ESTRO 2024 - Abstract Book

S3853

Physics - Image acquisition and processing

ESTRO 2024

State-of-the-art image quality for ion radiographs is typically obtained with single-event imaging, although it is currently too slow for practical image guidance and generally not compatible with clinical beam settings. A more accessible approach is to use integrated mode imaging, where signals are acquired from individual pencil beams and each beam is analysed independently. In this work, we report and compare the image quality of integrated mode iRads for protons (pRads) and carbon ions (cRads).

Material/Methods:

iRads were acquired at the Marburger Ionenstrahl-Therapiezentrum (Marburg, Germany). The detector is a plastic volumetric scintillator equipped with three CCD cameras capturing orthogonal views of the 3D energy deposition in the scintillator (fig 1a), and covers a 20x20 cm 2 field of view (FOV). pRad and cRad transmission data were obtained using a 15x15 cm 2 field and a beam spacing of 1 mm. cRads were acquired at 344.62 MeV/u and spot size (full width at half maximum, FWHM) of 5.2 mm at the isocentre, while pRads were obtained at 180.06 MeV and a spot size of 9.4 mm (FWHM). Four phantoms were scanned: 9 tissue-substitute Gammex inserts (fig 2a), 9 custom 3-D printed line pair (lp) modules ranging from 0.5 to 10 lp/cm (fig 1b), an anthropomorphic Alderson head phantom (lateral and front views, fig 1c), and a custom 3-D printed low contrast module consisting of 20 small circular holes with depths of d =[1,3,5,8] and diameters of ⌀ =[1,2,4,7,10] mm (fig 2b). To reproduce realistic scattering conditions, a 10 cm solid water slab was stacked in front of the line pair and low contrast modules phantoms. For image reconstruction, the position of each pencil beam is inferred from the two perpendicular lateral views of the Bragg peak, and WET maps are reconstructed by identifying multiple candidate Bragg peaks and performing a weighted WET reprojection with a point spread function consistent with Fermi-Eyges theory [3]. To evaluate WET accuracy as a function of beam spacing, iRads of Gammex inserts were also reconstructed for beam spacings of 2 to 7 mm by sub-sampling the 1 mm datasets. Exemplar iRads for each phantom and ion species are shown in figures 1 and 2. Image resolution and low contrast detectability are systematically improved with cRads compared to pRads. pRads resolved up to 1.5 lp/cm modules and low contrast holes of up to ⌀ =10 mm and d=8 mm. cRads resolved up to 5 lp/cm modules, and low contrast holes of up to ⌀ =7 mm and d=3 mm. Radiographs of both views of the Alderson head show improved contrast, especially around bony structures, for carbon ions. pRads are found to be blurrier due to increased multiple Coulomb scattering. The WET accuracy, evaluated on Gammex inserts (fig 2a) was found to be similar for both species (around 0.5 mm on average per insert, not shown). WET accuracy is found to be stable (<2% relative variation) for a wide range of beam spacings (1 to 7 mm, fig 2a). Large beam spacings lead to increased under-sampling artefacts for cRads compared to pRads. Results: