SELECT2 (Randomized Controlled Trial to Optimize Patient’s Selection for Endovascular Treatment in Acute Ischemic Stroke) was a phase 3, international, randomized, open-label clinical trial with adaptive enrichment design and blinded end-point assessment.¹⁷ An academic steering committee oversaw the conduct of the trial. The first author designed the trial and wrote the first draft of the manuscript, and three of the authors performed the data analyses. The trial protocol (available with the full text of this article at NEJM.org) was approved by the local institutional review board at each participating site before enrollment began. All enrolled patients or their legally authorized representatives provided written informed consent. The trial was conducted in accordance with the principles of the Declaration of Helsinki and the International Council for Harmonisation Good Clinical Practice guidelines. Adverse-event monitoring and adjudication was performed by an independent medical safety monitor. The data and safety monitoring board (which was composed of independent vascular neurologists, neurointerventionalists, and statisticians) oversaw patient safety and the interim analyses. The authors vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol.
SELECT2（Randomized Controlled Trial to Optimize Patient's Selection for Vesselscular Treatment in Acute Ischemic Stroke）是一项 3 期、国际、随机、开放标签临床试验，采用适应性富集设计和盲法终点评估。 ¹⁷ 一个学术指导委员会监督了试验的进行。第一作者设计了试验并撰写了手稿的初稿，其中三位作者进行了数据分析。试验方案（本文全文见 NEJM.org）在入组开始前已由每个参与地点的当地机构审查委员会批准。所有入组患者或其合法授权代表均提供书面知情同意书。该试验是根据《赫尔辛基宣言》和国际协调委员会良好临床实践指南的原则进行的。不利事件监测和裁决由独立的医疗安全监测员进行。数据和安全监测委员会（由独立的血管神经学家、神经介入学家和统计学家组成）负责监督患者安全和中期分析。作者保证数据的准确性和完整性以及试验对方案的保真度。

The trial was funded by a grant from Stryker Neurovascular to the University Hospitals Cleveland Medical Center and University of Texas Houston McGovern Medical School. Stryker Neurovascular did not provide trial equipment, and the design of the trial did not mandate the use of Stryker Neurovascular products. Stryker Neurovascular had no role in the design or execution of the trial, the analysis of the data, or the writing or reviewing of the manuscript. Confidentiality agreements were in place between University Hospitals Cleveland Medical Center (one of the participating centers) and Stryker Neurovascular.
该试验由Stryker Neurovascular 向大学医院克利夫兰医学中心和德克萨斯大学休斯顿麦戈文医学院资助。Stryker Neurovascular 没有提供试验设备，试验设计也没有强制要求使用 Stryker Neurovascular 产品。Stryker Neurovascular 在试验的设计或执行、数据分析、手稿的撰写或审查中没有任何作用。克利夫兰大学医院医学中心（参与中心之一）与史赛克神经血管公司之间签订了保密协议。

The trial was conducted at 31 sites across the United States, Canada, Europe, Australia, and New Zealand. Eligible patients were 18 to 85 years of age and had acute ischemic stroke due to occlusion of the internal carotid artery (either cervical or intracranial) or the M1 segment (main trunk) of the middle cerebral artery or both. All patients underwent a standardized imaging evaluation with noncontrast CT and, depending on site preference, either CT perfusion imaging or diffusion-weighted MRI. Eligible patients had to have a large ischemic core on noncontrast CT (defined as an ASPECTS value of 3 to 5; on a scale from 0 to 10, with lower values indicating larger infarction) or an estimated ischemic-core volume of 50 ml or greater on CT perfusion imaging (defined as a relative cerebral blood flow of <30% as determined with RAPID automated software [iSchemaView], which was installed at each trial site). For sites that used noncontrast CT and diffusion-weighted MRI as the standard for baseline imaging, the apparent diffusion coefficient value was used to determine the ischemic-core volume. A large ischemic core was considered to be a lesion with a volume of 50 ml or greater with an apparent diffusion coefficient value of less than 620×10⁻⁶ mm² per second as determined with the use of RAPID software. There was no upper limit for ischemic-core volume. The ischemic core volume for patients enrolled with CT perfusion was based on automated measurement of cerebral blood flow volume. It was appreciated during the trial that in some cases, the region of hypodensity on noncontrast CT was larger than the CT perfusion core volume. Some ischemic core volumes were therefore substituted by planimetric assessment of the volume of hypodensity on noncontrast CT by core laboratory personnel who were unaware of the treatment assignments.
该试验在美国，加拿大，欧洲，澳大利亚和新西兰的31个地点进行。符合条件的患者年龄在 18 至 85 岁之间，由于颈内动脉（颈动脉或颅内动脉）或大脑中动脉 M1 段（主干）闭塞或两者兼而有之而发生急性缺血性卒中。所有患者均接受标准化影像学评估，包括非增强 CT，并根据部位偏好进行 CT 灌注成像或弥散加权 MRI。符合条件的患者必须在非增强 CT 上具有较大的缺血核心（定义为 ASPECTS 值为 3 至 5;在 0 到 10 的范围内，较低的值表示更大的梗死）或 CT 灌注成像估计的缺血核心体积为 50 ml 或更大（定义为相对脑血流量为 <30%，由安装在每个试验地点的 RAPID 自动化软件 [iSchemaView] 确定）。对于使用非增强 CT 和弥散加权 MRI 作为基线成像标准的部位，表观弥散系数值用于确定缺血核心体积。大缺血核心被认为是体积为 50 ml 或更大的病变，其表观扩散系数值小于 620×10 ⁻⁶ mm ² /s，如使用 RAPID 软件确定。缺血核心体积没有上限。接受 CT 灌注的患者的缺血核心容积基于脑血流量的自动测量。在试验过程中，人们发现，在某些情况下，非增强 CT 上的低密度区域大于 CT 灌注核心体积。 因此，一些缺血性核心体积被不了解治疗任务的核心实验室人员在非增强 CT 上对低密度体积的平面评估所取代。

Endovascular thrombectomy was expected to begin within 24 hours after the onset of stroke. Other eligibility criteria included a prestroke score on the modified Rankin scale of 0 or 1 (indicating no disability), ascertained at the time of randomization, and no evidence of intracranial hemorrhage on neuroimaging. Scores on the modified Rankin scale range from 0 to 6, with higher scores indicating greater disability. Intravenous thrombolytic drugs (alteplase or tenecteplase) were administered to eligible patients who were first assessed within 4.5 hours after the onset of stroke. A screening log was kept at all trial sites for patients who had a large-vessel occlusion and met the criteria for a large ischemic core on any imaging method but were not enrolled in the trial; the log included documentation of the reason for exclusion.
血管内血栓切除术预计将在卒中发作后 24 小时内开始。其他合格标准包括在随机分组时确定的改良 Rankin 量表的卒中前评分为 0 或 1（表示没有残疾），并且神经影像学上没有颅内出血的证据。修改后的 Rankin 量表的分数范围为 0 到 6，分数越高表示残疾程度越高。对在卒中发作后 4.5 小时内首次接受评估的符合条件的患者给予静脉溶栓药物（阿替普酶或替奈普酶）。在所有试验地点都保存了筛查日志，以检查患有大血管闭塞且符合任何成像方法上大缺血核心标准但未参加试验的患者;该日志包括排除原因的文档。

Patients were randomly assigned in a 1:1 ratio to undergo endovascular thrombectomy and receive standard medical care or to receive standard medical care alone. Randomization was performed with the use of a central, Web-based module with an adaptive (minimization) procedure to balance groups across key clinical and imaging characteristics: age, National Institutes of Health Stroke Scale (NIHSS) score at presentation (with scores ranging from 0 to 42 and higher scores indicating worse neurologic deficits), occlusion location, time window (the interval beween the time that the patient was last known to be well and randomization), ischemic-core volume estimate, ASPECTS value, presence or absence of target perfusion–diffusion mismatch profile (mismatch ratio [the ratio of critically hypoperfused tissue to the ischemic-core estimate] of ≥1.8 with a mismatch volume [the volumetric difference between critically hypoperfused tissue and the ischemic-core estimate] of ≥15 ml), affected brain hemisphere, and participating center.
患者以 1：1 的比例随机分配接受血管内血栓切除术并接受标准医疗护理或仅接受标准医疗护理。使用基于 Web 的中央模块进行随机化，该模块具有自适应（最小化）程序，以平衡关键临床和影像学特征的组：年龄、就诊时的美国国立卫生研究院卒中量表 （NIHSS） 评分（分数范围为 0 到 42，分数越高表示神经功能缺损越严重）、闭塞位置、时间窗口（患者最后一次已知良好和随机化的时间间隔）， 缺血核心体积估计值、ASPECTS 值、是否存在目标灌注-弥散错配曲线（错配比 [严重灌注不足组织与缺血核心估计值之比] ≥1.8，错配体积 [严重灌注不足组织与缺血核心估计值之间的体积差异] 为 ≥15 ml）、受影响的脑半球和参与中心。

Endovascular thrombectomy was performed with stent retrievers, aspiration devices, or both of various manufacturers, depending on trial site. All the patients received medical care according to institutional protocols in accordance with the guidelines of the American Heart Association–American Stroke Association, European Stroke Organization, and the Stroke Foundation (Australia and New Zealand), including guidelines regarding blood-pressure management, critical care, and in-hospital and outpatient rehabilitation.^9-11 Decisions to proceed with decompressive hemicraniectomy in patients with severe brain swelling were made in accordance with local practices, which could include intravenous thrombolytic and oral antiplatelet agents. Patients with tandem occlusions or isolated cervical internal carotid artery occlusions were allowed in the trial; details regarding treatment guidance for these occlusions are provided in the protocol.
血管内血栓切除术使用支架取回器、抽吸装置或不同制造商的两者进行，具体取决于试验地点。所有患者均根据美国心脏协会-美国卒中协会、欧洲卒中组织和卒中基金会（澳大利亚和新西兰）的指南接受医疗护理，包括有关血压管理、重症监护以及院内和门诊康复的指南。 ^9-11 根据当地惯例，决定对严重脑肿胀患者进行减压半颅切除术，其中可能包括静脉溶栓和口服抗血小板药物。试验允许串联闭塞或孤立性颈内动脉闭塞的患者;有关这些闭塞的治疗指南的详细信息，请参阅方案。

The primary outcome was the ordinal score on the modified Rankin scale at 90 days. Scores of 6 (indicating death) and 5 (indicating that the patient is bedridden and constant care is needed) were merged for purposes of the analysis to avoid considering a shift from a score of 6 to 5 as a substantial improvement in functional status. The initial plan included, at the request of regulatory agencies, a modified Rankin scale score of 0 to 2 as a primary outcome; however, the power analysis was based on a single primary outcome, and a modified Rankin scale score of 0 to 2 was changed to be the first secondary outcome in the previously published trial protocol.¹⁷
主要结局是 90 天时改良 Rankin 量表的序数评分。为了分析的目的，将 6 分（表示死亡）和 5 分（表示患者卧床不起，需要持续护理）合并，以避免将 6 分转变为 5 分作为功能状态的实质性改善。应监管机构的要求，初始计划包括修改后的 Rankin 量表评分 0 至 2 作为主要结局;然而，功效分析基于单一的主要结局，并且修改后的 Rankin 量表评分 0 至 2 被更改为先前发表的试验方案中的第一个次要结局。 ¹⁷

Secondary outcomes were functional independence (a score on the modified Rankin scale of 0 to 2) at 90 days after randomization (with a window of ±15 days), independent ambulation (a score on the modified Rankin scale of 0 to 3) at 90 days after randomization (with a window of ±15 days), procedural complications, successful reperfusion (defined as grade 2b to 3 on the modified Treatment in Cerebral Ischemia Scale; range, 0 to 3, with higher grades indicating increased reperfusion [grade 2b indicates reperfusion of ≥50% of the occluded middle cerebral artery territory and grade 3 indicates reperfusion of 100% of the occluded middle cerebral artery territory at the end of the thrombectomy procedure]), discharge location, early neurologic improvement (defined as a reduction of ≥8 points in the NIHSS score from baseline to 24 hours after presentation or a score of 0 to 1), and quality of life as assessed with the use of domain-specific Neuro-QoL measures (mobility, depression, social participation, and cognitive aspects) transformed to respective T scores (with higher scores indicating better performance in a given domain, except for depression, for which higher scores indicate worse performance). Safety outcomes included symptomatic intracranial hemorrhage,¹⁸ death, neurologic worsening (an increase of ≥4 points in the NIHSS score within 24 hours after presentation), and procedural complications. Additional details regarding outcome definitions are provided in the protocol.
次要结局是随机分组后90天（窗口期为±15天）的功能独立性（改良Rankin量表评分为0至2分）、随机分组后90天（窗口期为±15天）的独立行走（改良Rankin量表评分为0至3分）、手术并发症、成功再灌注（定义为改良脑缺血量表治疗的2b至3级; 0 至 3 级，较高等级表示再灌注增加 [2b 级表示再灌注 ≥50% 的闭塞大脑中动脉区域，3 级表示血栓切除术结束时 100% 的闭塞大脑中动脉区域再灌注]）、出院位置、早期神经系统改善（定义为 NIHSS 评分从基线到就诊后 24 小时降低 ≥8 分或评分为 0 至 1 分）， 以及使用特定领域的 Neuro-QoL 测量（流动性、抑郁、社会参与和认知方面）评估的生活质量，转化为相应的 T 分数（分数越高表示在给定领域表现更好，抑郁症除外，抑郁症的分数越高表示表现越差）。安全性结局包括有症状的颅内出血、 ¹⁸ 死亡、神经系统恶化（就诊后24小时内NIHSS评分增加≥4分）和手术并发症。方案中提供了有关结果定义的其他详细信息。

Prespecified subgroups were defined according to age, NIHSS score at presentation, time to randomization, mismatch profile, affected hemisphere, occlusion site, and geographic location (U.S. vs. non-U.S. sites). The subgroup analyses assessed the odds that the trial patients in the thrombectomy group would have better functional recovery at 90 days than patients assigned to the medical-care group, but the trial was not powered to allow conclusions from these data.
根据年龄、就诊时的 NIHSS 评分、随机化时间、错配概况、受影响的半球、闭塞部位和地理位置（美国与非美国部位）定义预先指定的亚组。亚组分析评估了血栓切除术组的试验患者在90天时比分配到医疗护理组的患者有更好的功能恢复的几率，但该试验无法从这些数据中得出结论。

Clinical assessments were performed at baseline and at 24 hours, 5 to 7 days (or at discharge, if discharge occurred earlier), 30 days, and 90 days after randomization. Trained, certified assessors who were unaware of trial-group assignments and imaging results collected 30-day and 90-day outcomes. The score on the modified Rankin scale was assessed at in-person visits or by means of telephone interviews¹⁹ with the patient or a surrogate.
临床评估在基线和随机分组后 24 小时、5 至 7 天（或出院时，如果出院发生得更早）、30 天和 90 天进行。训练有素的认证评估员不知道试验组的分配和成像结果，收集了30天和90天的结果。改良的 Rankin 量表的分数是在面对面访问或通过与患者或代理人的电话访谈 ¹⁹ 进行评估的。

All patients underwent noncontrast CT, CT angiography or magnetic resonance angiography, and CT perfusion imaging or perfusion–diffusion MRI, with automated processing with the use of RAPID software at baseline to assess trial eligibility. All imaging (including imaging performed at baseline, angiography performed during thrombectomy, and follow-up neuroimaging) was reviewed by the central imaging core laboratory at the University of Texas McGovern Medical School and the University of Melbourne for central adjudication of ASPECTS values, occlusion location, perfusion measures, angiographic procedure success, hemorrhages, and follow-up infarct volumes. The imaging methods are described in the Supplementary Appendix, available at NEJM.org.
所有患者均接受非增强CT、CT血管造影或磁共振血管造影，以及CT灌注成像或灌注弥散MRI，并在基线时使用RAPID软件进行自动处理，以评估试验资格。德克萨斯大学麦戈文医学院和墨尔本大学的中央影像学核心实验室对所有影像学检查（包括基线时进行的影像学检查、血栓切除术期间的血管造影术和随访神经影像学检查）进行审查，以集中裁定 ASPECTS 值、闭塞位置、灌注措施、血管造影手术成功率、出血和随访梗死体积。成像方法在补充附录中进行了描述，可在 NEJM.org 上获得。

The original trial design included two prespecified interim analyses, which would be performed when 200 patients and 380 patients enrolled in the trial had completed the 90-day follow-up. The analyses included an efficacy boundary (z score of >2.604) and a futility boundary (z score of ≤1.897). The data and safety monitoring board reviewed the results of the first interim analysis of the data for 200 patients and recommended that the trial continue. In light of the publication of the RESCUE-Japan LIMIT trial,¹⁴ which showed the efficacy of endovascular thrombectomy in patients with a large stroke who had been selected for enrollment primarily on the basis of MRI results, the board requested to review the data for the efficacy and safety outcomes after 300 patients enrolled in the trial had completed the 90-day follow-up. The review by the data and safety monitoring board of these data showed that the prespecified efficacy boundary had been crossed in favor of endovascular thrombectomy. Therefore, on September 9, 2022, the board recommended that enrollment be stopped. The statistical analysis plan was finalized on November 20, 2022, before the outcome data was released by the independent data management core on November 21, 2022.
最初的试验设计包括两项预先指定的中期分析，这些分析将在200名患者和380名参加试验的患者完成90天随访时进行。分析包括疗效边界（z 评分 >2.604）和无效边界（z 评分 ≤1.897）。数据和安全监测委员会审查了对200名患者数据的第一次中期分析的结果，并建议继续进行试验。鉴于 RESCUE-Japan LIMIT 试验的发表， ¹⁴ 该试验显示了血管内血栓切除术对主要根据 MRI 结果被选中入组的大脑卒中患者的疗效，委员会要求在 300 名参加试验的患者完成 90 天随访后审查疗效和安全性结果的数据。数据和安全监测委员会对这些数据的审查表明，已经越过了预先设定的疗效界限，有利于血管内血栓切除术。因此，董事会于 2022 年 9 月 9 日建议停止注册。统计分析计划于2022年11月20日敲定，结果数据于2022年11月21日由独立数据管理核心发布。

On the basis of data from the Optimizing Patient Selection for Endovascular Treatment in Acute Ischemic Stroke (SELECT) prospective cohort study,¹⁵ the trial was designed to have a maximum sample of 560 patients and 90% power at a two-sided alpha level of 5% to detect a mean standardized difference of 0.34 for a shift across outcomes on the modified Rankin scale between groups. The primary analysis was performed on the basis of the intention-to-treat principle and used a two-sided Wilcoxon–Mann–Whitney test for superiority to assess the distribution of scores on the modified Rankin scale at the 90-day follow-up. The effect size was determined with the use of the Wilcoxon–Mann–Whitney measure (probability) of superiority and assumption-free generalized odds ratios with 95% confidence intervals. A Wilcoxon–Mann–Whitney probability of superiority value greater than 0.5 and a generalized odds ratio greater than 1 indicated improvement of outcomes in the thrombectomy group.^20,21 The analyses were conducted to control the overall two-sided type I error rate at 0.05, accounting for a cumulative alpha of 0.018 spent at the interim analyses.
根据优化急性缺血性卒中血管内治疗的患者选择 （SELECT） 前瞻性队列研究的数据， ¹⁵ 该试验被设计为在 5% 的双侧 α 水平下的最大样本为 560 名患者和 90% 的功效，以检测平均标准化差异 0.34 对于改良的 Rankin 量表在组间结果的变化。初步分析基于意向治疗原则进行，并使用双侧 Wilcoxon-Mann-Whitney 优效性检验来评估 90 天随访时改良 Rankin 量表的分数分布。使用 Wilcoxon-Mann-Whitney 优效度量（概率）和无假设的广义比值比和 95% 置信区间来确定效应量。Wilcoxon-Mann-Whitney 优效概率值大于 0.5 且广义比值比大于 1 表明血栓切除术组的结局有所改善。 ^20,21 进行分析是为了将整体双侧 I 型错误率控制在 0.05，占中期分析中花费的累积 alpha 为 0.018。

Because the analyses used an assumption-free generalized odds ratio, no test of proportionality was performed. All effects for binary secondary and safety outcomes were estimated as relative risks and are reported with respective 95% confidence intervals. Because there was no prespecified plan for the adjustment of the widths of confidence intervals for the secondary outcomes or for multiple comparisons in the subgroup analyses, the confidence intervals should not be used for hypothesis testing. Missing data were imputed with the use of multiple imputation under the missing-at-random assumption with chained equations. Additional details regarding statistical analyses and handling of missing data are provided in the statistical analysis plan (available with the protocol) and in the Supplementary Appendix.
由于分析使用了无假设的广义比值比，因此未进行相称性检验。二元、次要和安全性结局的所有影响均被估计为相对风险，并以相应的 95% 置信区间进行报告。由于没有预先指定的计划来调整次要结局的置信区间宽度或亚组分析中的多重比较，因此置信区间不应用于假设检验。在随机缺失假设下，使用链式方程的多重插补来插补缺失数据。有关统计分析和处理缺失数据的其他详细信息，请参阅统计分析计划（随协议提供）和补充附录。

From September 2019 through September 2022, at the time the trial was stopped, 958 patients had been screened; 352 of these patients were eligible, provided informed consent, and were enrolled. The main reasons for exclusion were multiple or ineligible clot locations, ineligible age of the patient, a time since the onset of stroke of more than 24 hours, or the presence of imaging abnormalities in addition to the cerebral infarction. A total of 178 patients were assigned to the thrombectomy group, and 174 were assigned to the medical-care group (Fig. S1 in the Supplementary Appendix). Approximately 60% of the thrombectomy procedures were performed with the patient under general anesthesia. A total of 3 patients were lost to follow-up before 90 days, and 1 patient withdrew consent at the 90-day follow-up. Two patients crossed over from medical care to endovascular thrombectomy; thrombectomy was attempted in all the patients who were assigned to the thrombectomy group. Eleven patients with an ASPECTS value greater than 5 as adjudicated by the core laboratory and an ischemic-core volume of less than 50 ml who were enrolled on the basis of a site-assessed ASPECTS value of 3 to 5 were excluded from the per-protocol analysis.
从 2019 年 9 月到 2022 年 9 月，在试验停止时，已经筛查了 958 名患者;这些患者中有 352 名符合条件，提供了知情同意书，并被纳入了组。排除的主要原因是多个或不合格的凝块位置、患者年龄不合格、卒中发作时间超过 24 小时，或除脑梗死外还存在影像学异常。血栓切除组共178例，内科护理组174例（补充附录图S1）。大约 60% 的血栓切除术是在全身麻醉下对患者进行的。共有 3 例患者在 90 天前失访，1 例患者在 90 天随访时撤回同意。2例患者从医疗护理过渡到血管内血栓切除术;所有被分配到血栓切除组的患者都尝试了血栓切除术。11 名 ASPECTS 值大于 5 且缺血核心体积小于 50 ml 的患者根据现场评估的 ASPECTS 值为 3 至 5 入组，被排除在符合方案的分析之外。

Baseline demographic, clinical, and imaging characteristics were similar in the two trial groups (Table 1). The median age was 66.5 years (interquartile range, 58 to 75), the median NIHSS score was 19 (interquartile range, 15 to 23), the median interval between the time the patient was last known to be well and randomization was 9.31 hours (interquartile range, 5.66 to 15.33), the median ASPECTS value was 4 (interquartile range, 3 to 5). In 48 of 352 patients, planimetric measurement of noncontrast CT hypodensity by the core laboratory was used to estimate ischemic core volume rather than the prespecified method; 6 patients fulfilled the prespecified reason of suboptimal perfusion and 42 had underestimated core volume based on CT perfusion, as judged by the core laboratory. The mean volumes for each group reported here are from the prespecified method of measurement and are smaller than if the substituted planimetric method had been used. The median estimated ischemic-core volume was 74 ml (interquartile range, 50 to 111.5) in the thrombectomy group and 77 ml (50.3 to 104) in the medical-care group. A total of 145 patients (41.2%) were women, and 50 patients (14.2%) were Black. A total of 97 of 339 patients (28.6%) awoke with symptoms of stroke, and 211 of 352 patients (59.9%) were transferred from outside hospitals or emergency departments to participating trial centers. Intravenous thrombolysis was administered in 37 patients (20.8%) in the thrombectomy group and in 30 patients (17.3%) in the medical-care group; alteplase was used in 32 of the 37 patients in the thrombectomy group and in 28 of the 30 patients in the medical-care group. The representativeness of the trial population to patients with acute ischemic stroke in the countries in which the trial was conducted is provided in Table S13.
两个试验组的基线人口统计学、临床和影像学特征相似（表1）。中位年龄为 66.5 岁（四分位距，58 至 75 岁），NIHSS 评分中位数为 19（四分位距，15 至 23 分），患者最后一次已知健康与随机分组之间的中位间隔为 9.31 小时（四分位距，5.66 至 15.33），中位 ASPECTS 值为 4（四分位距，3 至 5）。在 352 例患者中的 48 例中，使用核心实验室对非增强 CT 低密度的平面测量来估计缺血核心体积，而不是预先指定的方法;根据核心实验室的判断，6 例患者满足了预先指定的灌注欠佳原因，42 例患者根据 CT 灌注低估了核心容积。此处报告的每组的平均体积来自预先指定的测量方法，并且比使用替代的平面法时要小。血栓切除术组的中位估计缺血核心体积为 74 ml（四分位距，50 至 111.5），医疗组为 77 ml（50.3 至 104）。共有 145 名患者 （41.2%） 为女性，50 名患者 （14.2%） 为黑人。339例患者中共有97例（28.6%）醒来时出现中风症状，352例患者中有211例（59.9%）从医院外或急诊科转移到参与试验中心。静脉溶栓治疗组37例（20.8%），内科治疗组30例（17.3%）;血栓切除术组 37 例患者中有 32 例使用阿替普酶，医疗护理组 30 例患者中有 28 例使用阿替普酶。 表S13提供了试验人群在进行试验的国家中对急性缺血性卒中患者的代表性。

At 90 days, the median score on the modified Rankin scale was 4 (interquartile range, 3 to 6) in the thrombectomy group and 5 (interquartile range, 4 to 6) in the medical-care group (Figure 1). The Wilcoxon–Mann–Whitney probability of superiority was 0.60 (95% confidence interval [CI], 0.55 to 0.65), and the generalized odds ratio favoring endovascular thrombectomy was 1.51 (95% CI, 1.20 to 1.89; P<0.001). In the prespecified tipping-point analysis, in which missing scores on the modified Rankin scale were imputed as 6 in the thrombectomy group and as 0 in the medical-care group, the Wilcoxon–Mann–Whitney probability of superiority was 0.59 (95% CI, 0.53 to 0.64), and the generalized odds ratio was 1.42 (95% CI, 1.13 to 1.78) (Table S11).
在 90 天时，血栓切除术组改良 Rankin 量表的中位评分为 4（四分位距，3 至 6），医疗护理组为 5（四分位距，4 至 6）（图 1）。Wilcoxon-Mann-Whitney 优效概率为 0.60（95% 置信区间 [CI]，0.55-0.65），有利于血管内血栓切除术的广义比值比为 1.51（95% CI，1.20-1.89;P<0.001）。在预先设定的临界点分析中，改良的Rankin量表上的缺失分数在血栓切除术组中归因为6，在医疗组中归因为0，Wilcoxon-Mann-Whitney优效概率为0.59（95%CI，0.53至0.64），广义比值比为1.42（95%CI，1.13至1.78）（表S11）。

Functional independence at 90 days (a score on the modified Rankin scale of 0 to 2) was observed in 20.3% of the patients in the thrombectomy group and in 7.0% of the patients in the medical-care group (relative risk, 2.97; 95% CI, 1.60 to 5.51). Independent ambulation (a score on the modified Rankin scale of 0 to 3) at 90 days occurred in 37.9% of the patients in the thrombectomy group and in 18.7% of the patients in the medical-care group (relative risk, 2.06; 95% CI, 1.43 to 2.96). Results for the other secondary outcomes are shown in Table 2 and were generally supportive of those of the primary analysis. There was no prespecified plan for adjustment of the widths of confidence intervals for multiplicity in comparing secondary outcomes between trial groups, and no definite conclusions can be drawn from these data. Successful reperfusion of the target vessel occurred in 142 patients (79.8%) in the thrombectomy group.
血栓切除术组 20.3% 的患者和 7.0% 的医疗护理组患者在 90 天时观察到功能独立（改良 Rankin 量表评分为 0-2）（相对危险度，2.97;95% CI，1.60-5.51）。血栓切除术组和医疗护理组分别有37.9%和18.7%的患者在90天时独立行走（改良的Rankin量表评分为0-3）（相对危险度为2.06;95%CI，1.43-2.96）。其他次要结局的结果如表2所示，总体上支持主要分析的结果。在比较试验组之间的次要结局时，没有预先指定的计划来调整多重性置信区间的宽度，并且无法从这些数据中得出明确的结论。血栓切除术组 142 例 （79.8%） 成功再灌注目标血管。

Symptomatic intracranial hemorrhage occurred in 1 patient (0.6%) in the thrombectomy group and in 2 patients (1.1%) in the medical-care group (relative risk, 0.49; 95% CI, 0.04 to 5.36) (Table 3). Parenchymal hematoma was observed in 5 patients (2.8%) in the thrombectomy group and in 3 patients (1.7%) in the medical-care group (relative risk, 1.63; 95% CI, 0.39 to 6.73). At 90 days, 68 of 177 patients (38.4%) in the thrombectomy group and 71 of 171 patients (41.5%) in the medical-care group had died (relative risk, 0.91; 95% CI, 0.71 to 1.18).
血栓切除组1例（0.6%）和医疗护理组2例（1.1%）发生症状性颅内出血（相对危险度，0.49;95%CI，0.04-5.36）（表3）。血栓切除组5例（2.8%）和内科治疗组3例（1.7%）均存在实质血肿（相对危险度，1.63;95%CI，0.39-6.73）。90天时，血栓切除术组177例患者中有68例（38.4%）死亡，医疗护理组171例患者中有71例（41.5%）死亡（相对危险度为0.91;95%CI，0.71-1.18）。

Early neurologic worsening occurred in 44 patients (24.7%) in the thrombectomy group and in 27 patients (15.5%) in the medical-care group (relative risk, 1.59; 95% CI, 1.03 to 2.45). In a post hoc analysis, from which no conclusions can be drawn, early neurologic worsening was associated with worse functional outcomes at 90 days (Wilcoxon–Mann–Whitney probability of superiority, 0.37 [95% CI, 0.31 to 0.43]; generalized odds ratio, 0.58 [95% CI, 0.45 to 0.74]). Patients who had neurologic worsening had larger ischemic-core lesions at baseline (median volume, 107 ml [interquartile range, 67 to 158] among patients with neurologic worsening vs. 77 ml [interquartile range, 60 to 100] among patients without neurologic worsening).
血栓切除组44例（24.7%）和内科治疗组27例（15.5%）发生早期神经系统恶化（相对危险度，1.59;95%CI，1.03-2.45）。在一项无法得出结论的事后分析中，早期神经系统恶化与 90 天时较差的功能结局相关（Wilcoxon-Mann-Whitney 优效概率，0.37 [95% CI，0.31-0.43];广义比值比，0.58 [95% CI，0.45-0.74]）。神经系统恶化的患者在基线时有较大的缺血核心病变（神经系统恶化患者的中位体积为 107 ml [四分位距，67 至 158]，而无神经系统恶化的患者为 77 ml [四分位距，60 至 100]）。

Procedural complications occurred in 33 patients (18.5%) in the thrombectomy group. Complications at the arterial access site included occlusion (in 3 patients [1.7%]), hematoma (in 1 patient [0.6%]), and infection (in 1 patient [0.6%]). In addition, 10 patients (5.6%) had vascular dissections, 7 (3.9%) had arterial perforation, and 11 (6.2%) had intraprocedural vasospasm. Two patients had both arterial access-site and intracranial vascular complications. The results of a post hoc analysis of clinical and safety outcomes among patients who had or did not have procedural complications are reported in Table S4.
血栓切除组 33 例患者 （18.5%） 发生手术并发症。动脉通路部位的并发症包括闭塞（3 例患者 [1.7%]）、血肿（1 例患者 [0.6%]）和感染（1 例患者 [0.6%]）。此外，10例（5.6%）有血管夹层，7例（3.9%）有动脉穿孔，11例（6.2%）有术中血管痉挛。2例患者同时出现动脉通路部位和颅内血管并发症。表S4报告了对有或没有手术并发症的患者的临床和安全性结果的事后分析结果。

The results of subgroup analyses are shown in Figure 2 and were generally supportive of the primary analysis. The results of the per-protocol analysis (which included 336 patients) and the as-treated analysis (which included 352 patients) were consistent with those of the primary intention-to-treat analysis (Tables S6, S7, S9, and S10). A sensitivity analysis that used site-rated ASPECTS categories showed similar treatment effects to those based on core laboratory ratings (Table S12).
亚组分析的结果如图2所示，总体上支持主要分析。符合方案的分析（包括336名患者）和治疗后分析（包括352名患者）的结果与主要意向性分析的结果一致（表S6、S7、S9和S10）。使用现场评定的SPECTS类别的敏感性分析显示，治疗效果与基于核心实验室评定的治疗效果相似（表S12）。

The results among the patients who had been enrolled on the basis of having a low ASPECTS value were similar to those of the primary analysis, as were the results among patients who had been enrolled on the basis of having large ischemic-core volumes (Figure 2). The results in the subgroup of patients with both an ASPECTS value of 5 or less and an estimated ischemic-core volume of 70 ml or greater were also similar to the overall results (Wilcoxon–Mann–Whitney probability of superiority, 0.61 [95% CI, 0.54 to 0.68]; generalized odds ratio, 1.58 [95% CI, 1.19 to 2.09]). This direction of effect persisted in patients with an ischemic-core volume greater than 100 ml and greater than 150 ml, although functional independence (a score on the modified Rankin scale of 0 to 2) was less frequent in these patients than in those with smaller ischemic cores.
在SPECTS值较低的基础上入组的患者的结果与初步分析的结果相似，在缺血核心体积较大的基础上入组的患者的结果也是如此（图2）。ASPECTS 值均为 5 或更低且估计缺血核心体积为 70 ml 或更大的患者亚组的结果也与总体结果相似（Wilcoxon-Mann-Whitney 优效概率，0.61 [95% CI，0.54-0.68];广义比值比，1.58 [95% CI，1.19-2.09]）。这种作用方向在缺血核心体积大于 100 ml 和大于 150 ml 的患者中持续存在，尽管这些患者的功能独立性（改良的 Rankin 量表评分为 0 至 2）比缺血核心较小的患者少见。