Clinical Integration of a Nozzle-Mounted Quad-Camera Prompt Gamma Imaging for Proton Range Verification
Abstract
Purpose
To integrate the first nozzle-mounted, quad-camera prompt gamma imaging (PGI) system on a clinical proton therapy gantry and to evaluate its efficacy for proton range verification under realistic treatment delivery conditions.
Methods
A PGI system consisting of four position-sensitive solid-state Compton camera modules were mounted to a range shifter attached to the gantry snout of a proton therapy treatment nozzle. The camera modules were placed in a rectangular geometry centered on the proton beam axis with each module comprises four cadmium zinc telluride (CdZnTe) detector crystals. Each camera module was independently powered and connected via Ethernet using static IPs, with custom software enabling synchronized and real-time data acquisition during beam delivery. Measurements were performed under clinical pencil-beam scanning (PBS) delivery, including single-energy spots and SOBP fields at 170 and 245 MeV, gantry angles of 90° and 270°, and SOBP doses of 2 and 7.5 Gy. The detected prompt-gamma scatter events were used to reconstruct prompt-gamma images at a resolution of 2.5×5×5 mm3 using a GPU-based kernel-weighted back-projection (KWBP) algorithm with Compton-line filtering. PGI’s sensitivity to range shifts was assessed for varying voxel size, reconstruction volume, and kernel bandwidth based on beam-axis prompt-gamma profiles.
Results
The nozzle-mounted quad-camera PGI system operated stably during clinical proton beam delivery. Controlled range-shift inserts (0–10 mm) yielded reproducible prompt-gamma emission localization across all runs, and KWBP-based reconstruction enabled identification of 7–10 mm range shifts in a clinically delivered proton beam.
Conclusion
This study demonstrated the feasibility of the first nozzle-mounted, quad-camera Compton PGI system integrated into a clinical proton therapy gantry. Result showed the system’s ability to generate PG images to detect range shifts during proton beam delivery. Future studies are warranted to optimize PGI through novel algorithms, such as deep-learning, to further improve its precision for proton range verification.