ESTRO 2021 Abstract Book

S744

ESTRO 2021

Brain patients were selected to reduce the influence of anatomy variations and focus on image intensities. For three patients, single field uniform dose pencil beam scanning proton plans were optimized on the original planning CT (planCT) and on a scCBCT. The two plans per patient were irradiated on patient-specific 3D printed head phantoms (see figure 1) with film (all patients) and gel (one patient) dosimetry inserts. The resulting dose distributions were compared based on proton range differences (PRDs) (80% distal fall-off). Additionally, every plan was recalculated on the respective other image modality. The conventional recalculation approach calculates the PRDs between the original and the recalculated planCT dose distribution on scCBCT. Our proposed simulation approach compares the original dose to the dose of the scCBCT- optimized plan recalculated on the planCT. This mimics the measurements we performed, where the PRDs are determined from the measured dose distributions of both plans on the same phantom. The PRDs between the two optimized plans on scCBCT and planCT were applied as a correction to the simulation and the measurement PRDs to enable a direct comparison to the recalculation approach.

Figure 1: Photo of a head phantom (A), the film holder (B) and an empty gel cylinder (C).

Results The simulation approach was found to be equivalent to the recalculation approach in terms of deviations of the median values (< 0.1 mm) and interquartile ranges (< 0.3 mm, IQRs) of the PRDs. High agreement was observed between the PRDs of the simulation and the film measurement (see evaluation of the PRDs in the film plane in Figure 2). The median values and IQR of the PRD maps comparing the gel measurement, simulation and recalculation approach showed a deviation of less than 0.7 mm and 0.5 mm, respectively.