An updated variable RBE model for proton therapy.

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiobiology

Background:

• Clinical proton therapy traditionally uses a constant relative biological effectiveness (RBE) of 1.1.
• Increasingly, variable RBE models are used for treatment plan evaluation, but existing models show significant variation and lack updated in vitro data.
• There is a need for refined RBE models that incorporate current experimental findings.

Purpose of the Study:

• To compile an updated in vitro proton RBE database.
• To analyze assumptions of current phenomenological RBE models.
• To propose a new, up-to-date RBE model for clinical evaluation.

Main Methods:

• Collected a proton RBE database of 471 data points from literature, incorporating linear-quadratic model parameters and linear energy transfer (LET).
• Performed statistical analyses to validate RBE model assumptions and proposed new functions for RBEmax and RBEmin.
• Refitted existing models to the database and compared their performance against the newly developed model.

Main Results:

• Statistical analysis indicated the RBEmax function's intercept should be a free fitting parameter, leading to higher RBE estimates.
• RBEmin increased with LET, and its dependency on reference radiation fractionation sensitivity (α/βx) did not significantly improve model performance.
• The new RBE model exhibited the lowest Root Mean Square Error (RMSE) and highest R2 score, outperforming refitted existing models.

Conclusions:

• An updated RBE model, based on the largest available phenomenological model database, was successfully developed.
• The proposed model demonstrates superior performance compared to refitted published RBE models.
• This new model serves as a valuable tool for evaluating RBE effects in clinical proton therapy.