Nanoparticle Non-Destructive Testing for Arterial Stenosis: A Nanoparticle-Assisted Blood Flow Measurement Framework

This is the authors' abstract. We don't add key points for this paper.

Accurate early detection of arterial stenosis remains challenging because subtle hemodynamic alterations are difficult to capture using conventional imaging. This paper presents a nanoparticle non-destructive testing (NP-NDT) framework for quantitative stenosis assessment based on outlet flow redistribution in vascular bifurcations. The framework combines computational fluid dynamics (CFD), benchtop experiments, and machine learning to characterize the stenosis-induced outlet asymmetry and infer the type, degree, and location of the stenosis. Non-Newtonian blood flow is modeled in Y-shaped bifurcations, and experimental validation is performed in a 3D-printed vascular phantom using particle-based flow tracing under comparable laminar-flow conditions. The results show a stronger outlet redistribution for branch-vessel stenosis than for trunk stenosis, with consistent trends between CFD and experiments. The framework achieves an AUC of 0.9136 and a mean absolute error of approximately 0.03 for estimating the degree of stenosis, while all regression tasks achieve a mean absolute error below 0.12. Uncertainty analysis further supports the repeatability of the measurement. These results demonstrate that outlet redistribution provides a physically interpretable indicator for the early detection and quantitative assessment of arterial stenosis.