Out-of-field neutron doses in proton therapy: from experimental characterization to a practical calculation tool

Objective. To experimentally characterize out-of-field neutron dose in a pencil beam scanning (PBS) proton therapy treatment room using multiple detectors, to analyze its dependence on proton energy, gantry angle, distance to isocenter, and field size, and to develop a practical calculation tool for neutron dose estimation.Approach. Two extended-range ambient detectors, bubble detectors (BDs), track-etch, thermoluminescent (TLDs), and three electronic personal dosimeters were used. Measurements were taken at up to 21 points for five proton energies, four gantry angles (θ), four field sizes, four measurement lines, and distances of 1-4.7 m from the isocenter. Room symmetry was evaluated. At each point,H*(10) orH_p(10) =a∗Epbwas assessed, and measurements were also performed for a clinical irradiation, with the resulting doses compared against weighted sums of contributions from individual energy layers. Bland-Altman analyses were applied.Main results. The room was symmetric for θ = 270°-90°, while doses at 0° were on average 8% lower than at 180°. Neutron doses for the spot field and 10 × 10 cm²field were interchangeable, with 20 × 20 and 30 × 30 cm²fields differing by 2%-22%. Dose dependence on proton energy followed a power-law, with best fits for ambient detectors and moderate performance for BDs and TLDs. Linear superposition held for ambient and BDs, validating that neutron doses can be approximated as weighted sums of contributions from individual energy layers. A Python-based calculator for out-of-field neutron doses, provided as supplementary material, was developed for six detectors.Significance. This study delivers an easy-to-use Python tool for first-order estimation of out-of-field neutron doses in a PBS proton therapy room for any treatment plan or location. The tool is based on the linear superposition model, is validated against measurements, and accounts for detector-dependent uncertainties.