摘要：

目的: 探讨术中平板CT对急性前循环大血管闭塞患者机械取栓术后颅内出血的预测价值。方法: 纳入2020年3月至2023年6月在蚌埠医科大学第二附属医院行机械取栓术的106例急性前循环大血管闭塞患者，均行术中平板CT扫描，记录CT值、高密度征和占位效应。根据术后是否发生颅内出血分成颅内出血组（29例）和无颅内出血组（77例），单因素和多因素Logistic回归分析筛查术后发生颅内出血的影响因素，绘制受试者工作特征（ROC）曲线，评估术中平板CT对术后颅内出血的预测效能。结果: Logistic回归分析显示，高血压是急性前循环大血管闭塞患者机械取栓术后不发生颅内出血的保护因素（OR=0.340，95% CI：0.123~0.941；P=0.038），入院时美国国立卫生研究院卒中量表评分高（OR=1.068，95% CI：1.010~1.129；P=0.021）以及术中平板CT高密度征（OR=9.942，95% CI：2.631~37.567；P=0.001）、占位效应（OR=23.564，95% CI：7.266~79.421；P=0.000）和CT值≥90 HU（OR=9.714，95% CI：1.982~47.605；P=0.005）是术后发生颅内出血的危险因素。ROC曲线显示，术中平板CT值≥90 HU、占位效应及二者联合预测术后颅内出血的曲线下面积分别为0.82（95% CI：0.734~0.889，P=0.000）、0.75（95% CI：0.661~0.833，P=0.000）和0.76（95% CI：0.668~0.838，P=0.000），灵敏度为75.86%、58.62%和58.62%，特异度为88.31%、92.21%和93.51%，尤以CT值≥90HU的预测效能最高（t=48.777，P=0.000；t=43.681，P=0.000）。结论: 急性前循环大血管闭塞患者机械取栓术中平板CT呈高密度征如CT值≥90 HU、占位效应可以预测颅内出血，且CT值≥90 HU的预测效能较高。

关键词: 缺血性卒中, 血栓切除术, 手术后并发症, 颅内出血, Logistic模型, ROC曲线, 体层摄影术, X线计算机

Abstract:

Objective: To explore the predictive value of intraoperative flat CT scan for intracranial hemorrhage after mechanical thrombectomy in patients with acute anterior circulation large vessel occlusion (AAC-LVO). Methods: A total of 106 patients with AAC-LVO who underwent mechanical thrombectomy at The Second Affiliated Hospital of Bengbu Medical University was conducted from March 2020 to June 2023. Intraoperative flat CT scan was performed among above patients to record CT values, high-density sign and mass effect. According to whether occurred intracranial hemorrhage after surgery, those patients were divided into intracranial hemorrhage group (n=29) and non-intracranial hemorrhage group (n=77). Univariate and multivariate stepwise Logistic regression analyses were used to screen the influencing factors of intracranial hemorrhage after mechanical thrombectomy. Receiver operating characteristic (ROC) curve was drawn to evaluate the predictive efficacy of intraoperative flat CT scan for postoperative intracranial hemorrhage. Results: Logistic regression analysis showed that hypertension was a protective factor for patients with AAC-LVO who did not experience intracranial hemorrhage after mechanical (OR=0.340, 95%CI: 0.123-0.941; P=0.038). A high National Institutes of Health Stroke (NIHSS) score at admission (OR=1.068, 95%CI: 1.010-1.129; P=0.021), as well as intraoperative flat CT high-density sign (OR=9.942, 95%CI: 2.631-37.567; P=0.001), mass effect (OR=23.564, 95%CI: 7.266-79.421; P=0.000), and CT value ≥ 90 HU (OR=9.714, 95%CI: 1.982-47.605; P=0.005) were risk factors for postoperative intracranial hemorrhage. The ROC curve showed that the area under the curve (AUC) for intraoperative flat CT value ≥ 90 HU, mass effect and their combined prediction of postoperative intracranial hemorrhage were 0.82 (95%CI: 0.734-0.889, P=0.000), 0.75 (95%CI: 0.661-0.833, P=0.000) and 0.76 (95%CI: 0.668-0.838, P=0.000), with sensitivity were 75.86%, 58.62% and 58.62%, and specificity were 88.31%, 92.21% and 93.51%. The predictive efficiency was highest for CT value ≥ 90 HU (t=48.777, P=0.000; t=43.681, P=0.000). Conclusions: In patients with AAC-LVO undergoing mechanical thrombectomy, intraoperative flat CT high-density sign such as CT value ≥ 90 HU and mass effect can predict postoperative intracranial hemorrhage, and the predictive value of CT value ≥ 90 HU is relatively high.

Key words: Ischemic stroke, Thrombectomy, Postoperative complications, hemorrhages, Logistic models, ROC curve, Tomography, X-ray computed