Development and benchmarking of a dose rate engine for raster‐scanned FLASH helium ions

Abstract

AbstractBackgroundRadiotherapy with charged particles at high dose and ultra‐high dose rate (uHDR) is a promising technique to further increase the therapeutic index of patient treatments. Dose rate is a key quantity to predict the so‐called FLASH effect at uHDR settings. However, recent works introduced varying calculation models to report dose rate, which is susceptible to the delivery method, scanning path (in active beam delivery) and beam intensity.PurposeThis work introduces an analytical dose rate calculation engine for raster scanned charged particle beams that is able to predict dose rate from the irradiation plan and recorded beam intensity. The importance of standardized dose rate calculation methods is explored here.MethodsDose is obtained with an analytical pencil beam algorithm, using pre‐calculated databases for integrated depth dose distributions and lateral penumbra. Dose rate is then calculated by combining dose information with the respective particle fluence (i.e., time information) using three dose‐rate‐calculation models (mean, instantaneous, and threshold‐based). Dose rate predictions for all three models are compared to uHDR helium ion beam (145.7 MeV/u, range in water of approximatively 14.6 cm) measurements performed at the Heidelberg Ion Beam Therapy Center (HIT) with a diamond‐detector prototype. Three scanning patterns (scanned or snake‐like) and four field sizes are used to investigate the dose rate differences.ResultsDose rate measurements were in good agreement with in‐silico generated distributions using the here introduced engine. Relative differences in dose rate were below 10% for varying depths in water, from 2.3 to 14.8 cm, as well as laterally in a near Bragg peak area. In the entrance channel of the helium ion beam, dose rates were predicted within 7% on average for varying irradiated field sizes and scanning patterns. Large differences in absolute dose rate values were observed for varying calculation methods. For raster‐scanned irradiations, the deviation between mean and threshold‐based dose rate at the investigated point was found to increase with the field size up to 63% for a 10 mm × 10 mm field, while no significant differences were observed for snake‐like scanning paths.ConclusionsThis work introduces the first dose rate calculation engine benchmarked to instantaneous dose rate, enabling dose rate predictions for physical and biophysical experiments. Dose rate is greatly affected by varying particle fluence, scanning path, and calculation method, highlighting the need for a consensus among the FLASH community on how to calculate and report dose rate in the future. The here introduced engine could help provide the necessary details for the analysis of the sparing effect and uHDR conditions.