实用医学杂志 ›› 2025, Vol. 41 ›› Issue (2): 264-270.doi: 10.3969/j.issn.1006-5725.2025.02.017

• 医学检查与临床诊断 • 上一篇    

CT灌注成像联合人工智能评估脑梗死后再灌注损伤的临床意义

芦苇1,张盼1,覃玉术2   

  1. 1.中国贵航集团三〇二医院,医学影像科,(贵州 安顺 561000 )
    2.中国贵航集团三〇二医院,神经内科,(贵州 安顺 561000 )
  • 收稿日期:2024-07-25 出版日期:2025-01-25 发布日期:2025-01-26
  • 基金资助:
    贵州省卫生健康委科学技术基金项目(gzwkj2022-111)

Clinical significance of CT perfusion imaging combined with artificial intelligence in evaluating reperfusion injury after cerebral infarction

Wei LU1,Pan ZHANG1,Yushu. QIN2   

  1. 1.*Department of Medical Imaging,China Guihang Group 3, Hospital, Anshun 561000,Guizhou,China
  • Received:2024-07-25 Online:2025-01-25 Published:2025-01-26

摘要:

目的 分析CT灌注成像联合人工智能(AI)评估脑梗死后再灌注损伤的意义。 方法 选取2019年1月至2023年10月医院收治的106例脑梗死患者,依据患者溶栓14 d后是否发生再灌注损伤分为再灌注损伤组、无灌注损伤组。比较再灌注损伤组、无灌注损伤组CT灌注成像、AI参数,分析影响脑梗死患者溶栓后再灌注损伤的因素,分析CT灌注成像参数联合AI预测脑梗死患者溶栓治疗后再灌注损伤的价值。 结果 31例出现再灌注损伤,其余75例无灌注损伤。再灌注损伤组的脑血流量(CBF)、平均CT值、熵水平均低于无灌注损伤组(P < 0.05),脑血容量(CBV)、注对比剂平均通过时间(MTT)、达峰时间(TTP)、峰度均高于无灌注损伤组(P < 0.05)。logistic回归分析显示NIHSS评分(OR = 5.228,95%CI:2.151 ~ 12.705)、CBF(OR = 3.777,95%CI:1.554 ~ 9.180)、CBV(OR = 3.699,95%CI:1.522 ~ 9.989)、平均CT值(OR = 4.125,95%CI:1.697 ~ 10.024)是影响脑梗死患者溶栓治疗后再灌注损伤的影响因素(P < 0.05)。ROC曲线显示,CBF、CBV、平均CT值及三者联合预测脑梗死患者溶栓治疗后再灌注损伤的灵敏度分别为67.74%、70.97%、77.42%、87.10%,特异度分别为70.67%、74.67%、77.33%、90.67%,AUC分别为0.665、0.667、0.744、0.908。 结论 CT灌注成像联合AI评估脑梗死后再灌注损伤效能良好。

关键词: CT灌注成像, 人工智能, 脑梗死, 再灌注损伤

Abstract:

Objective To analyze the significance of CT perfusion imaging combined with artificial intelligence (AI) in evaluating reperfusion injury after cerebral infarction. Methods 106 patients with cerebral infarction admitted to the hospital from January 2019 to October 2023 were prospectively selected as the study objects. Patients were divided into reperfusion injury group and no perfusion injury group according to whether reperfusion injury occurred after 14 days of thrombolytic therapy. CT perfusion imaging and AI parameters were compared between reperfusion injury group and non-perfusion injury group. The factors affecting reperfusion injury in cerebral infarction patients after thrombolytic therapy were analyzed. The value of CT perfusion imaging parameters combined with AI in predicting reperfusion injury after thrombolytic therapy in cerebral infarction patients was analyzed. Results 31 cases had reperfusion injury, and the other 75 cases had no perfusion injury. CBF, average CT value and entropy level in the reperfusion injury group were lower than those in the non-perfusion injury group (P < 0.05), CBV, MTT, TTP and kurtosis were higher than those in the non-perfusion injury group (P < 0.05). Logistic regression analysis showed that NIHSS (OR = 5.228, 95%CI: 2.151 ~ 12.705), CBF(OR = 3.777, 95%CI: 1.554 ~ 9.180), CBV(OR = 3.699, 95%CI: 1.522 ~ 9.989) and average CT value (OR = 4.125, 95%CI: 1.697 ~ 10.024) were the influencing factors of reperfusion injury in cerebral infarction patients after thrombolytic therapy (P < 0.05). ROC curve results showed that the sensitivity of CBF, CBV, average CT value and their combination in predicting reperfusion injury after thrombolytic therapy in cerebral infarction patients were 67.74%, 70.97%, 77.42%, 87.10%, and the specificity were 70.67%, 74.67%, 77.33%, 90.67%, AUC values were 0.665, 0.667, 0.744 and 0.908. Conclusion CT perfusion imaging combined with AI is effective in evaluating reperfusion injury after cerebral infarction.

Key words: CT perfusion imaging, artificial intelligence, cerebral infarction, reperfusion injury

中图分类号: