Division of Neurology, University of Alberta, Edmonton, Canada.
Neuroscience and Mental Health Institute, University of Alberta, Edmonton, Canada.
Ann Clin Transl Neurol. 2024 Nov;11(11):2967-2976. doi: 10.1002/acn3.52207. Epub 2024 Oct 7.
Early diagnosis of large vessel occlusion (LVO) in acute stroke often requires CT angiography (CTA). Automated CT perfusion (CTP) software, which identifies blood flow abnormalities, enhances LVO diagnosis and patient selection for endovascular thrombectomy (EVT). This study evaluates the sensitivity of automated CTP images in detecting perfusion abnormalities in patients with acute ischemic stroke (AIS) and LVO or medium vessel occlusion (MeVO), compared to CTA.
We screened acute ischemic stroke patients presenting within 24 h who underwent CT, CTA, and CTP as per institutional protocol. RAPID AI software processed CTP images, while neuroradiologists reviewed CTA for intracranial arterial occlusions. Sensitivity, specificity, and accuracy of automated CTP maps in detecting occlusions were assessed.
Of 790 screened patients, 31 were excluded due to lack of RAPID CTP data or poor-quality scans, leaving 759 for analysis. The median age was 71 years (IQR: 61-81), with 47% female. Among them, 678 had AIS, and 81 had AIS ruled out. CTA identified arterial occlusion in 562 patients (74%), with corresponding CTP abnormalities in 537 patients (Tmax > 6 sec). In the 197 without occlusion, CTP was negative in 161. Automated CTP maps had a sensitivity of 95.55% (CI 95: 93.50-97.10%), specificity of 81.73% (CI 95: 75.61-86.86%), negative predictive value of 98.22% (CI 95: 97.39-98.79%), positive predictive value of 63.54% (CI 95: 56.46-70.09%), and overall accuracy of 85.18% (CI 95: 82.45-87.64%).
Automated CTP maps demonstrated high sensitivity and negative predictive value for LVOs and MeVOs, suggesting their usefulness as a rapid diagnostic tool, especially in settings without expert neuroradiologists.
急性脑卒中患者常需行 CT 血管造影（CTA）以早期诊断大血管闭塞（LVO）。自动 CT 灌注（CTP）软件可以识别血流异常，从而提高 LVO 诊断和血管内血栓切除术（EVT）的患者选择。本研究评估了自动 CTP 图像在检测急性缺血性脑卒中（AIS）合并 LVO 或中等血管闭塞（MeVO）患者的灌注异常方面的敏感性，并与 CTA 进行了比较。
我们筛选了在 24 小时内按机构方案行 CT、CTA 和 CTP 的急性缺血性脑卒中患者。RAPID AI 软件处理 CTP 图像，神经放射科医师则对 CTA 进行颅内动脉闭塞评估。评估了自动 CTP 图在检测闭塞方面的敏感性、特异性和准确性。
在 790 例筛选出的患者中，有 31 例因缺乏 RAPID CTP 数据或扫描质量差而被排除，759 例患者纳入分析。中位年龄为 71 岁（IQR：61-81），女性占 47%。其中，678 例患者为 AIS，81 例患者为排除 AIS。CTA 在 562 例患者中发现动脉闭塞（74%），相应的 CTP 异常在 537 例患者中发现（Tmax>6 秒）。在 197 例无闭塞的患者中，CTP 阴性的有 161 例。自动 CTP 图的敏感性为 95.55%（95%CI：93.50-97.10%），特异性为 81.73%（95%CI：75.61-86.86%），阴性预测值为 98.22%（95%CI：97.39-98.79%），阳性预测值为 63.54%（95%CI：56.46-70.09%），总体准确率为 85.18%（95%CI：82.45-87.64%）。
自动 CTP 图对 LVO 和 MeVO 具有较高的敏感性和阴性预测值，表明其作为一种快速诊断工具的有用性，尤其是在没有神经放射科专家的情况下。
