Validation of a Cable-Free and User-Friendly thermoAcoustic Range Verifier for a 2 Gy Proton Field Delivered to the Liver of an Anthropomorphic Phantom

Why clinicians should know about it

• Picked for Medical Physics (paper of the day, 30 August 2026): Thermoacoustic range verifier validates proton range

Abstract

PURPOSE: Detection and offline correction of proton range errors due to incorrect Hounsfield Unit-to-stopping power conversion with minimal (5-10 minute) extension of the patient-facing clinical workflow using a cable-free thermoAcoustic Range Verifier (tARV). MATERIALS: The accelerator (Mevion Hyperscan) delivered at a conventional dose rate. To generate quantitative results, the target was a multi-modality anthropomorphic phantom (Sun Nuclear) composed primarily of hydrogels with known stoichiometry. The tARV includes six sonar receivers packed around the active imaging face of a wireless ultrasound probe (Clarius PAHD3). METHODS: Two dual-field plans were computed using Raystation 2023b's Monte Carlo engine. "Clinical MC" and "stoichiometric MC" plans were generated using clinical settings and by applying stoichiometry and range measurements (IBA Zebra) of phantom hydrogels. We evaluated the posterior-anterior field that delivered approximately 2 Gy (maximum 27 proton pulses/spot) to seven energy layers (135.6-162.8 MeV). The tARV was positioned anteriorly and distal to the Bragg peak, nearly in-line with the posterior-to-anterior beam trajectory, and rotated to collect a sagittal ultrasound image. Acoustic paths to receivers 5-6 were frequently obstructed by ribs. Measured thermoacoustic pulses generated by each spot were averaged to increase SNR. Time shifts between measured and simulated pulses were used to compute range errors. Range errors were averaged with respect to energy layer. Clinical MC dose was recomputed by shifting each beamlet by the layer's average range error, and imported back into the treatment planning system. RESULTS: Overall, thermoacoustic estimates of range errors for clinical and stoichiometric dosemaps were 6.7±1.7mm and 0.5±1.6mm, respectively. DVHs of recomputed and stoichiometric MC dosemaps were indistinguishable from each other but were clearly distinguishable from the clinical MC DVH. CONCLUSION: Using thermoacoustic range estimates to recompute largely corrected the clinical dosemap. This preliminary, but quantitative study demonstrates that the tARV warrants and is poised for efficient evidence development, such as noninterventional patient safety and offline adaptive studies.

Abstract as published, via PubMed.