| 王敏可,王甜甜,黄金标,林优优,谢继承.全脑CT灌注联合深度学习全模型迭代算法用于评估急性缺血性脑卒中[J].中国医学影像技术,2025,41(4):515~519 |
| 全脑CT灌注联合深度学习全模型迭代算法用于评估急性缺血性脑卒中 |
| Whole brain CT perfusion combined with artificial intelligence iterative reconstruction for evaluating acute ischemic stroke |
| 投稿时间：2024-10-29 修订日期：2025-02-25 |
| DOI：10.13929/j.issn.1003-3289.2025.04.002 |
| 中文关键词: 脑卒中 体层摄影术,X线计算机 灌注成像 血管造影术 |
| 英文关键词:stroke tomography, X-ray computed perfusion imaging angiography |
| 基金项目: |
|
| 摘要点击次数: 595 |
| 全文下载次数: 162 |
| 中文摘要: |
| 目的 观察全脑CT灌注(CTP)动脉期图像联合深度学习全模型迭代算法(AIIR)用于评估急性缺血性脑卒中(AIS)的价值。方法 前瞻性对50例AIS行全脑CTP与CT血管造影(CTA),之后分别行AIIR(CTP-AIIR组)及混合迭代重建(HIR,CTP-HIR组)并提取CTP动脉期CTA图像;对常规CTA行HIR(CTA-HIR组)。以5分法对各组图像质量进行主观评分;测量基底动脉干、颈内动脉海绵状段、大脑中动脉M1段及脑实质噪声,计算3支动脉信噪比(SNR)和对比度噪声比(CNR)。以数字减影血管造影为参考标准,评估CTP及CTA定位责任血管的准确性。结果 CTP-AIIR组噪声、血管边缘锐利度、小血管可见性和整体可诊断性4项主观评分均显著高于,3支动脉及脑实质噪声均低于,而SNR、CNR均高于CTP-HIR和CTA-HIR组(P均＜0.017)。基于CTP-AIIR组与CTA-HIR组判断责任血管的准确率均为96.00%(48/50),与基于CTP-HIR组(44/50,88.00%)差异无统计学意义(P=0.142)。结论 全脑CTP动脉期图像结合AIIR评估AIS可获得与常规CTA相当的图像质量及诊断效能。 |
| 英文摘要: |
| Objective To observe the value of whole brain CT perfusion (CTP)-derived arterial phase images combined with artificial intelligence iterative reconstruction (AIIR) for evaluating acute ischemic stroke (AIS). Methods Fifty AIS patients were prospectively enrolled, and whole brain CTP followed by routine CT angiography (CTA) were performed. CTP images were reconstructed using AIIR (CTP-AIIR group) and hybrid iterative reconstruction (HIR, CTP-HIR group), respectively, and CTA images were derived from CTP at the arterial phase. Meanwhile, routine CTA images were obtained using HIR (CTA-HIR group). Then image quality was subjectively evaluated with a 5-point scale. The noise of basilar artery trunk, cavernous segment of internal carotid artery, M1 segment of middle cerebral artery and the brain parenchyma were calculated, and signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of the above 3 arteries were measured. Taken digital subtraction angiography as the reference standard, the accuracy of CTP and CTA for localizing the responsible vessel of AIS was assessed. Results In CTP-AIIR group, the subjective scoring of 4 subjective metrics, including image noise, sharpness of vessel margin, small vessel visibility and overall diagnosability were significantly higher than, the noise of 3 arteries and brain parenchyma were lower than, and SNR and CNR of 3 arteries were higher than those in CTP-HIR group and CTA-HIR group (all P＜0.017). The diagnostic accuracy of responsible vessel based on CTP-AIIR group and CTA-HIR group were both 96.00% (48/50), not significantly different with that in CTP-HIR group (44/50, 88.00%) (P=0.142). Conclusion Whole brain CTP-derived arterial phase CTA combined with AIIR could achieve comparable image quality and diagnostic performance of routine CTA for evaluating AIS. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
