Comparison of 4 CBCT-IGRT geometric QA phantoms within a single institutional workflow

To compare 4 commercial cone-beam CT (CBCT) geometric quality assurance phantoms within a common workflow, focusing on positioning/repositioning accuracy and agreement between the laser-defined setup reference and the CBCT imaging isocenter.

MIMI, QUASAR Penta-Guide, ISOCube, and MZ CUBE were evaluated monthly for 5 months on a Versa HD system with XVI CBCT and a HexaPOD couch. Two observers independently completed each measurement. Three-dimensional post-correction translational residuals (Δdpos/rep) and laser-defined setup reference-CBCT imaging isocenter distances (Δdiso) were calculated. Phantom comparisons used observer-averaged values from the 5 monthly sessions. Rotational correction and CBCT image quality were assessed descriptively with MZ CUBE.

Mean Δdpos/rep values ranged from 0.19 to 0.31 mm, with mean absolute observer differences of 0.05-0.12 mm. No overall difference was found among the 4 phantoms (P = .547). Mean Δdiso values ranged from 0.68 to 0.87 mm, and all measurements were below 2 mm. An overall difference was observed for Δdiso (P = .019), although no pairwise comparisons were performed. With MZ CUBE, mean post-correction translational and rotational residuals were 0.24 mm and 0.37°, respectively, and image-quality measures showed no consistent temporal trend.

The 4 phantoms showed similarly small post-correction translational residuals. In our setting, MZ CUBE additionally supported rotational and image-quality checks.

Direct comparison within a common workflow showed similar translational performance, while the scope of additional testing depended on the phantom configuration and locally available accessories.