KEY POINTS
- The international expert group reviewed the clinical application of the three main carbon-ion RBE model families: the mixed beam model, the microdosimetric kinetic model, and the local effect model. Their implementations differ not only mathematically but also in reference radiation, biological parameters, scaling procedures, and clinical conventions.
- RBE is not a fixed property of carbon ions. It varies with dose per fraction, α/β ratio, LET, ion energy, biological endpoint, tissue type, oxygenation, vascularization, cell-cycle state, and potentially dose rate.
- A reported dose in Gy(RBE) does not necessarily represent the same biological or absorbed dose across centres. LEM-based systems generally calculate an RBE-weighted dose directly, whereas MBM- and modified-MKM-based systems apply an additional clinical scaling procedure.
- Even when the same absorbed-dose distribution is evaluated, different models can predict materially different target and normal-tissue effects. Conversely, optimizing an identical prescribed Gy(RBE) value with different models may produce different physical-dose gradients and hotspot locations.
- Translation between dose-prescription systems is technically feasible but cannot rely on one universal conversion factor. Conversion depends on the model pair, dose level, fractionation, field arrangement, target depth, spread-out Bragg peak width, LET distribution, modulation, and organ-specific dose-volume endpoint.
- Some centres already use dual-model assessment. MedAustron and CNAO, for example, evaluate selected plans with both LEM I and modified MKM, iteratively adjusting absorbed-dose distributions and avoiding distal high-LET regions in critical structures.
- The authors recommend reporting the complete dose-prescription system, including RBE model, parameter settings, clinical scaling conventions, absorbed dose, and relevant dose-volume metrics. Clinical outcomes cannot be pooled reliably when publications report only nominal Gy(RBE).
CLINICAL TAKEAWAY
Carbon-ion prescriptions from different institutions should not be treated as equivalent merely because the same Gy(RBE) value is reported. Multicentre trials, protocol transfer, and meta-analyses require explicit model-aware dose translation and preferably access to the underlying absorbed-dose distribution.
SOURCE
International Journal of Radiation Oncology, Biology, Physics