Siting a FLASH radiotherapy beam in the animal house, without a bunker

A bunker is the standard answer for siting a radiotherapy accelerator: a room thick enough to stop stray radiation. For a small-animal FLASH program, the standard answer breaks down. A dedicated shielded room is too expensive to add to a building already configured as a vivarium, and the experiments have to happen where the mice are.

A team at the Centre for Advanced Preclinical in vivo Research in Italy took a different approach. "Preclinical" here means work on animal models, not patients, the stage that sits between physics-bench experiments and the first human trials. The team installed a 9 MeV electron FLASH accelerator inside one of their animal-housing rooms, then used FLUKA.CERN 4-4.1 to simulate the geometry, building layout, and environmental dose equivalent. The model treats the room itself as a controlled space: lead shielding cuts scattered dose, and the beam only comes on when no staff are present, in a day/night zoning pattern the team describes as the key operational lever.

FLASH itself is an experimental, ultra-high-dose-rate form of radiotherapy. Conventional radiotherapy delivers dose at roughly a few gray per minute; FLASH runs at hundreds or thousands of gray per second, and the biological effect on tissue appears to differ from conventional dosing in ways preclinical researchers are still mapping out. The interest is in the dose-rate effect: at FLASH intensities, normal tissue seems to tolerate radiation better than predicted, while tumor control stays the same or improves. Moving the technique from physics labs into small-animal work is the step that lets those biological questions get asked, and the question of where to put the beam is the bottleneck.

The dose arithmetic in the Italian paper is the load-bearing part. The simulation found that during daytime, with the beam on while staff are working nearby, exposure in unrestricted areas stays below 0.250 mSv per year. That is the limit set for members of the public under Italian Legislative Decree 101/2020, which transposes the European Basic Safety Standards. At night, when only the beam is on for quality-control runs, the team allows higher local dose rates because no one is in the room to receive them.

The day/night rule is operationally simple. During working hours the accelerator is in standby; staff can enter the room for animal care, cage changes, and routine husbandry. Outside working hours the beam comes on for experimental runs and quality-control checks, when the room is unoccupied and the dose outside the room is the only thing the safety case has to manage. The pattern works because dose falls off sharply with distance and the room is treated as the controlled boundary, not the building.

Workers, researchers, and members of the public all stay under 1 mSv per year, the same ceiling the decree sets for non-classified staff. Accelerator operators are classified as occupationally exposed and tracked under that rule. For the mice themselves, the team adopted a precautionary 1 mSv-per-year limit and used the simulation to verify the housing layout keeps them under it. That precautionary limit is one of the more reusable pieces of the package: a safety committee can challenge it, defend it, and adjust it.

Lead shielding is the obvious half of the design; the day/night zoning is the part that does the work no wall can. A traditional bunker absorbs dose whether or not anyone is in the building. The Italian group's pattern accepts the higher local dose during unoccupied hours and relies on the simulation to prove the occupied-hours dose stays inside the public limit. The same pattern, retuned for a different room plan and a different beam energy, is the kind of design a safety committee can actually evaluate.

Most FLASH work to date has run in physics labs or in hospitals with existing radiotherapy infrastructure. The Italian group's design is one of the first published worked examples of a safety case for installing a FLASH beam in a building not built as a radiotherapy facility, and the parameters are explicit enough to be retuned for a different room plan. The regulatory anchor, Italian Legislative Decree 101/2020, is jurisdiction-specific, and a different country would need a different reference, but the structure of the argument transfers.

The study has the limits of a single-site simulation. The geometry is the Italian group's, and a different building will have different scattering paths, so the dose numbers will move. The 9 MeV electron beam is a specific choice, and a higher-energy or different-mode accelerator would change the shielding calculation. The paper gives the next preclinical center a starting template: the day/night zoning, the lead-shielding design, the dose limits, and the regulatory anchor that makes the package defensible in front of a safety authority.

The work is published in Radiation Protection Dosimetry. The next site that wants to try FLASH now has a defensible way to ask the same siting question.