An arXiv preprint reports steeper survival curves under proton irradiation than under 225 kV X-rays in two laboratory cell models, a pattern the authors interpret as higher cytotoxicity. The work was designed as an initial test of whether the Bern Medical Cyclotron could support preclinical proton radiobiology. The authors conclude that conventional-dose-rate in vitro work was feasible, while saying the platform still needs refinement and validation.
How the comparison was made
To make the comparison, the team used 8.14(29) MeV protons at conventional dose rates and intended doses from 0 to 8 Gy, with 225 kV X-rays as the reference beam. The models were HaCaT human keratinocytes and B16-F10 mouse melanoma cells. The researchers measured clonogenic survival, meaning whether irradiated cells could still form colonies, and fitted the results with a linear-quadratic model. They calculated relative biological effectiveness, or RBE, by comparing the proton and X-ray doses at 0.1 survival, or 10%.
The reported advantage was model-dependent
The main numerical comparison was made at that 0.1 survival point. For B16-F10 cells, the reported RBE was 1.34. The authors described this as proton radiation being 34% more effective than X-rays and requiring a lower dose for the same survival reduction. For HaCaT cells, the RBE was 1.21, described as a moderate proton advantage. No uncertainty was reported for either RBE value, so the figures should be read as results from these tested cell lines and conditions, not as general values for every tissue or proton treatment.
One finding is internally inconsistent
The study also contains an unresolved disagreement about which model was more radiosensitive. In the Results section, HaCaT cells are described as having a steeper survival decline and being more radiosensitive than B16-F10 cells after both radiation types. The abstract says the opposite, identifying B16-F10 as the most sensitive model while listing the higher RBE for that cell line. The supplied analysis does not reconcile the two statements. That makes the overall sensitivity ranking uncertain and leaves the underlying data and fits important for follow-up.
Measurement uncertainty matters
Measurement uncertainty is another important part of the story. For a 55 mm field, the ionization-chamber calibration yielded 3.98 MGy/C. The paper reports a film-related dose-overestimation factor of 1.191(73), field uniformity of 7.7%, and overall proton-dose uncertainty of 12.8%. Because the RBE calculation compares doses at the same survival level, those calibration and uncertainty figures matter when judging how precisely the proton and X-ray responses can be compared.
A practical setup with a narrow scope
The experimental arrangement had a practical complication: in the vertical proton-flask setup, cells temporarily lost direct contact with their culture medium. Six control flasks were kept in the irradiation position for the longest exposure to quantify that effect. The clonogenic dataset came from two independent experiments with 12 flasks in total, with data points shown as standard error of the mean and horizontal error bars representing dose uncertainty. That makes this an initial platform test rather than a broad biological survey.
The conclusion stops short of clinical claims
The authors’ conclusion is correspondingly narrow. They say the Bern Medical Cyclotron can provide a foundation for future preclinical studies using in vitro proton radiobiology at conventional dose rates, but further refinement and validation are needed. The work did not establish biological FLASH or SFRT capability, which the analysis identifies as future questions rather than tested outcomes. It also cannot answer how tumors, normal tissues or whole organisms would respond, because the evidence came from two cell-line models.
The fitted curves also varied by cell line
A separate model output showed why results may depend on the cell line. The reported alpha/beta ratio was 1.09 for protons versus 3.2 for X-rays in B16-F10, but 2.58 for protons versus 1.8 for X-rays in HaCaT. Uncertainty for these ratios was not reported. The figures describe fitted survival behavior in this experiment; they do not by themselves settle the sensitivity contradiction or establish treatment rules.
Paper data and sources
Original title: Using a radiopharmaceutical production cyclotron for radiobiological research
Authors: Eva Kasanda, Jan Gruber, Pierluigi Casolaro et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-09-29
DOI: Not available
Original paper · Full text