Reproducibility and positioning sensitivity of CT beam width measurements using a pencil ionization chamber and...

To evaluate the reproducibility and precision of CT beam width measurements using a pencil ionization chamber and radiopaque mask and to assess the robustness of the technique against clinically realistic setup errors in the superior-inferior (SI) and anterior-posterior (AP) directions.

Beam width was measured on a GE Discovery RT590 CT scanner at three nominal collimations (10, 15, 20 mm), three tube potentials (80, 100, 120 kV), and three tube currents (50, 100, 150 mA). Three repeated exposures per setting were acquired using a Radcal Accu-Gold pencil ionization chamber with a radiopaque mask to calculate beam width. Coefficients of variation and Levene's test were used to test precision. To assess the robustness to setup error, the chamber was offset from the isocenter at 0, 1, 3, and 5 mm in the SI direction at 20-mm collimation across all kV/mA combinations and at 0 and 5 mm in the AP direction at 10- and 20-mm collimation. A linear mixed-effects model was used to test the sensitivity of beam width to SI setup errors across kV/mA combinations. Slopes of the overall deviance from the global mean vs offset were obtained for each combination with 95% confidence intervals, and pairwise slope differences were evaluated using a Tukey-adjusted statistical test. Confidence intervals were used to assess AP setup errors at two different nominal collimations.

Across all kV/mA combinations, mean measured radiation beam width was 13.04, 16.90, and 20.60 mm for nominal 10-, 15-, and 20-mm collimations, with inter-protocol coefficients of variation of 0.57%, 0.48%, and 0.25%, respectively. Intra-protocol standard deviations across triplicate exposures were 0.04 to 0.09 mm. SI offset slopes from 0 - 3 mm ranged from -0.07 to -0.12 mm (per mm of offset) with overlapping 95% confidence intervals. There was no significant interaction with kV/mA (p > 0.05). AP offsets of 5 mm produced beam-width changes of 0.09 mm at 10 mm collimation and 0.15 mm at 20 mm collimation.

The pencil ionization chamber and radiopaque mask technique produced highly reproducible CT beam width estimates for potentials up to 120 kV and at or above 50 mA that were independent of tube potential, tube current, and robust to clinically plausible setup inconsistencies.