Assessment of set up accuracy using an evacuated cushion for proton beam therapy

Abstract

Purpose or Objective Effective, reproducible patient positioning is paramount for radiotherapy. With the introduction of a new proton beam therapy (PBT) service in January 2019, evacuated cushions (vacbags) have been implemented for patients receiving treatment to the limbs, abdomen, thorax and pelvis. This work reports on an audit reviewing set up times, image review times and positional reproducibility to confirm the suitability of the vacbag as an immobilisation tool in PBT. Material and Methods The first ten patients treated on vacbags were included in this study. Patients were initially imaged with kilovoltage (kV) 2-dimensional (2D) image pairs as a gross error check (GEC), followed by a conebeam CT (CBCT) from which positional errors were corrected and verified with a final 2DkV image before proceeding to treatment. Images were compared to the planning CT to determine translational and rotational displacements. Treatment and image review times were also reported to audit the new service. Results Compared to the planning CT, translational displacements were satisfactory with 95% <5mm and 100% <10mm. Greater variation was seen in rotational displacements, however, on average only 3% of all exceeded the 2⁰ tolerance. Agreement between 2D and 3D imaging were within the institutions defined GEC tolerance of <1cm and <2⁰. This agreement was reached in 99% of translational and 91% of rotational displacements. A Bland-Altman analysis demonstrated that there was no statistical significance for the bias between the 2D and 3D measurements. In total 181 CBCT images were assessed against confirmation 2D images. Agreement was achieved in 89.1% of translational (2mm tolerance) and 94.3% of rotational (1⁰ tolerance) orientations. Treatment delivery times in minutes (mins) were reduced from a mean of 55 mins for the first patient to 47 mins for the most recent patient. Average Image review time for the GEC 2D orthogonal images also reduced from 6 mins to 3 mins, however the trend for review time of CBCT remained an average of <10 mins (range 7 mins – 12mins). Conclusion Initial results show that vacbags are suitable for accurate patient positioning and minimising translational displacements. Rotational displacements appeared more random, especially pitch. As such it has been determined, the use of a 6⁰ of freedom bed, combined with daily imaging, is paramount for accurate set up. Future investigations will include a comparison of this data to external institutional data and further review of the effects of contour change on patient position and set up accuracy.