Abstract
Octavius(®) -4D is a very effective device in radiotherapy treatment quality assurance (QA), due to its simple set-up and analysis package. However, even if it is widely used, its main characteristics and criticalities were only partially investigated. Taking start from its commissioning, the aim of this work was to study the main dependencies of the device response. The outcome dependence was studied comparing results by different delivery techniques [Intensity Modulated Radiation Therapy, IMRT (n = 29) and RapidArc, RA (n = 15)], anatomical regions [15 head/neck, 19 pelvis and 10 pancreas] and linear accelerators [DHX (n = 14) and Trilogy (n = 30)]. Moreover, the agreement dependency on the section of the phantom was assessed. Plan evaluations obtained by 2D, 3D, and volumetric γ-index (both local and global) were also compared. Generally, high dose gradient resulted critically managed by the assembly, with a smoother effect in RA technique. Worse agreements emerged in the 2D γ-index vs those of 3D and volumetric (P < 0.001), that were instead statistically comparable in global metric (P > 0.300). Volumetric plan evaluation was coherent with the average of passing rates on the 3 phantom axes (r ≥ 0.9), but transversal section provided best agreements vs sagittal and coronal ones (P < 0.050). The three studied districts furnished comparable results (P > 0.050) while the two LINACs provided different agreements (P < 0.005). The study pointed out that the phantom transversal section better fits the planned dose distribution, so this should be accounted when a two-dimensional evaluation is needed. Moreover, the major reliability of the 3D metric with respect to the 2D one, as it better agrees with the dosimetric evaluation on the whole volume, suggests that it should be preferred in a two-dimensional evaluation. Better agreements, obtained with RA vs IMRT technique, confirm that Octavius(®) -4D is specifically conceived for rotational delivery. Lastly, the assembly resulted sensitive to different technology.