ESTRO 2024 - Abstract Book

S4750

Physics - Quality assurance and auditing

ESTRO 2024

On the MR-linac online adaptive treatments can be performed by taking the target and organ at risk deformation into account (adapt-to-shape procedure (ATS)) based on the online MRI. As a result, the online plan has better a tumor coverage. To determine the geometrical accuracy of the online ATS workflow of the Unity MR-linac, an end-to-end test is performed. In this study we determine the 3D geometrical accuracy of the online ATS procedure using 3D gel dosimeters. Furthermore, we assess the reproducibility of the end-to-end test performed on the ATS workflow.

Material/Methods:

All measurements are performed on the same Unity MR-linac (Elekta AB, Stockholm, Sweden) to assure that calibration and machine differences are eluded. The Prime head phantom (RTsafe, Athens, Greece), with a gel dosimeter mounted inside the phantom, is used for the end-to-end test. A pre-treatment MRI scan of the Prime phantom is acquired on a 1.5 T MRI (Philips Healthcare, Best, the Netherlands) in combination with a head coil. Body contours are delineated as well as two anti-parallel cone shaped PTVs to obtain unambiguous high dose regions with sharp gradients. These PTVs are positioned such that the shapes of the PTVs have the high dose regions within the volume of the cylindrical gel dosimeter accordingly. The pre-treatment plan is created using the Monaco Treatment Planning System (v. 5.51.10, Elekta AB, Stockholm, Sweden (TPS)). The maximum prescribed dose for the pre-treatment plan is approximately 10 Gy for both PTVs. For online adaptation the ATS workflow is used. All online contours are deformably registered except the PTVs, which are registered rigidly. In order to ensure constant positioning of the Prime phantom, an index bar is used. Stable fixation is assured by making use of clinical treatment aids. Gels manufactured by RTsafe are used during the measurements. To determine the reproducibility of the end-to-end test, measurements are performed five times, on five different days, within two weeks. For each measurement a new gel is used. The gels are handled according to the manufacture’s guidelines, and scanned approximately 24 hours after dose delivery. The scans are performed on a 1.5 T MRI scanner employing a scan protocol designed by RTsafe in combination with a head coil. The online treatment plan is recalculated using an isometric voxel size of 1.0 mm 3 and a statistical uncertainty of 0.5% per segment (pre-treatment plan has 148 segments). A 3D dose grid (resolution is 1.0 x 1.0 x 1.0 mm 3 ) from the gel measurement is obtained from RTsafe and registered in 3D to the calculated online dose distribution to determine the geometrical accuracy of the online workflow. The reproducibility of the end-to-end workflow is assessed by determining the standard deviations for the x, y, z directions (IEC61217) and the 3D vector lengths for the five different measurements.

Results:

The geometrical deviations in the x, y and z directions (figure 1) were small with the largest deviation of 0.3 mm in the y direction. The mean geometric deviations of the x, y and z were -0.1 ± 0.1 (1SD) mm, 0.1 ± 0.1 (1SD) mm and 0.0 ± 0.1 (1SD) mm, respectively. The 3D vector length had a mean deviation of 0.2 ± 0.1 (1SD) mm.