An assessment of the efficiency of methods for measurement of the computed tomography dose index (CTDI) for cone beam (CBCT) dosimetry by Monte Carlo simulation.

The IEC has introduced a practical approach to overcome shortcomings of the CTDI100 for measurements on wide beams employed for cone beam (CBCT) scans. This study evaluated the efficiency of this approach (CTDIIEC) for different arrangements using Monte Carlo simulation techniques, and compared CTDIIEC to the efficiency of CTDI100 for CBCT. Monte Carlo EGSnrc/BEAMnrc and EGSnrc/DOSXYZnrc codes were used to simulate the kV imaging system mounted on a Varian TrueBeam linear accelerator. The Monte Carlo model was benchmarked against experimental measurements and good agreement shown. Standard PMMA head and body phantoms with lengths 150, 600, and 900 mm were simulated. Beam widths studied ranged from 20-300 mm, and four scanning protocols using two acquisition modes were utilized. The efficiency values were calculated at the centre (εc) and periphery (εp) of the phantoms and for the weighted CTDI (εw). The efficiency values for CTDI100 were approximately constant for beam widths 20-40 mm, where εc(CTDI100), εp(CTDI100), and εw(CTDI100) were 74.7  ±  0.6%, 84.6  ±  0.3%, and 80.9  ±  0.4%, for the head phantom and 59.7  ±  0.3%, 82.1  ±  0.3%, and 74.9  ±  0.3%, for the body phantom, respectively. When beam width increased beyond 40 mm, ε(CTDI100) values fell steadily reaching ~30% at a beam width of 300 mm. In contrast, the efficiency of the CTDIIEC was approximately constant over all beam widths, demonstrating its suitability for assessment of CBCT. εc(CTDIIEC), εp(CTDIIEC), and εw(CTDIIEC) were 76.1  ±  0.9%, 85.9  ±  1.0%, and 82.2  ±  0.9% for the head phantom and 60.6  ±  0.7%, 82.8  ±  0.8%, and 75.8  ±  0.7%, for the body phantom, respectively, within 2% of ε(CTDI100) values for narrower beam widths. CTDI100,w and CTDIIEC,w underestimate CTDI∞,w by ~55% and ~18% for the head phantom and by ~56% and ~24% for the body phantom, respectively, using a clinical beam width 198 mm. The CTDIIEC approach addresses the dependency of efficiency on beam width successfully and correction factors have been derived to allow calculation of CTDI∞.