Internal Carotid Artery Tortuosity: an Incremental Predictor of Poor Functional Outcome After Mechanical Thrombectomy

BACKGROUND: Preoperative prognostic prediction for patients undergoing mechanical thrombectomy (MT) remains suboptimal. Several studies have explored the association between internal carotid artery (ICA) tortuosity and post-MT functional outcomes with inconsistent results, and whether ICA tortuosity provides incremental prognostic value beyond conventional preoperative predictors remains unexplored. METHODS: We retrospectively analyzed 338 consecutive patients receiving MT for isolated middle cerebral artery (MCA) occlusion between January 2019 and December 2024. ICA tortuosity was qualitatively assessed on preprocedural CTA using standard post-processing workstations and classified into four subgroups: non-tortuosity, extracranial ICA-only, cavernous ICA-only, and both-tortuosity. The primary outcome was major disability or death (modified Rankin Scale [mRS] score 4-6) at discharge, and principal secondary outcome was major disability or death at 3 months. Additional secondary outcomes included favorable functional outcome (mRS 0-2), successful reperfusion (final eTICI ≥2b), angiography-to-recanalization time (ART), and hemorrhagic complications. Multivariable logistic regression assessed associations between tortuosity type and outcomes. Incremental predictive value was evaluated using AUC, NRI, and IDI. Mediation analyses were performed to explore potential pathways. RESULTS: The Both-tortuosity group was independently associated with the primary outcome (65.9% vs 30.6%; adjusted OR = 4.384, 95% CI: 1.790-11.193, P = 0.001), and was similarly associated with 3-month major disability or death (adjusted OR = 5.141, 95% CI 1.954-14.077, P = 0.001). Adding ICA tortuosity type to the conventional preoperative model did not significantly improve the AUC (from 0.731 to 0.756, P = 0.109), but yielded significant risk reclassification improvements (NRI 14.1%, P < 0.001; IDI 3.8%, P < 0.001); the NRI remained significant after bootstrap internal validation, and consistent reclassification benefit was observed across four additional conventional models. Mediation analysis revealed ART mediated 29.8% of the overall tortuosity-outcome effect (P = 0.004) and 39.0% of extracranial ICA tortuosity (P = 0.021), while cavernous ICA tortuosity effect was mediated by successful reperfusion (28.7%, P = 0.022). CONCLUSIONS: ICA tortuosity type is an independent predictor of poor early functional outcomes after MT for isolated MCA occlusion and provides incremental prognostic value in terms of risk reclassification. Tortuosity at different ICA segments may exert effects through distinct pathways.