Abstract TP187: Revalidation of the BAND Score

The BAND score is a previously validated predictor of stroke outcomes at 90 days for patients treated with successful thrombectomy. This score is calculated using baseline disability, age, NIHSS, and delay from last known well (LKW). Each point confers a 10% increase in poor outcome probability. The BAND score had been created to predict outcomes for anterior circulation strokes with TICI 2b or better reperfusion scores. We hypothesized that the BAND score could also accurately predict stroke outcomes for posterior circulation strokes. We performed a retrospective chart review to obtain thrombectomy data from our institution in 2024. There were 130 patients that had recorded data for 90-day mRS outcomes. Demographics for good versus poor outcomes were compared using appropriate statistical tests. Accuracy of BAND score predictions were analyzed using ROC curves with area under the curve (AUC) calculations as well as multivariable logistical regression analysis. Anterior circulation and posterior circulation strokes were analyzed to assess differences between subgroups. Patients with poor outcomes were more likely to have higher BAND score, initial NIHSS, 24-hour NIHSS, and age based on Bonferroni correction for significant p-values. Multivariable regression analysis demonstrated that Higher BAND scores and 24hr NIHSS were less likely to be associated with a good outcome at 90 days. AUC for each ROC curve (all strokes, anterior circulation, posterior circulation) was approximately 0.7 suggesting fair predictive value for the BAND score for anterior and posterior circulation strokes outcomes at 90 days. Small subgroup size did not allow for analysis of poor reperfusion scores. We found that the BAND score is a reasonable tool to predict 90-day thrombectomy outcomes. Anterior and posterior circulation strokes groups had similar predictive accuracy. BAND scores of 0-2 and 8-10 had the greatest predictive accuracy. Additional studies may be warranted to further improve accuracy for patients with BAND scores of 3-7 and to identify strong predictive factors for the 0-2 and 8-10 score cohorts.