Quantitative comparison of vessel deformation among four stent retrievers using a strain sensor equipped cerebrovascular model

BACKGROUND: Mechanical thrombectomy with stent retrievers is the standard treatment for acute large vessel occlusion stroke, but device induced vessel deformation during retrieval has not been quantified in a spatially resolved manner. METHODS: A polyvinyl alcohol (PVA) hydrogel cerebrovascular model reproducing the internal carotid artery to M1-M2 segments was equipped with eight strain sensors (TML FLA-03-11, gauge length 0.3 mm) at anatomically defined locations, corresponding to presumed perforator origins. Four stent retrievers (two sheet-type (Solitaire and pRESET) and two segmented-type (eNVi-SR and Embotrap)) were deployed and retrieved under identical conditions (three trials each). Primary outcomes were early peak strain and late mean strain. The Friedman test with post hoc Wilcoxon signed rank tests (Bonferroni correction) was used. RESULTS: The highest strain concentrations were consistently observed at the M1 outer curvature and distal perforator bearing M1 segment. Global early peak strain differed significantly among devices (P=0.025): Solitaire 56.7±5.3, pRESET 63.2±14.1, eNVi-SR 68.8±7.3, and Embotrap 84.9±1.9. Embotrap was significantly higher than Solitaire (P=0.037) and pRESET (P=0.019). For global late mean strain, Embotrap (29.3±2.9) was significantly higher than Solitaire (18.0±4.3; P=0.007). The relaxation ratio did not differ among devices (P=0.73). CONCLUSIONS: Strain sensor equipped PVA hydrogel models enabled reproducible, spatially resolved comparison of device induced vessel deformation. Embotrap induced significantly greater deformation than Solitaire and pRESET, particularly at perforator bearing segments.