Detecting Proton Beam Accuracy Through Prompt Gamma Imaging

Proton therapy is now a mainstream radiotherapy modality, offering clinicians a way to concentrate dose within a tumour while sparing healthy tissue. The defining physical advantage of protons is the Bragg peak, a sharp deposition of energy at a depth set by the beam's initial energy. Knowing where that peak actually stops inside the patient is the central clinical question, because uncertainties of even a few millimetres can compromise target coverage or irradiate organs at risk.

Medical physicists have spent decades refining the tools used to plan and deliver proton treatments. An article on the organisation's professional mandate captures how such bodies connect research advances to clinical practice, and range verification sits at the intersection of physics, instrumentation and patient safety.

Prompt gamma imaging offers a direct way to confront range uncertainty. It exploits radiation emitted at the moment protons interact with tissue, produced within nanoseconds of the nuclear reaction and therefore suitable for in-room, near-real-time monitoring. The Banff meeting was one venue where competing detection strategies were compared side by side.

For delegates travelling from Australian centres, the technique resonates with growing local investment in particle therapy. The Peter MacCallum Cancer Centre in Melbourne commissioned Australia's first hospital-based proton facility, while the Australian Bragg Centre for Proton Therapy in Adelaide has moved from announcement into construction. Practitioners in Sydney, Brisbane and Perth face the same challenges as their European counterparts, while the regulatory environment shaped by ARPANSA and the Therapeutic Goods Administration adds documentation requirements that emerging technologies must satisfy.

The physical origin of prompt gammas

When a proton beam traverses tissue, it slows through Coulomb interactions with atomic electrons and, less often, collides directly with atomic nuclei. Those collisions excite target nuclei to short-lived states that de-excite by emitting gamma rays with characteristic energies between 1 and 10 MeV. Because the cross section peaks near the end of the proton range, the spatial distribution of prompt emission closely follows the depth-dose curve.

The clinically useful emissions come from light elements in soft tissue, particularly oxygen, carbon and nitrogen. Lines such as the 4.44 MeV photon from carbon-12 and the 6.13 MeV line from oxygen-16 provide a spectroscopic fingerprint, and the emission depth correlates with the proton's residual range.

A second class of emissions comes from positron-emitting fragments of the same reactions, which form the basis of offline PET verification. Understanding the relative yields, angular distributions and time structure of these photons is essential for designing detectors that cope with the noisy environment of a clinical treatment room.

Detection technologies for prompt gamma imaging

Detector architectures have matured from prototypes toward clinical systems. Slit-camera designs use a collimating slit to project emission onto a position-sensitive detector, producing a one-dimensional profile whose centroid shifts when range deviates by as little as two or three millimetres.

Detector families investigated for this application include:

• Collimated scintillator arrays paired with silicon photomultipliers or conventional photomultiplier tubes
• Compton cameras based on CdZnTe, Si or LaBr3 crystals
• Multi-slit and knife-edge cameras with high-z collimators
• Time-of-flight systems that exploit the nanosecond time structure of prompt emission
• Coded-aperture designs that balance efficiency against spatial resolution

Each architecture trades spatial resolution against energy resolution, detection efficiency and tolerance to the secondary radiation field generated during irradiation. The choice depends on whether the goal is in-room online monitoring or post-treatment dosimetric reconstruction.

Range uncertainty and clinical translation

Residual range uncertainty is typically quoted at three to five percent for single-energy proton beams, arising from CT Hounsfield-to-stopping-power conversion, organ motion during delivery, and setup variation. In pencil-beam scanning, range uncertainties can translate into under-dosage of the distal target edge or over-dosage of adjacent critical structures.

Verification methods that confirm the actual range give clinicians an actionable signal, a basis either to adapt the plan immediately or to relax dosimetric margins in subsequent fractions. Prompt gamma imaging stands apart by delivering a measured profile within seconds of beam delivery, without waiting for radioactive decay or offline reconstruction.

PET of positron-emitting fragments and proton radiography remain active areas, but neither delivers a measured profile within seconds, which is precisely what prompt gamma systems provide.

Verification in the Australian context

Australia has been an active contributor to the field, with physicists publishing simulation studies, characterising detectors, and collaborating with European groups at OncoRay and the Institut Curie. Researchers at the University of Sydney and at the Peter MacCallum Cancer Centre have explored prompt gamma monitoring for hypofractionated regimens, where each fraction carries heavier dosimetric consequences and verification becomes more rather than less valuable.

The Australian regulatory framework shaped by ARPANSA's Code of Practice for Radiation Protection in Medical Uses of Ionizing Radiation, alongside Therapeutic Goods Administration oversight of treatment planning software, places strong documentation requirements on new technologies. Local commissioning teams favour solutions that benchmark against reference dosimetry and fit within existing quality assurance schedules.

Integration with Australian workflows has practical implications for staffing, since Medicare funding for particle therapy is still being negotiated and many centres pair public hospital infrastructure with private treatment packages. A verification tool that reduces margins without increasing workload aligns closely with the funding logic Australian public hospitals operate under, and physicists here are accustomed to arguing their case in economic terms before introducing the more elegant physics.

Integration with treatment workflows

Bringing prompt gamma imaging into routine clinical practice requires more than a working detector. Geometric calibration of the camera relative to the beam coordinate system, regular quality assurance checks, and streamlined reconstruction software all have to fit within the time pressures of a treatment fraction.

Workflow models proposed in the literature include:

• Pre-treatment verification, where the plan is delivered to a phantom and the measured profile is compared to a prediction
• In-vivo monitoring during the actual fraction, with measured profiles overlaid on the planned distribution
• Post-treatment review, where the recorded profile feeds back into adaptive planning

Each model requires a different level of integration with the treatment control system. In-vivo monitoring is the most ambitious because it implies active intervention if range deviations exceed a clinical threshold, raising questions about alarm fatigue, operator training and the legal framework for pausing treatment.

Future directions and open questions

The trajectory of prompt gamma imaging points toward systems that combine multiple detection modalities and exploit time-of-flight information more aggressively. Hybrid designs pairing a Compton camera with a multi-slit collimator have shown promise in Monte Carlo simulations and experimental beam tests, offering wide spatial coverage and reasonable spectroscopic resolution.

A second direction is integration with motion management. Intrafractional motion of thoracic and abdominal tumours complicates interpretation of any measured profile, and correlating prompt gamma data with external surrogate signals or on-board volumetric imaging is active work. Several Australian groups contribute through collaborations spanning diagnostic imaging, adaptive replanning and surface-guided radiotherapy.

Open questions shaping research include how plans should be adapted when a range deviation is detected, what minimum detectable shift justifies clinical action or allows margin reduction, and what level of evidence regulators such as the TGA will require before clearing these devices. Funding discussions in mixed public-private systems such as Australia also stand to be reshaped by prompt gamma monitoring.

For attendees visiting Banff for the COMP 2014 Annual Scientific Meeting, the science discussed during the day will colour how they explore the surrounding landscape between sessions. A walk along the Bow River after a session on prompt gamma kinematics makes a curious kind of sense, because the physical principles behind range uncertainty draw on the same continuum of nuclear, atomic and environmental physics that defines the meeting's broader identity. Delegates can consult a guide to between-session sightseeing for ideas that pair well with evening poster sessions and morning plenaries, and anyone with verification data or detector prototypes to share is encouraged to bring their work to the range monitoring working group sessions on site.

As prompt gamma imaging moves from physics demonstration to clinical deployment, the conversations started at gatherings such as COMP 2014 will guide how this powerful tool is integrated into the next generation of proton therapy practice. Medical physicists across Australian proton centres, alongside their international collaborators, are well positioned to shape the standards and evidence base that will determine where this technology lands first, and where it eventually becomes routine for patients everywhere.