Session Item

Intra-fractional motion has a detrimental effect on proton therapy delivery for tumours in the thorax and abdomen, an effect that may be mitigated by real-time motion-management methods such as respiratory gating. Dosimetric studies on complex intra-fractional motion, interplay effects and motion mitigation have been limited to simulations or 1D point, 2D planar, and 2.5D dose measurements. In preparation for a clinical trial, this study demonstrates how anthropomorphic deformable 3D dosimeters embedded in a deformable abdominal phantom can be applied to measure the effects of complex motion and gating on clinically relevant pencil beam scanning (PBS) proton treatments in the liver.

A deformable abdominal phantom was modified to hold an anthropomorphic 3D dosimeter shaped as a human liver (Figure 1). A 3-field PBS proton plan was made on the exhale phase of a 4DCT scan using multi-field robust optimization (±5 mm shift along each axis including ±3.5 % range uncertainty). The plan was normalized to deliver a mean fraction dose of 12 Gy to a clinically realistic clinical target volume (CTV) in the liver. Two batches, each with two radiochromic silicone-based 3D dosimeters, were made. In both batches, one dosimeter was irradiated in a stationary position in the abdominal phantom. The second dosimeter was irradiated while the abdominal phantom moved sinusoidally with a 10 mm peak-to-peak amplitude without (Batch 1) and with (Batch 2) respiratory gating using a 50% duty cycle around the exhale phase monitored by an external marker block (Varian RPM system). The dosimeters were read out using an optical CT scanner which provided a 3D distribution of the radiation induced change in optical attenuation coefficients (Δα).  In this study, Δα was used as a metric for the dose, which allowed direct comparison between the dynamic dose and the static dose in both batches. The motion-experiment datasets were image registered for alignment and evaluated using a global 3D gamma analysis with the stationary experiment as a reference for all voxels with signal above 8% of the maximum Δα. Additionally, a Δα-volume histogram of the CTV was constructed for all experiments normalized to the mean Δα in the CTV for the stationary scenario.

The gamma pass rates for the motion experiments were 68% (3%/3 mm) and 44% (2%/2 mm) without gating and 96% (3%/3 mm) and 85% (2%/2 mm) with gating (Figure 2A and 2B). Furthermore, Δα-volume histograms (Figure 2C) show better CTV coverage in the gated than non-gated scenario. Further experiments are needed to investigate the influence of interplay effects, CTV and motion magnitude.

An anthropomorphic 3D dosimeter was for the first time embedded in a deformable abdominal phantom, and successfully measured the motion-induced dose perturbation of liver proton PBS treatments.