Dosimetric and geometric end-to-end accuracy of a magnetic resonance guided linear accelerator

Abstract

The introduction of real-time imaging by magnetic resonance guided linear accelerators (MR-Linacs) enabled adaptive treatments and gating on the tumor position. Different end-to-end tests monitored the accuracy of our MR-Linac during the first year of clinical operation. We report on the stability of these tests covering a static, adaptive and gating workflow. Film measurements showed gamma passing rates of 96.4% ± 3.4% for the static tests (five measurements) and for the two adaptive tests 98.9% and 99.99%, respectively (criterion 2%/2mm). The gated point dose measurements in the breathing phantom were 2.7% lower than in the static phantom.

Keywords: End-to-end test; Gating; MR-Linac; Treatment adaption.

Fig. 1

Set-up for the three different end-to-end tests. On top, the phantoms are shown, first the spherical phantom followed by the in-house made silicon phantom and the thorax motion phantom. Below, the characteristics of the static, adaptive and gating measurements are indicated. In the next row the dose distribution in the TPS is displayed. Then, the longitudinal measured and planned dose profiles are shown for the static and adaptive tests. For the gating tests the calculated longitudinal dose profile is combined with the point dose measurements.

Fig. 2

Point dose measurements in the static phantom, in the breathing phantom but without gating, in the breathing phantom with a 3 mm gating window around the GTV and in the breathing phantom with a 2 mm gating window around a surrogate tracking structure. The red dot in the first column corresponds to the dose in the TPS.