Monte Carlo-based prediction models for nanodroplet-mediated proton range verification

Objective.Radiation-sensitive nanodroplets were recently proposed forin vivoproton range verification. Specifically, the ultrasound contrast generated by radiation-induced droplet vaporization enables visualization of radiation deposition. Here, we developed and validated anin silicomodel to predict the nanodroplet radiation response from the original treatment plan, ultimately enabling contrast-mediated treatment verification.Approach.By combining the theory of radiation-induced nucleation of superheated emulsions with Monte Carlo simulations, we determined the effective fluence of ionizing particles contributing to the nanodroplet vaporization response. This fluence we assumed to be linearly proportional to the number of vaporization events. The resulting model performance was compared to experimental observations in phantoms and healthy rats. Main results.The modeled effective fluence closely mimicked the vaporization distributions in phantoms, with a sub-millimeter range retrieval performance at body temperature. We also demonstrated the possibility to apply the model directly to preclinical computed tomography scans, yielding promising qualitative agreement with the radiation-induced contrast.Significance.The presented prediction model forms an essential step towards clinical implementation of nanodroplet-mediated proton range verification. Moreover, the simplicity and generalizability of the model enables its use for a wide range of radiotherapy applications.