A Sensitivity Study on the Effect of Voxel Human Model Deformation and Radionuclide Accumulation for Internal Dose Assessment in Nuclear Emergency

Current internal dose assessments in nuclear emergencies rely on static, upright voxel phantoms, often neglecting realistic human postures and physiological factors—such as breathing rates specific to emergency scenarios—that influence radionuclide intake and biokinetics. We present a voxel deformation method based on an improved as-rigid-as-possible (ARAP) algorithm incorporating a novel smoothing term to generate anatomically consistent stooping and swivelling models. Coupled with Geant4 Monte Carlo simulations using the full decay spectra of radionuclides relevant to simulated nuclear accident scenarios (i.e., 131I and 137Cs), and incorporating scenario-specific respiratory parameters into activity calculations, our results demonstrate that body posture significantly influences internal dose distributions: for 137Cs, the specific absorbed fraction (SAF) of the liver increases by up to 24.9% in the stooping posture, while swivelling induces variations of up to 15.1%. In contrast, dose metrics for 131I show minimal sensitivity to posture (<5%). These findings highlight the importance of incorporating realistic postures and context-aware physiological parameters into emergency dosimetry. Our method enables behaviorally realistic internal dose reconstruction and provides a robust foundation for integrating human motion and respiratory data into rapid triage and risk assessment protocols.