TU‐C‐BRA‐01: Progress in Calculations of KQ for TG‐51

The calculated values of k Q in the TG‐51 protocol are based on analytic calculations which make use of various tabulated values of measured and calculated factors such as the water to air stopping‐power ratio, P wall , P repl and P cel (for details see Ch 9 of the 2009 AAPM Summer School book, Med. Phys. Publishing, Madison, WI). Since the publication of TG‐51 there have been improved calculations of almost all the factors required in TG‐51. This talk will briefly review those improved calculations which are based on Monte Carlo calculations of ratios of the dose to the cavity of the ion chamber. However, with the development of the EGSnrc Monte Carlo system (Kawrakow, Med Phys 27(2000) 499) it became possible to accurately calculate the response of ion chambers, not just the ratios of the dose to the cavity. This made it possible to accurately do ab initio Monte Carlo calculations of k Q . As computers became faster per unit cost and clever variance reduction techniques for doing ion chamber calculations in complex geometries were implemented (Wulff et al, Med Phys 35(2008)1328) these calculations became both accurate and feasible (albeit with large clusters of computers). Values of k Q have been calculated this way for a total of 33 commonly used cylindrical ion chambers (Muir and Rogers, Med Phys 37(2010)5939). Detailed estimates of the systematic uncertainty in these calculations have been made and range between 0.6% and 1.0%, depending on the assumptions made. The largest component is the uncertainty (0.5%) in the assumed constancy with beam quality of (W/e) air , which relates the energy deposited in the cavity to the charge released in the air. In a detailed comparison of the calculated k Q values to the extensive high‐ quality measurements by McEwen (Med Phys 37(2010)2179), Muir et al (submitted, 2011) found the mean percentage differences between the calculations and the experiments are 0.08(0.17), 0.07(0.32) and 0.23(0.31) in 6, 10 and 25 MV beams respectively (bracketed values are rms deviations). These discrepancies are well within the stated uncertainties of the measurements (about 0.3% to 0.4%) and the calculations (about 0.3% to 0.4% ignoring W/e uncertainties and assuming correlated uncertainties in photon cross section). These comparisons can be used to set an upper limit of 0.4% on the variation of (W/e) air with beam quality between 60 Co and 25 MV beams (95% confidence). More importantly, the close agreement with experiment gives confidence in the accuracy of the Monte Carlo calculated values of k Q with a 68% confidence uncertainty of between 0.4% and 0.5%. Learning Objectives 1. Understand the basis of calculated k Q values in the original TG‐51 protocol 2. Become aware of the improvements made in many of the required correction factors in the last decade 3. Understand how ab initio calculations of k Q are done 4. Understand the uncertainties involved in ab initio calculations of k Q