迟发性脑缺血是动脉瘤性蛛网膜下腔出血（aSAH）患者致残及死亡的主要原因之一，早期预测并及时治疗对提高患者存活率及预后至关重要。以往研究常基于灌注成像预测aSAH后迟发性脑缺血。然而，不同后处理软件及算法的差异使得灌注成像在aSAH后迟发性脑缺血预测中的准确性存在争议，且灌注成像不能获取壁剪切应力和局部血管压力梯度等血流动力学参数。基于计算流体动力学可获得壁切应力等血流动力学参数。前期研究表明壁切应力及血管内压力梯度等参数与微血栓形成、血脑屏障的破坏及侧枝循环形成程度有关。机器学习算法可提高aSAH后迟发性脑缺血预测模型效能。因此，本项目拟基于计算流体动力学研究整体与局部基线血流动力学及早期改变与aSAH后迟发性脑缺血发生的关系；继而基于血流动力学参数及临床信息，联合机器学习构建并验证aSAH后迟发性脑缺血的早期风险预测模型。该研究预期为及时诊治迟发性脑缺血提供重要的参考价值。

Delayed cerebral ischemia (DCI) is a major cause of death and disability in patients with aneurysmal subarachnoid hemorrhage (aSAH). Early prediction and timely treatment are essential to improve the survival rate and prognosis of patients with aSAH. Previous studies were often based on perfusion imaging to predict DCI after aSAH. However, due to the differences of post-processing software and algorithms, the accuracy of perfusion imaging in predicting DCI after aSAH is controversial. Moreover, perfusion imaging cannot obtain wall shear stress (WSS) and pressure gradient. Hemodynamic parameters, such as WSS, can be obtained by computational fluid dynamics (CFD). Previous studies showed that WSS and intravascular pressure gradient were related to microthrombosis, blood-brain barrier damage and collateral circulation. Machine learning algorithm can improve the prediction model of DCI after aSAH. Therefore, our study intends to explore the potential relationship between global and local baseline hemodynamics and early changes and DCI after aSAH based on CFD. And then developing and validating the individualized prediction model for risk of DCI after aSAH based on hemodynamics, routine clinical information and machine learning. The goal is to provide an important reference value for the timely diagnosis and treatment for DCI after aSAH.