Examining the Prognostic Implications of Early Digital Subtraction Angiography Signs in Patients Undergoing Anterior Circulation Thrombectomy.

Chen, Yongping; Wu, Chuyue; Chen, Shengli; Zhang, Lina; Yang, Zhenjie; He, Lei; Huang, Yu · World Neurosurg · 2025

prospective_cohort · Level II

Where this comes from

Abstract

The study aimed to evaluate the predictive value of early digital subtraction angiography (DSA) signs for hemorrhagic transformation (HT) and short-term outcomes in patients undergoing endovascular thrombectomy. Patients who underwent mechanical thrombectomy from January 2021 to December 2023 were enrolled. The correlation between DSA signs, clinical factors, and adverse outcomes within 72 hours postsurgery were analyzed. Distinct prediction models for unfavorable prognoses were formulated and the predictive parameters of each model were compared. Among 230 patients, 77 (33.5%) developed HT within 72 hours, while 114 (53.0%) exhibited poor functional outcomes (modified Rankin scale ≥ 3 points) at 90 independent predictors of HT included age, admission National Institutes of Health Stroke Scale score, balloon angioplasty during endovascular thrombectomy, early venous filling (EVF), and prior intravenous thrombolysis. Poor 90-day outcomes were independently associated with time from recanalization to puncture, admission National Institutes of Health Stroke Scale, EVF, presence of EVF or basal ganglia blush (BGB) sign, and HT (P < 0.05). Of the 4 prediction models that were developed, the prediction model that integrated baseline clinical data with digital EVF demonstrated the highest accuracy (area under the curve = 0.867; sensitivity = 0.871; specificity = 0.826) and significantly outperformed the model based solely on clinical data (area under the curve = 0.723; P < 0.05). EVF was shown to be an effective predictor for HT, while BGB had a significant correlation with the 90-day adverse prognosis. These 2 DSA indicators target "disease occurrence" and "long-term patient prognosis," respectively, in HT-related predictive models. The inclusion of EVF and BGB in prognostic models enhances precision.

Medical subject headings