Dosimetric validation of Monte Carlo and analytical dose engines with raster-scanning $^1$H, $^4$He, $^{12}$C, and $^{16}$O ion-beams using an anthropomorphic phantom

With high-precision radiotherapy on the rise towards mainstream healthcare, comprehensive validation procedures are essential, especially as more sophisticated technologies emerge. In preparation for the upcoming translation of novel ions, case-/disease-specific ion-beam selection and advanced multi-particle treatment modalities at the Heidelberg Ion-beam Therapy Center (HIT), we quantify the accuracy limits in particle therapy treatment planning under complex heterogeneous conditions for the four ions ($^1$H, $^4$He, $^{12}$C, $^{16}$O) using a Monte Carlo Treatment Planning platform (MCTP), an independent GPU-accelerated analytical dose engine developed in-house (FRoG) and the clinical treatment planning system (Syngo RT Planning). Attaching an anthropomorphic half-head Alderson RANDO phantom to entrance window of a dosimetric verification water tank, a cubic target spread-out Bragg peak (SOBP) was optimized using the MCTP to best resolve effects of anatomic heterogeneities on dose homogeneity. Subsequent forward calculations were executed in FRoG and Syngo. Absolute and relative dosimetry was performed in the experimental beam room using 1D and 2D array ionization chamber detectors. Mean absolute percent deviation in dose (|%Δ|) between predictions and PinPoint ionization chamber measurements were within ∼2% for all investigated ions for both MCTP and FRoG. For protons and carbon ions, |%Δ| values were ∼4% for Syngo. For the four ions, 3D-γ analysis (3%/3mm criteria) of FLUKA and FRoG presented mean passing rates of 97.0( ± 2.4)% and 93.6( ± 4.2)%. FRoG demonstrated satisfactory agreement with gold standard Monte Carlo simulation and measurement, superior to the commercial system. Our preclinical trial landmarks the first measurements taken in anthropomorphic settings for helium, carbon and oxygen ion-beam therapy.