Stopping power monitoring during proton therapy by means of prompt gamma timing: first experimental results with a homogeneous phantom

TL;DR

This work addresses uncertainties in proton therapy dose delivery by introducing SER-PGT, a method that uses prompt gamma timing to infer both particle range and stopping power. The authors implement a multi-detector SER-PGT system with LGAD-based proton timing and LaBr$_3$:Ce photon detection, and they apply a 2D spatiotemporal reconstruction to experimental data from a 226.9 MeV synchrotron beam on a homogeneous PMMA phantom. They report an average stopping-power error of $8\% \pm 3\%$ relative to NIST PSTAR and a SPR of $2\% \pm 2\%$ relative to water at 100 MeV, with range-difference detection for a 4 cm air-gap at a precision of $3$ mm (SD); a 3.8 cm mean range shift was observed between geometries. The results demonstrate the feasibility of recovering range and stopping power from particle kinematics and prompt gamma measurements, offering a potential route toward in-vivo stopping-power verification in proton therapy and informing future clinical validation and hardware improvements.

Abstract

Proton therapy's full potential is limited by uncertainties that prevent optimal dose distribution. Monitoring techniques can reduce these uncertainties and enable adaptive treatment planning. Spatiotemporal Emission Reconstruction from Prompt-Gamma Timing (SER-PGT) is a promising method that provides insights into both particle range and stopping power, whose calculation would normally require knowledge about patient tissue properties that cannot be directly measured. We present the first experimental results using a 226.9 MeV synchrotron-proton beam impinging on a homogeneous phantom at a sub-clinical intensity (2 - 4 x 10^7 pps). SER-PGT uses data from a multi-detector setup: a thin and segmented Low Gain Avalanche Diode for proton detection and Lanthanum Bromide-based crystals for photon detection. The estimated stopping power profile showed an 8% +- 3% average error compared to NIST PSTAR values, and 2% +- 2% deviation relative to water at 100 MeV. Range assessment in a phantom with a 4 cm air-gap successfully identified the range shift with a 3 mm standard deviation. These results demonstrate the feasibility of using SER-PGT to recover both range and stopping power information through particle kinematics and PGT measurements.

Figures (6)

• Figure 1: (a): Schematic of the phantom without (yellow) or with (blue) the airgap, the reference detector (R, red), and the photon detector positions (1-14, green). (b): setup in treatment room with the photon detectors in positions 1 and 12.
• Figure 2: Waveform snapshots after baseline subtraction and signal inversion, showing the photon signal in a PMT (red) and two proton signals in a LGAD-channel (blue).
• Figure 3: Example of PGT spectra of the first subset with and without the air cavity for detector positions 1 (a) and 11 (b).
• Figure 4: Reconstructed spatiotemporal distributions using the first subset of PGT-data. Left: homogeneous phantom after 100 (top) and 1000 (bottom) iterations. Right: 4-cm air cavity after 100 (top) and 1000 iterations (bottom). No threshold was applied to these distributions. Spatial phantom boundaries are marked with red dashed lines.
• Figure 5: Range differences for homogeneous subsets (blue) and 4-cm air gap subsets (red), both centred at 0.08cm (± 0.3 cm). Differences between the air-gap and homogeneous subsets (green) show the expected 3.8 cm shift (± 0.3 cm).