ESTRO 2024 - Abstract Book

S3816

Physics - Image acquisition and processing

ESTRO 2024

An advanced electron density phantom (Model 1467, Sun Nuclear, Melbourne, USA) was configured into a single setup with two 8 cm long sections in the z-direction, where the 20 cm diameter cylindrical inner section represented an adult head, whereas the 40 cm x 30 cm elliptical outer section represented an adult abdomen. Four 16 cm long, 28.5 mm diameter insert rods of different bone tissue surrogate materials were positioned equidistantly on 15 cm diameter circle concentric with the central axis of the phantom (figure 1).

The phantom was scanned with SECT with three different CT scanners; Brilliance Big Bore (Philips Healthcare, Best, NL), Somatom Definition AS and go.Open Pro (Siemens Healthineers, Forchheim, DE) henceforth denoted S1, S2 and S3, respectively. The SECT scans were performed with clinical scan protocols used for proton radiotherapy treatment planning for H&N and/or CSI. Images from the H&N protocol of S3 were reconstructed with and without iterative bone beam hardening correction (iBHC). DECT scans were performed with S2 and S3, from which SPR images were created with the vendor proprietary DirectSPR algorithm. The CT-numbers of the bone inserts were extracted from volumes-of-interest (4 cm long, 20 mm diameter) in the contouring workspace in Eclipse 16.1 (Varian Medical systems, Palo Alto, USA). For the SECT images, the difference of the mean CT-numbers between the head and body geometries were calculated for each bone material rod. The theoretical reference SPRs of the bone materials were calculated with the Bethe formula using vendor-specified elemental compositions and batch specific mass densities. The SPR CT-numbers in the DirectSPR images were compared to the reference SPR. The SPR CT-numbers were expressed in SPR scaled Hounsfield units [SPR-HU] according to 1000*(SPR-1).

Results: