Abstract: Objective  To explore the risk factors associated with poor short-term prognosis of patients with middle cerebral artery occlusive acute ischemic stroke (MCAO-AIS) undergoing endovascular treatment and to develop a prediction model. 
Methods  Clinical, imaging, surgical data, and postoperative laboratory examination results of MCAO-AIS patients who underwent endovascular treatment at Xuzhou Central Hospital from February 2022 to May 2024 were retrospectively analyzed. Patients were divided into the good prognosis group and the poor prognosis group based on their mRS score on postoperative day 90. Factors influencing poor prognosis were explored by univariate analysis, collinearity analysis, and multivariate logistic regression analysis, and the prediction model was constructed. The predictive performance of the model was assessed and validated using the ROC curve, calibration curve, and goodness-of-fit test. 
Results  A total of 245 patients were included, with 91 in the good prognosis group and 154 in the poor prognosis group. Multivariate logistic regression analysis showed that higher NIHSS scores at the onset (OR 1.017, 95%CI 1.026-1.119, P=0.002), longer duration of surgery (OR 1.014, 95%CI 1.004-1.023, P=0.004), postoperative cerebral edema (OR 11.396, 95%CI 4.884-26.591, P<0.001), postoperative pneumonia (OR 5.609, 95%CI 2.179-14.436, P<0.001), higher levels of fibrinogen (FIB) (OR 1.877, 95%CI 1.214-2.903, P=0.005), and higher levels of S100 calcium binding protein β (S100β) (OR 1.013, 95%CI 1.004-1.022, P=0.003) were independent risk factors for poor prognosis, while the use of mechanical thrombectomy using stent retriever combined with aspiration under intracranial support catheter assistance (SWIM) combined with angioplasty (OR 0.140, 95%CI 0.045-0.429, P=0.001) was an independent protective factor. ROC curve analysis showed that the AUC of the model constructed based on the seven independent predictor variables determined by multivariate logistic regression analysis was 0.934 (95%CI 0.905-0.964), which had a good predictive efficacy. The sensitivity of the model was 0.844 and the specificity was 0.890 when the optimal cutoff value was 0.665. The Hosmer-Lemeshow test showed that the prediction model was well calibrated, and the calibration plot showed that the model prediction curve and the actual curve had a high degree of fitting. 
Conclusions  The model constructed based on the NIHSS score at the onset, operative time, postoperative cerebral edema, postoperative pneumonia, FIB, S100β, and SWIM combined with angioplasty has a good predictive efficacy for poor prognosis on day 90 after endovascular treatment in patients with MCAO-AIS.

Key words: Acute ischemic stroke; Middle cerebral artery occlusion; Endovascular treatment; Prognosis; Prediction model

摘要： 目的 探讨大脑中动脉闭塞性急性缺血性卒中（middle cerebral artery occlusive acute ischemic stroke，MCAO-AIS）患者血管内治疗短期预后不良的相关危险因素，并开发预测模型。
方法 回顾性分析2022年2月—2024年5月在徐州市中心医院接受血管内治疗的MCAO-AIS患者的临床、影像和手术资料，以及术后实验室检查结果。根据患者在术后第90天的mRS评分，将患者分为预后良好组和预后不良组。通过单因素分析、共线性分析及多因素logistic回归分析来探索患者预后不良的影响因素，并构建预测模型。采用ROC曲线、校准曲线和拟合优度检验对模型的预测性能进行评估和验证。
结果 共纳入245例患者，其中预后良好组91例，预后不良组154例。多因素logistic回归分析显示，发病时NIHSS评分较高（OR 1.017，95%CI 1.026～1.119，P=0.002）、手术时间较长（OR 1.014，95%CI 1.004～1.023，P=0.004）、术后脑水肿（OR 11.396，95%CI 4.884～26.591，P＜0.001）、术后肺炎（OR 5.609，95%CI 2.179～14.436，P＜0.001）、纤维蛋白原（fibrinogen，FIB）水平较高（OR 1.877，95%CI 1.214～2.903，P=0.005）及S100钙结合蛋白β（S100 calcium binding protein β，S100β）水平较高（OR 1.013，95%CI 1.004～1.022，P=0.003）是患者预后不良的独立危险因素，而采用颅内支撑导管辅助取栓支架联合抽吸行机械取栓术（mechanical thrombectomy using stent retriever combined with aspiration under intracranial support catheter assistance，SWIM）联合血管成形术（OR 0.140，95%CI 0.045～0.429，P=0.001）是其独立保护因素。ROC曲线分析显示，基于多因素logistic回归分析确定的7个独立预测变量构建的模型AUC为0.934（95%CI 0.905～0.964），具有较好的预测效能。模型的最佳截断值为0.665，此时模型的敏感度为0.844，特异度为0.890。Hosmer-Lemeshow检验显示预测模型校准良好，校准图显示模型预测曲线和实际曲线贴合度较高。
结论 基于发病时NIHSS评分、手术时间、术后脑水肿、术后肺炎、FIB、S100β及SWIM联合血管成形术这7个指标构建的模型对MCAO-AIS患者血管内治疗后第90天的不良预后有较好的预测效能。

关键词: 急性缺血性卒中; 大脑中动脉闭塞; 血管内治疗; 预后; 预测模型