertainty introduced in calculated k(Q) factors by possible variation of W/e with beam energy is investigated further.

The validity of the experimental set of k(Q) factors relies on the accuracy of the NE2571 reference chamber measurements to which k(Q) factors for all other ion chambers are correlated. The stability of NE2571 absorbed dose to water calibration coefficients is determined and comparison to other experimental k(Q) factors is analyzed. Reliability of Monte Carlo calculated k(Q) factors is assessed through comparison to other publications that provide Monte Carlo calculations of k(Q) as well as an analysis of the sleeve effect, the effect of cavity length and self-consistencies between graphite-walled Farmer-chambers. Comparison between the two data sets is given in terms of the percent difference between the k(Q) factors presented in both publications.

Monitoring of the absorbed dose calibration coefficients for the NE2571 chambers over a period of more than 15 yrs exhibit consistency at a level better than 0.1%. Agreement of the NE2571 k(Q) factors with a quadratic fit to all other experimental data from standards labs for the same chamber is observed within 0.3%. Monte Carlo calculated k(Q) factors are in good agreement with most other Monte Carlo calculated k(Q) factors. Expected results are observed for the sleeve effect and the effect of cavity length on k(Q). The mean percent differences between experimental and Monte Carlo calculated k(Q) factors are -0.08, -0.07, and -0.23% for the Elekta 6, 10, and 25 MV nominal beam energies, respectively. An upper limit on the variation of W/e in photon beams from cobalt-60 to 25 MV is determined as 0.4% with 95% confidence. The combined uncertainty on Monte Carlo calculated k(Q) factors is reassessed and amounts to between 0.40 and 0.49% depending on the wall material of the chamber.

Excellent agreement (mean percent difference of only 0.13% for the entire data set) between experimental and calculated k(Q) factors is observed. For some chambers, k(Q) is measured for only one chamber of each type--the level of agreement observed in this study would suggest that for those chambers the measured k(Q) values are generally representative of the chamber type.