External Validation of the R2eD AVM Score to Predict the Likelihood of Rupture Presentation of Brain Arteriovenous Malformations.

BACKGROUND The R2eD AVM scoring system has been recently proposed to predict the risk of hemorrhagic presentation of brain arteriovenous malformations (AVMs). OBJECTIVE To perform the first external validation of this R2eD AVM score. METHODS Over a retrospective 10-yr period from 2009 to 2019, 122 patients with brain AVMs fit the inclusion criteria. All components required to calculate the R2eD AVM score had to be available. Patient demographics and score calculations were recorded, as well as the circumstances of their presentation. Multivariable analysis was performed to assess predictors of rupture in the study cohort. Next, area under the receiver operating characteristics and linear-by-linear association were employed to assess the effectiveness of the models. RESULTS The mean age of the entire cohort was 45.2 yr, and 8% of patients were nonwhite. A total of 48 (39.3%) patients presented with AVM rupture. In univariable analysis, all 5 categories demonstrated P < .15 association with rupture risk. In multivariable analysis, race (nonwhite), size <3 cm, and exclusive deep venous drainage remained in the model with an area under the curve (AUC) of 0.677. The utilization of the R2eD AVM score yielded an AUC of 0.711. CONCLUSION Even though mitigated by sample size, this study demonstrated acceptable external validity of the R2eD AVM score.