Venous tortuosity as a novel biomarker of rupture risk in arteriovenous malformations: ARI score.

BACKGROUND: Risk of rupture in arteriovenous malformations (AVMs) varies considerably among series. Hemodynamic factors, especially within the venous side of the circuit, seem to be responsible but are not yet well defined. We analyzed tortuosity in the draining vein as a potential new marker of rupture in AVMs, and propose a simple index to predict AVM bleeding. METHODS: A retrospective analysis of the venous angioarchitecture of brain AVMs was carried out at our center from 2013 to 2021, with special attention to venous tortuosity. After univariate analysis, the features of interest were combined to construct several predictive models using multivariate logistic regression. The best model proposed was the new AVM rupture index (ARI), which was then validated in an independent cohort. RESULTS: 68 AVMs were included in the first step and 32 in the validation cohort. Venous tortuosity, expressed as at least one curve >180°, was a significant predictor of rupture (p=0.023). The proposed bleeding index consisted of: venous tortuosity (any curve of >180°), single draining vein, and paraventricular/infratentorial location. It seems to be a robust evaluation tool, with an area under the receiver operating characteristic (AUROC) curve of 0.806 (95% CI 0.714 to 0.899), consistently replicated in the independent sample (AUROC 0.759 (95% CI 0.607 to 0.911)), and with an inter-rater kappa coefficient of 0.81 . CONCLUSIONS: Venous tortuosity may serve as a predictor of bleeding in AVMs that warrants further investigation. This likely new marker was one of the three elements of the proposed ARI. ARI outperformed the predictive accuracy of previous scores, and remained consistent in an independent cohort.