Defense Date

2026

Document Type

Thesis

Degree Name

Master of Science

Department

Medical Physics

First Advisor

Siyong Kim

Second Advisor

Tianjun Ma

Third Advisor

Chris Bartee

Abstract

Background: Delivery of ultra-high dose rate radiation at mean dose rates > 40 Gy/s, also known as FLASH, shows great promise in widening the therapeutic window by decreasing the normal tissue complication probability without compromising tumor control. However, several obstacles stand in the way of the clinical translation of FLASH: the underlying mechanisms of the FLASH effect remain unknown, development of accurate control systems for FLASH linacs is hindered by challenges in performing real-time dosimetry, and research capacity is limited by access to a small number of FLASH-capable machines.

Purpose: This thesis aims to make improvements to a FLASH linac control system based on a commercially available scintillation detector. This system involves minimal modifications to the detector, making it easy to replicate by others. In addition, a novel scintillation detector is proposed that would be capable of automatically correcting for the sensitivity degradation that occurs as a result of the accumulation of radiation damage.

Methods: A decommissioned clinical linac had been previously converted to deliver a 6 MeV electron FLASH beam. Further modifications were made to the FLASH linac to stabilize beam output, and the control system was refined to improve the accuracy of pulse counting and real-time dose calculation. The scintillation detector was cross-calibrated against EBT- XD film. The control system was thoroughly evaluated in both control modes. Dose per pulse dependency of the scintillation detector was also characterized. Extensive literature review was conducted in order to identify scintillation compounds that could be used to create a scintillation detector capable of automatic degradation compensation.

Results: The FLASH linac achieved a mean dose rate of 127.5 ± 14.91 Gy/s at isocenter. In pulse-counting mode, the control system halted the beam within 1 pulse of the target value in 93.6% of tests and in the worst case delivered 3 extra pulses. This overshoot was attributed to latency downstream of the control system. In dose-calculation mode, the control system demonstrated delivery accuracy within 1.45 ± 0.38 Gy when the scintillator was at the depth of maximum dose, and 1.11 ± 0.81 Gy when it was on the surface of a phantom, mimicking an in-vivo dosimetry scenario. The detector response was linear in the dose per pulse range of 0.11 – 0.78 Gy/pulse. Several avenues for further improvement of the controller were identified and discussed. Four promising scintillation compounds were identified for further study as candidates for a new scintillation detector.

Conclusions: The control system performed sufficiently well to be implemented either as a primary control system for FLASH linacs used in pre-clinical research or as a backup in-vivo dose monitor capable of halting the beam in the case of primary control failure, thus preventing significant overdose to a patient. This achieved the goal of improving research capacity as other researchers could easily replicate this system. The identified scintillation compound candidates are expected to lead to a more robust scintillation detector for FLASH.

Rights

© Matthew Lee Richeson

Is Part Of

VCU University Archives

Is Part Of

VCU Theses and Dissertations

Date of Submission

5-6-2026

Available for download on Friday, May 05, 2028

Share

COinS