Session Item

Pencil beam scanning (PBS) proton therapy offers individualized conformal dose distributions. However, intra-fractional motion in the abdomen and thorax is a major challenge and can severely degrade the dose distribution.
Building up to a clinical protocol for liver tumours in our proton centre, we are developing an anthropomorphic dosimetry phantom, containing clinically realistic deformable three-dimensional (3D) dosimetry inserts, onto which complex deformations can be applied.
The aim of this study was to establish a method for producing high precision anthropomorphic 3D dosimeters to fit into this phantom. 

The dosimeter moulds were created using a 3D printed organ as a negative. The shape of an adult-size liver was divided into three parts to minimize the optical pathlength through the individual dosimeter parts and 3D printed using a polylactic acid filament. The parts were sanded down, coated with a layer of primer, and spray-painted to provide a smooth surface (Figure 1).
A casting basin was made from sculpting clay and filled with silicone moulding rubber. The 3D print was suspended in the basin and the moulding rubber was left to cure for 24 hours.
The 3D dosimeters were prepared using transparent silicone kit, chloroform, and the radiation-sensitive component leucomalachite green, poured into the moulds and cured for 72 hours prior to irradiation. A specially designed nut was suspended in the dosimeter during curing for later attachment to the read-out system.
The dosimeters were x-ray computed tomography (CT) scanned and imported to the treatment planning system (TPS). The treatment plan consisted of six conventional 6 MV photon fields (15 mm x 15 mm) evenly spaced on a 180-degree arc with a prescribed dose of 5 Gy normalized to 100% in target maximum. A four-degrees-of-freedom couch correction was performed using a cone-beam CT.
The optical density (OD) of the dosimeters was read-out pre- and post-irradiation using an optical CT scanner with 1000 projections and reconstructed using the ordered-subsets-convex total-variation method.
The TPS dose distribution was compared to the experimental dose distributions using a 3%/3 mm 3D gamma analysis including only voxels above 10% of the maximum TPS dose. 

The silicone rubber moulds were successfully made with great detail reproducing all visible filament lines from the 3D prints and had the advantage that it was easy to separate the 3D dosimeters from the moulds compared to rigid moulds. Most of the dose was delivered in just one of the liver-dosimeter parts (Figure 2) and here we achieved a gamma pass rate of 87%. The main challenges in the optical read-out were the dosimeter edges along with the signal-to-noise ratio.

The described method successfully produced reusable anthropomorphic moulds for 3D dosimeters. Furthermore, the liver irradiation showed a fair correlation between the TPS calculation and experimental measurements and indicates its suitability for future proton irradiations.