ESTRO 2021 Abstract Book

S405

ESTRO 2021

derived from paediatric proton therapy patients (11.5-year-old (Ewing’s sarcoma), 18-year-old (salivary sarcoma), 5-year-old (glioma)). For each phantom, the dose was optimised on SECT-predicted RSPs, then forward calculated on theoretically calculated RSPs (theoretically delivered dose) and DECT-predicted RSPs. Range and dose differences are assessed. Results With SECT, soft tissues have mean RSP errors and standard deviations (µ±σ) up to (2.00±4.12)% observed in newborns, compared to (0.24±1.14)% in adult males. Mean RSP prediction errors for bones are up to (- 3.42±4.74)% in paediatrics, much higher than in adult males with (0.03±0.62)%, see Fig 1a for all age groups. With DECT, mean errors reduce to (0.17±0.13)% and (0.23±0.22)% in newborns (soft/bones), see Fig 1b. In the glioma phantom (Fig 2), we observe average range overshoots of (8.17±3.26) mm (beam 1) and (6.16±0.72) mm (beam 2) introduced by the erroneous SECT-predicted RSP, causing an additional dose >2 Gy (4% of prescribed dose) to the healthy brain distal to the target volume. DECT reduces the range errors to (0.05±1.32) mm and (0.07±0.37) mm, respectively, and the dose errors to <0.5 Gy. Similar observations are made in the Ewing’s and salivary sarcoma phantoms, with dose errors >5 Gy (9.9% prescribed dose) and range errors up to (7.80±1.19) mm distal to the target, reduced to (0.12±0.62) mm with DECT.