摘要：

研究背景: 前期研究虽确立65 ml作为前循环轻型缺血性卒中患者药物治疗预后的阈值，但其在血管内治疗（EVT）和后循环缺血性卒中患者中是否适用仍存质疑。本研究依托国际卒中灌注注册研究（INSPIRE）的更新扩展队列，纳入血管内治疗和后循环大动脉闭塞性卒中患者，重新校准预测早期神经功能恶化和远期预后不良的CT灌注成像（CTP）脑低灌注区体积最佳截断值，并探讨脑灌注高负荷患者血管内治疗的有效性与安全性。方法: 纳入INSPIRE研究更新扩展队列193例轻型大动脉闭塞性卒中患者，涵盖血管内治疗患者72例（37.31%）和后循环大动脉闭塞性卒中患者41例（21.24%），均行CTP检查，予以药物治疗或血管内治疗，以早期神经功能恶化和发病90 d重残或死亡[改良Rankin量表（mRS）评分5 ~ 6分]为临床结局，采用单因素和多因素Logistic回归分析筛查临床结局的影响因素，绘制受试者工作特征（ROC）曲线预测早期神经功能恶化和远期预后不良的CTP脑低灌注区体积最佳截断值，在模型中引入“脑低灌注区体积×治疗方式（药物治疗或血管内治疗）”的交互项，检验治疗方式对不同脑灌注负荷患者预后的影响差异。结果: 早期神经功能恶化发生率为14.51%（28/193），发病90 d重残或死亡发生率为10.38%（19/183）。Logistic回归分析显示，入院时脑低灌注区体积较大是早期神经功能恶化的危险因素（OR = 1.010，95%CI：1.000 ~ 1.020；P = 0.010），脑低灌注区体积≥ 75 ml是早期神经功能恶化（OR = 4.890，95%CI：1.890 ~ 12.610；P = 0.000）以及发病90 d重残或死亡（OR = 7.740，95%CI：2.290 ~ 26.200；P = 0.000）的共同危险因素。ROC曲线显示，入院时脑低灌注区体积预测早期神经功能恶化及发病90 d重残或死亡的曲线下面积分别为0.71（95%CI：0.610 ~ 0.820，P = 0.000）和0.72（95%CI：0.600 ~ 0.850，P = 0.000），最佳截断值均为75 ml。以75 ml为阈值分为低负荷组（< 75 ml，122例）和高负荷组（≥ 75 ml，71例），高负荷组早期神经功能恶化风险是低负荷组的4.89倍[28.17%（20/71）对6.56%（8/122），P = 0.000]，发病90 d重残或死亡风险是低负荷组的7.74倍[22.54%（16/71）对2.46%（3/122），P = 0.000]；两组症状性颅内出血风险相当（P = 0.062）。进一步将低负荷组和高负荷组分为药物组和EVT组2个亚组，无论是脑灌注低负荷还是脑灌注高负荷，药物组与EVT组患者早期神经功能恶化发生率、发病90 d神经功能预后良好（mRS评分0 ~ 2分）率和症状性颅内出血发生率差异均无统计学意义（P > 0.05）。结论: 在引入血管内治疗和后循环大动脉闭塞性卒中的广义队列中，入院时脑低灌注区体积仍是早期神经功能恶化和远期预后不良（发病90 d重残或死亡）的最强独立预测因素，最佳截断值修正为75 ml；对于脑灌注高负荷（≥ 75 ml）患者，积极行血管内治疗可能面临“获益缺失、风险增加”的困境。临床决策不应仅依赖美国国立卫生研究院卒中量表评分，还应结合CTP脑低灌注区体积进行精细化分层，警惕对高风险患者的过度干预。

关键词: 缺血性卒中, 动脉闭塞性疾病, 体层摄影术，X线计算机, 灌注成像, 预后, 血流动力学, Logistic模型, ROC曲线

Abstract:

Background: Previous studies identified 65 ml as the threshold of CT perfusion imaging (CTP) hypoperfusion lesion volume for predicting poor outcomes in medically managed patients with mild anterior circulation ischemic stroke. However, whether this threshold remains applicable to patients undergoing endovascular therapy (EVT) and those with posterior circulation ischemic stroke remains uncertain. Based on the updated and expanded cohort from the International Stroke Perfusion Imaging Registry (INSPIRE), this study included patients with EVT and posterior circulation large vessel occlusion (LVO) stroke, recalibrated the optimal CTP hypoperfusion lesion volume cutoff for predicting the early neurologic deterioration (END) and the long-term poor outcome, and explored the efficacy and safety of EVT in patients with high hypoperfusion-volume burden. Methods: A total of 193 minor stroke patients with LVO from the updated INSPIRE cohort were included. Among them, 72 patients (37.31%) received EVT and 41 patients (21.24%) had posterior circulation LVO stroke. All patients underwent CTP and received either medical therapy or EVT. The clinical outcomes were END and severe disability or death at 90 d, defined as a modified Rankin Scale (mRS) score of 5-6. Univariate and multivariate Logistic regression analyses were used to identify factors associated with clinical outcomes. Receiver operating characteristic (ROC) curves were used to determine the optimal cutoff value of CTP hypoperfusion lesion volume for predicting END and long-term poor outcome. Stratified analyses were further performed to compare clinical outcomes between medical therapy and EVT in patients with different hypoperfusion-volume burdens. Results: END occurred in 28 (14.51%) of 193 patients, and severe disability or death at 90 d occurred in 19 (10.38%) of 183 patients. Logistic regression analysis showed that larger hypoperfusion lesion volume at admission was a risk factor for END (OR = 1.010, 95%CI: 1.000-1.020; P = 0.010). Hypoperfusion lesion volume ≥ 75 ml was a common risk factor for END (OR = 4.890, 95%CI: 1.890-12.610; P = 0.001) and severe disability or death at 90 d (OR = 7.740, 95%CI: 2.290-26.200; P = 0.001). ROC curve showed that the area under the curve (AUC) of hypoperfusion lesion volume at admission for predicting END and severe disability or death at 90 d were 0.71 (95%CI: 0.610-0.820, P = 0.000) and 0.72 (95%CI: 0.600-0.850, P = 0.000), respectively. The optimal cutoff value was 75 ml for both outcomes. Using 75 ml as the threshold, patients were divided into a low hypoperfusion-volume burden group (< 75 ml, n = 122) and a high hypoperfusion-volume burden group (≥ 75 ml, n = 71). The risk of END in the high hypoperfusion-volume burden group was 4.89 times than that of the low hypoperfusion-volume burden group [28.17% (20/71) vs.6.56% (8/122), P = 0.000], and the risk of severe disability or death at 90 d in the high hypoperfusion-volume burden group was 7.74 times than that of the low hypoperfusion-volume burden group [22.54% (16/71) vs. 2.46% (3/122), P = 0.000]. There was no significant difference in the risk of symptomatic intracranial hemorrhage (sICH) in 2 groups (P = 0.062). Stratified analysis showed that, regardless of low (< 75 ml) or high (≥ 75 ml) hypoperfusion-volume burden, there were no significant differences between the medical therapy subgroup and the EVT subgroup in the incidence of END, favorable functional outcome at 90 d (mRS score 0-2), or sICH (P > 0.05, for all). Conclusions: In this expanded cohort including patients undergoing EVT and those with posterior circulation LVO stroke, hypoperfusion lesion volume at admission remained an important predictor of END and long-term poor outcome, defined as severe disability or death at 90 d. The optimal cutoff value was recalibrated to 75 ml. For patients with high hypoperfusion-volume burden (≥ 75 ml), the benefit-risk balance of EVT should be carefully evaluated. Clinical decision-making should not rely solely on the National Institutes of Health Stroke Scale (NIHSS) score, but should also incorporate CTP hypoperfusion lesion volume for refined risk stratification and avoidance of unnecessary intervention in high-risk patients.

Key words: Ischemic stroke, Arterial occlusive diseases, Tomography, X-ray computed, Perfusion imaging, Prognosis, Hemodynamics, Logistic models, ROC curve

中图分类号: 

• R741

• R651