Acoustic time-of-flight for proton range verification in water Kevin C. Jones Department of Radiation Oncology, University of Pennsylvania, Philadelphia, Pennsylvania 19104 François Vander Stappen Ion Beam Applications SA, Louvain-la-Neuve 1348, Belgium Chandra M. Sehgal Department of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania 19104 Stephen Avery a) Department of Radiation Oncology, University of Pennsylvania, Philadelphia, Pennsylvania 19104 (Received 29 March 2016; revised 27 July 2016; accepted for publication 3 August 2016; published 25 August 2016) Purpose: Measurement of the arrival times of thermoacoustic waves induced by pulsed proton dose depositions (protoacoustics) may provide a proton range verification method. The goal of this study is to characterize the required dose and protoacoustic proton range (distance) verification accuracy in a homogeneous water medium at a hospital-based clinical cyclotron. Methods: Gaussian-like proton pulses with 17 μs widths and instantaneous currents of 480 nA (5.6 × 10 7 protons/pulse, 3.4 cGy/pulse at the Bragg peak) were generated by modulating the cyclotron proton source with a function generator. After energy degradation, the 190 MeV proton pulses irradiated a water phantom, and the generated protoacoustic emissions were measured by a hydrophone. The detector position and proton pulse characteristics were varied. The experimental results were compared to simulations. Different arrival time metrics derived from acoustic waveforms were compared, and the accuracy of protoacoustic time-of-flight distance calculations was assessed. Results: A 27 mPa noise level was observed in the treatment room during irradiation. At 5 cm from the proton beam, an average maximum pressure of 5.2 mPa/1 × 10 7 protons (6.1 mGy at the Bragg peak) was measured after irradiation with a proton pulse with 10%–90% rise time of 11 μs. Simulation and experiment arrival times agreed well, and the observed 2.4 μs delay between simulation and experiment is attributed to the difference between the hydrophone’s acoustic and geometric centers. Based on protoacoustic arrival times, the beam axis position was measured to within ( x ,y ) = (−2.0, 0.5) ± 1 mm. After deconvolution of the exciting proton pulse, the protoacoustic compression peak provided the most consistent measure of the distance to the Bragg peak, with an error distribution with mean = −4.5 mm and standard deviation = 2.0 mm. Conclusions: Based on water tank measurements at a clinical hospital-based cyclotron, protoa- coustics is a potential method for measuring the beam’s position ( x and y within 2.0 mm) and Bragg peak range (2.0 mm standard deviation), although range verification will require simulation or experimental calibration to remove systematic error. Based on extrapolation, a protoacoustic arrival time reproducibility of 1.5 μs (2.2 mm) is achievable with 2 Gy of total deposited dose. Of the compared methods, deconvolution of the excitation proton pulse is the best technique for extracting protoacoustic arrival times, particularly if there is variation in the proton pulse shape. C 2016 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4961120] Key words: pulsed proton beam, range verification, acoustic pulses, ionoacoustics, protoacoustics 1. INTRODUCTION Proton range uncertainty limits the benefits of proton therapy. The proton’s unique dose deposition, which reaches its maximum (the Bragg peak) at an energy-dependent depth before abruptly falling off, allows for reduced entrance and exit doses compared to photon therapy, which exhibits monotonic decrease after the first few centimeters. Over- shooting or under-shooting of the proton beam due to range uncertainty, however, has large associated risks because of the steep Bragg peak dose gradient. Range uncertainty is primarily due to patient setup, changes in patient anatomy, CT conversion, and treatment planning errors. 1 Improvements to treatment planning 1 or CT conversion 2 may reduce the plan range uncertainty, but they will not correct for setup error or patient anatomy changes (unless imaging and planning are performed before each fraction). If possible, the ideal approach to reducing proton range uncertainty is in vivo measurement of the proton penetration depth. 3,4 Toward this goal, positron emission tomography and prompt gamma monitoring are promising possibilities that measure activation by protons, but they are limited by the variable lifetimes of the emitting radioisotopes, energy threshold for activation, and cost. 3 Another potential method, 5213 Med. Phys. 43 (9), September 2016 0094-2405/2016/43(9)/5213/12/$30.00 © 2016 Am. Assoc. Phys. Med. 5213