Comparison of Two Modified Linear Accelerators for Use in FLASH Clinical Trials

Abstract

<h3>Purpose/Objective(s)</h3> This study is the first direct comparison of two converted clinical linear accelerators producing ultra-high dose rate electron beams in the same way. It sought to quantify the differences in the spatial and temporal properties of the beams in relation to potential clinical impacts. <h3>Materials/Methods</h3> FLASH beams were produced on Varian Trilogy and 21EX linear accelerators by selecting the 10MV photon mode, retracting the target and setting the carousel to an empty port to give access to the pristine electron beam. Lateral profiles of an open beam with no accessories were measured by radiographic film (depth = 2cm, SSD = 100cm) in Solid Water. Percent depth dose (PDD) was measured along the central axis with radiographic film placed at varying depths in Solid Water. The temporal pulse structure and per-pulse output was recorded with a photomultiplier tube that amplified the signal from an optical fiber placed outside the beam with film providing cumulative absolute dosimetry. Pulse-to-pulse dosimetry was further confirmed with a diode detector centered in the beam under 1cm buildup. The practical range was calculated using PDD data. Beam width was reported as the FWHM of a Gaussian fit to the profile. Dose per-pulse (DPP) was calculated using film and the number of pulses detected with the PMT. Mean dose rate was calculated using the DPP and the accelerator rep rate. <h3>Results</h3> As reported in the included table, spatial beam characteristics were similar between the machines with the Trilogy showing a slightly wider profile in inline and crossline directions. The DPP on the Trilogy was 19.6% higher than the 21EX at 0.90 Gy/pulse with a mean dose rate of 322 Gy/s, compared to the 270 Gy/s on the 21EX. The Trilogy accelerator exhibited more stability during the initial pulses as indicated by the higher 25% quartile dose per-pulse. The difference between the 25% and 75% quartiles was 0.25 and 0.02 Gy/pulse on the 21EX and Trilogy, respectively. The 21EX exhibited additional low-dose outliers in the initial pulses and showed significant ramp-up and variability during the first 20 pulses of a 36-pulse delivery. The Trilogy was nearly stable after the first pulse. <h3>Conclusion</h3> For clinically relevant doses, the entire FLASH treatment may consist of only a few initial pulses. Stability across these initial pulses with high dose rates from the start are key to ensure the quality, repeatability, and safety of UHDR deliveries. Although significant variabilities have been observed between the two FLASH Linacs, the Trilogy demonstrated potential superiority in stability, which particularly matters for the clinical translation of FLASH-RT.