基于人工智能的腹主动脉瘤腔内修复术前术后CT综合分析
本研究开发并验证了一套人工智能流程,用于分析腹主动脉瘤(AAA)患者EVAR术前和术后的增强CT血管造影。
要点
- AI流程可自动分割术前术后CTA结构,Dice评分均>0.9。
- 内漏检测AUC内部0.87、外部0.85,具有良好泛化性。
- AI与手动测量的最大AAA直径误差≤2 mm,相关性R≥0.97。
核心结论
- AI流程能够对AAA患者进行准确、可泛化且可解释的术前评估和EVAR术后随访。
研究设计与样本
- 回顾性研究,在两个独立中心连续纳入接受择期EVAR且有术前术后CTA的患者。
- 开发了四个卷积神经网络(nnU-Net)自动识别术前和术后CTA结构并检测内漏,自动提取最大AAA直径及其进展并与专家手动标注比较。
主要发现
- 所有AI分割效果优秀(所有Dice评分>0.9),解剖标志识别准确,误差与观察者间变异相当。
- 内漏检测在内部和外部队列的AUC分别为0.87和0.85。
- AI与手动测量的最大AAA直径差异微小(中位误差≤2 mm),跨中心相关性优秀(所有R≥0.97,p<.001;组内相关系数≥0.97)。
- 随访直径变化中AI与手动测量的不一致性低于观察者间变异。
关键数值与推荐
- 中位年龄74岁(IQR 69,79),女性15例(4.0%),中位随访8.1个月(IQR 3.1,27.9),术前最大AAA直径59 mm(IQR 55,68),150例患者随访期间发生内漏。
临床怎么用
- AI流程可用于EVAR术前评估和术后随访的自动化影像分析,辅助监测内漏和动脉瘤增长。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVE: Endovascular aortic repair (EVAR) requires lifelong imaging surveillance to monitor potential complications such as endoleaks (EL) or aneurysm growth. This study aimed to develop and validate a comprehensive artificial intelligence (AI) pipeline for analysing pre- and post-EVAR contrast enhanced computed tomography angiograms (CTAs) in abdominal aortic aneurysm (AAA) patients. METHODS: Consecutive patients who underwent elective EVAR with available pre- and post-EVAR CTAs were retrospectively identified at two independent centres. Four convolutional neural networks (nnU-Net) were developed to automatically identify pre-EVAR (lumen, thrombus, six anatomic landmarks) and post-EVAR CTAs (lumen, excluded aorta, six anatomic and three prosthetic landmarks) and to detect EL. Maximum AAA diameter and its progression during follow up were automatically extracted and compared with manual annotations across experts and protocols. Data from the internal centre were used to train and validate the models, while data from the external centre served for external validation. RESULTS: Three hundred and seventy-two patients (1 133 CT scans) were included: 135 patients (295 CTAs) for model development and 237 (838 CTAs) for external validation. Overall, median age was 74 (interquartile range (IQR) 69, 79) years; 15 were women (4.0%), median follow up was 8.1 (IQR 3.1, 27.9) months, pre-EVAR maximum AAA diameter was 59 (IQR 55, 68) mm, and 150 patients developed an EL during follow up. All AI segmentations were excellent (all Dice scores > 0.9) and landmarks were accurately identified, with errors comparable with the interobserver variability. EL detection achieved an area under the curve of 0.87 and 0.85 in the internal and external cohorts, respectively. Differences between AI derived and manual maximum AAA diameters were minor (median error ≤ 2 mm), with excellent correlations across centres (all R ≥ 0.97, p < .001; intraclass correlation coefficient ≥ 0.97). Disagreement in follow up diameter change between AI and manual measurements was lower than the interobserver variability. CONCLUSION: The AI based pipeline enabled accurate, generalisable, and explainable assessment of AAA patients, including both pre-EVAR evaluation and post-EVAR follow up.
引用
- Garrido-Oliver J, Tenezaca-Sari X, Ribeiro T, et al. Comprehensive Artificial Intelligence-Based Analysis of Pre- and Post-operative Computed Tomography Scans in Patients Undergoing Endovascular Abdominal Aortic Repair. Eur J Vasc Endovasc Surg. 2026. PMID: 42759832. doi:10.1016/j.ejvs.2026.09.035
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。