'Dirty dose'-based proton variable RBE models - performance assessment on in vitro data.

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiobiology

Background:

• Clinical proton radiotherapy commonly uses a constant relative biological effectiveness (RBE) of 1.1.
• In vitro studies show variable RBE effects, leading to proposed models linking RBE parameters to dose-averaged linear energy transfer (LET).

Purpose of the Study:

• Introduce a novel variable RBE model based on the
• dirty dose
• concept.
• Simplify calculations and experimental assessments in proton therapy by focusing on dose within voxels exceeding a specific LET threshold.

Main Methods:

• Utilized published in vitro data on cell survival fraction, dose, and LET.
• Re-simulated experimental setups to extract dirty dose metrics.
• Developed and benchmarked LQ models using dirty dose fraction against conventional radiation quality metrics via RMSE.

Main Results:

• Dirty dose-based models demonstrated performance comparable to conventional metrics, with a RMSE of 0.38 for a 7 𝛾 threshold.
• Higher LET thresholds generally improved model performance.
• Models based on LET (𝛾) and dose-averaged LET (𝛾) showed RMSEs of 0.42 and 0.36, respectively.

Conclusions:

• Variable RBE models utilizing dirty dose metrics perform on par with current radiation quality metrics.
• The simplified calculations and potential for efficient experimental validation make dirty dose models a practical and conservative choice for future proton RBE modeling.