LETd-dependent RBE conversion between LEM1- and mMKM-based dose systems and characterisation of LETd-dependent uncertainty in carbon ion radiotherapy
In brief
LET-based conversion predicts carbon-ion dose differences within 3% to 4%
A new dose-averaged linear energy transfer framework converted relative biological effectiveness between two carbon-ion dose systems, predicting their difference within 1% when input differences were known and within 3% to 4% when estimated from LET. The analysis also found uncertainty in the model's predictions exceeded 100% in low-dose, low-LET regions, underscoring the need to account for uncertainty when evaluating treatment plans.
- Journal
- Physics in medicine and biology (Q1)
- Published
- 20 November 2026
- Study design
- Unclassified
- Evidence level
- Level 5, Expert Opinion (CEBM 5)
- Authors
- Yihan Jia, Martina Favaretto, Giuseppe Magro, Markus Stock, Piero Fossati, Dietmar Georg, et al.
- PMID
- 43063606
- DOI
- 10.1088/1361-6560/aec52e
Why clinicians should know about it
- Picked for Medical Physics (top studies of the week, 22 November 2026): Carbon-ion LET-dependent RBE conversion and uncertainty modeling
Abstract
Differences between LEM1- and mMKM-based clinical dose systems complicate the translation of RBE-weighted dose in carbon ion radiotherapy, while uncertainties associated with RBE model parameterisation complicate the interpretation of predicted RBE. This work establishes two complementary dose-averaged LET (LETd)-based frameworks to (i) analytically convert RBE between the two systems and (ii) characterise systematic fitting uncertainties and their propagation into RBE.
Methods. Monte Carlo simulations of clinical carbon‑ion pencil beams and SOBPs in water were performed using GATE 10/Geant4.11.3.2 within the IDEAL 2 platform. LETd, mixed-beam radiosensitivity αmix, and RBE distributions were computed for both clinical dose systems. Ion-specific αZ,Etables were obtained from saturation-corrected dose-averaged specific energy z*1Dand intrinsic coefficient αztables computed by internally developed PRUDENT code. Relationship between εαand εRBE, which quantify the mMKM-LEM1 relative differences in αmixand RBE, respectively, was derived and approximated using linear functions. The εα(LETd) relationship was fitted from simulated αmixand LETd. The full analytical framework was validated in SOBP target regions against computed εRBE. Systematic fitting uncertainty in mMKM, denoted as ηRBE, was characterized by comparing simulated αmixwith published in vitro αfitreference at equal LETd, and evaluated for physical dose D=0-8 Gy.
Main results. The εαranged from -20% at LETd=33 keV/μm to 18% at LETd=95 keV/μm in the target region of SOBP, well described by a logarithmic-logistic form. The linear approximation εRBE=k·εα+b predicted the εRBEwithin ±1% when εαwas known, and within 3-4% when fitted εα(LETd) was used. Large ηαwas observed, reaching 275% at LETd=22.5 keV/μm and decreasing to -22% at LETd=137 keV/μm. The resulting ηRBEexceeded 100% in the low-D-low-LETdregion.
Significance. This work provides the first closed-form, LETd-based frameworks for RBE conversion in LEM1- and mMKM-based clinical systems and LETd-related uncertainty quantification in the latter. The results reveal the mechanistic link between LETdand εαand between LETdand ηαarising from particle LET spectra, and support LETd-informed plan evaluation.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.