目的：探讨CT灌注成像（CT perfusion，CTP）结合Rapid自动分析软件参数对前循环大血管近端闭塞 （anterior circulation large vessel occlusion，ac-LVO）急性缺血性卒中（acute ischemic stroke，AIS）血管内治疗 （endovascular treatment，EVT）患者预后的评估和预测因素分析。方法：回顾性分析ac-LVO AIS并行 EVT合 并或不合并静脉溶栓治疗的患者，收集基线特征，及CTP成像后Rapid软件进行图像后处理得到的参数：包 括脑血流量< 30%的体积（VCBF＜30%），脑血流达峰时间＞4 s（VTmax＞4 s ）、脑血流达峰时间＞6 s（VTmax＞6 s ）、脑血流 达峰时间＞8 s（VTmax＞8 s ）、脑血流达峰时间＞10 s（VTmax＞10 s ）的低灌注区体积，低灌注强度比值（hypoperfusion intensity ratio，HIR）、错配体积（mismatch Volume，MMV）、错配率（mismatch ratio，MMR）、CBV 指数，及 Alberta卒中项目早期CT评分（Alberta stroke programme early CT score，ASPECTS）等，将患者术后90 d的随访 情况按照改良Rankin量表（modified Rankin scale，mRS）评分进行分组，mRS 0～2分为预后良好组，mRS 3～ 6分为预后不良组。比较2组患者临床资料及CTP参数的差异。采用多因素Logistic回归分析来探讨与预后 相关的独立预测因素，并采用受试者工作特征（ROC）曲线评价各预测因素对应的最佳阈值及其各自预测能 力。结果：纳入患者158例，其中预后良好组 62 例，预后不良组 96 例。多因素Logistic回归分析显示，VCBF＜30% （OR=1.014，95%CI 1.001～1.026，P=0.028）、HIR（OR=8.655，95%CI 1.800～41.620，P=0.007）、入院时收缩压 （OR=1.039，95% CI 1.017～1.062，P=0.001）、入院时 NIHSS（OR=1.125，95% CI 1.038～1.219，P=0.004）是 ac-LVO卒中患者EVT后90 d预后的独立影响因素。ROC分析显示，HIR、VCBF＜30%、入院时收缩压、入院时NIHSS评分的曲线下面积分别为0.738、0.704、0.658、0.600，预测不良预后的最佳截断值分别为：HIR≥0.355， VCBF＜30%≥25.5 mL，入院时收缩压≥151.5 mmHg，入院时NIHSS评分≥17分。结论：HIR、VCBF＜30%、入院时收 缩压、入院时NIHSS评分为ac-LVO AIS患者EVT 90 d预后的独立影响因素（P＜0.05）。HIR、VCBF＜30%对于 ac-LVO AIS患者EVT预后的预测能力优于NIHSS评分、入院时收缩压。

To explore prognostic and predictive values of CT perfusion imaging (CTP) combined with Rapid automatic analysis software parameters in patients suffering from acute ischemic stroke (AIS) with anterior circulation proximal large vessel occlusion (ac-LVO) and undertaking endovascular therapy (EVT). Methods: A retrospective analysis of AIS patients with ac-LVO underwent EVT with or without intravenous thrombolytic therapy was performed. The data used for analysis were baseline characteristics and the CTP parameters post-processed by Rapid software, including the volume of cerebral blood flow<30% (VCBF<30% ), volumes of hypoperfusion with blood flow time to max>4 s, 6 s, 8 s, 10 s(VTmax>4 s, VTmax>6 s, VTmax>8 s, VTmax>10 s), hypoperfusion intensity ratio (HIR), mismatch volume (MMV), mismatch ratio (MMR), CBV index and Alberta stroke program early CT score (ASPECTS score). According to the modified Rankin scale (mRS), the data of 90-day follow-up was classified as good prognosis group with mRS of 0～2 and poor prognosis group with the mRS of 3～6. The differences in clinical data and CTP analysis parameters between the two groups were analyzed. Multivariate logistic regression was used to analyze the relevant independent predictors, the corresponding optimal thresholds of each predictor and their prediction capabilities by using the receiver operating characteristic (ROC) curve. Results: There were 158 patients enrolled, including 62 in the good prognosis group and 96 in the poor prognosis group. Multivariate Logistic regression analysis showed that VCBF<30% (OR=1.014, 95%CI 1.000～1.028, P=0.028), HIR (OR=8.655, 95%CI 1.800～41.620, P=0.007), systolic blood pressure (SBP) at admission (OR=1.039, 95%CI 1.015～1.064, P=0.001) and NIHSS at admission (OR=1.125, 95% CI 1.038～1.219, P=0.004) were independent predictors of prognosis. The analysis of ROC showed that the area under the ROC curve of HIR, VCBF<30%, systolic blood pressure at admission and NIHSS score at admission were 0.738, 0.704, 0.658 and 0.600, respectively, and that the best cut-off values for predicting adverse prognosis were HIR ≥0.355, VCBF＜30%≥25.5 mL, SBP ≥ 151.5 mmHg on admission, and NIHSS's score≥17 on admission. Conclusion: HIR, VCBF<30%, SBP at admission and NIHSS score at admission were independent factors influencing patients with ac-LVO during the follow-up for 90 days. HIR, and VCBF<30% showed advantages over NIHSS and SBP in predicting the prognosis for patients with ac-LVO.