基于人工智能的算法在低剂量胸部CT中测量胸主动脉直径的验证
研究纳入ImaLife研究中240例低剂量非增强胸部CT受检者,比较AI与人工在8个指南推荐位置测量胸主动脉直径的表现,AI失败11例(4.6%),可分析229例;AI与人工测量无差异(32.7±6.4 mm vs 32.7±6.0 mm,p=0.70),AI重复性系数4.0 mm,AI与人工差异(1.3±1.6 mm)与读者间变异(1.4±1.4 mm)相当,仅在近端主动脉弓处AI差异更高(2.0±1.8 mm vs 0.9±0.9 mm,p<0.001)。
要点
- AI在低剂量非增强胸部CT上自动测量胸主动脉直径与人工测量无显著差异。
- AI失败率4.6%(11/240),均因分割错误,分析时需留意分割质量。
- 近端主动脉弓处AI与人工差异高于读者间变异(2.0±1.8 mm vs 0.9±0.9 mm)。
核心结论
- AI软件能准确测量胸主动脉直径,其与人工读者的差异与读者间变异相似,范围覆盖正常至扩张的主动脉。
研究设计与样本
- 研究纳入ImaLife研究中2017年8月至2022年5月接受低剂量非增强胸部CT的240名参与者(63±16岁,50%男性),随机抽取<50岁、≥80岁及胸主动脉直径≥40 mm各80人。
诊断标准与关键数值
- AI在8个指南推荐位置测量主动脉直径,与人工测量比较;AI重复性系数为4.0 mm;AI平均差异1.3±1.6 mm,读者间变异1.4±1.4 mm;近端主动脉弓处AI差异2.0±1.8 mm vs 读者间变异0.9±0.9 mm(p<0.001)。
临床应用与阅片要点
- AI与人工测量无系统偏倚,可用于低剂量非增强胸部CT的胸主动脉直径自动评估;近端主动脉弓处AI差异较高,阅片时需注意该位置的测量结果。
局限性
- AI评估因分割错误失败11例(4.6%)。
- 仅在近端主动脉弓处AI差异高于读者间变异(2.0±1.8 mm vs 0.9±0.9 mm,p<0.001)。
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)。
英文原文(摘要)· 点击展开
OBJECTIVES: To evaluate the performance of artificial intelligence (AI) software for automatic thoracic aortic diameter assessment in a heterogeneous cohort with low-dose, non-contrast chest computed tomography (CT).
MATERIALS AND METHODS: Participants of the Imaging in Lifelines (ImaLife) study who underwent low-dose, non-contrast chest CT (August 2017-May 2022) were included using random samples of 80 participants <50y, ≥80y, and with thoracic aortic diameter ≥40 mm. AI-based aortic diameters at eight guideline compliant positions were compared with manual measurements. In 90 examinations (30 per group) diameters were reassessed for intra- and inter-reader variability, which was compared to discrepancy of the AI system using Bland-Altman analysis, paired samples t-testing and linear mixed models.
RESULTS: We analyzed 240 participants (63 ± 16 years; 50 % men). AI evaluation failed in 11 cases due to incorrect segmentation (4.6 %), leaving 229 cases for analysis. No difference was found in aortic diameter between manual and automatic measurements (32.7 ± 6.4 mm vs 32.7 ± 6.0 mm, p = 0.70). Bland-Altman analysis yielded no systematic bias and a repeatability coefficient of 4.0 mm for AI. Mean discrepancy of AI (1.3 ± 1.6 mm) was comparable to inter-reader variability (1.4 ± 1.4 mm); only at the proximal aortic arch showed AI higher discrepancy (2.0 ± 1.8 mm vs 0.9 ± 0.9 mm, p < 0.001). No difference between AI discrepancy and inter-reader variability was found for any subgroup (all: p > 0.05).
CONCLUSION: The AI software can accurately measure thoracic aortic diameters, with discrepancy to a human reader similar to inter-reader variability in a range from normal to dilated aortas.
引用
- Hamelink II, de Heide EEJ, Pelgrim GJGJ, et al. Validation of an AI-based algorithm for measurement of the thoracic aortic diameter in low-dose chest CT. Eur J Radiol. 2023. PMID: 37659209. doi:10.1016/j.ejrad.2023.111067
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。