From the Departments of Medicine, Division of Artificial Intelligence in Medicine, Imaging, and Biomedical Sciences, Cedars-Sinai Medical Center, 6500 Wilshire Blvd, Los Angeles, CA 90048 (A.M.M., M.B., A.S., B.P.B., A.K., R.J.H.M., V.B., M.L., D.S.B., D.D., P.J.S.); Signal and Image Processing Institute, Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, Calif (A.S.); and Department of Cardiac Sciences, University of Calgary, Calgary, Alberta, Canada (R.J.H.M.).
Radiology. 2024 Sep;312(3):e240541. doi: 10.1148/radiol.240541.
Background Incidental extrapulmonary findings are commonly detected on chest CT scans and can be clinically important. Purpose To integrate artificial intelligence (AI)-based segmentation for multiple structures, coronary artery calcium (CAC), and epicardial adipose tissue with automated feature extraction methods and machine learning to detect extrapulmonary abnormalities and predict all-cause mortality (ACM) in a large multicenter cohort. Materials and Methods In this post hoc analysis, baseline chest CT scans in patients enrolled in the National Lung Screening Trial (NLST) from August 2002 to September 2007 were included from 33 participating sites. Per scan, 32 structures were segmented with a multistructure model. For each structure, 15 clinically interpretable radiomic features were quantified. Four general codes describing abnormalities reported by NLST radiologists were applied to identify extrapulmonary significant incidental findings on the CT scans. Death at 2-year and 10-year follow-up and the presence of extrapulmonary significant incidental findings were predicted with ensemble AI models, and individualized structure risk scores were evaluated. Area under the receiver operating characteristic curve (AUC) analysis was used to evaluate the performance of the models for prediction of ACM and extrapulmonary significant incidental findings. The Pearson χ test and Kruskal-Wallis rank sum test were used for statistical analyses. Results A total of 24 401 participants (median age, 61 years [IQR, 57-65 years]; 14 468 male) were included. In 3880 of 24 401 participants (16%), 4283 extrapulmonary significant incidental findings were reported. During the 10-year follow-up, 3389 of 24 401 participants (14%) died. CAC had the highest feature importance for predicting the three study end points. The 10-year ACM model demonstrated the best AUC performance (0.72; per-year mortality of 2.6% above and 0.8% below the risk threshold), followed by 2-year ACM (0.71; per-year mortality of 1.13% above and 0.3% below the risk threshold) and prediction of extrapulmonary significant incidental findings (0.70; probability of occurrence of 25.4% above and 9.6% below the threshold). Conclusion A fully automated AI model indicated extrapulmonary structures at risk on chest CT scans and predicted ACM with explanations. ClinicalTrials.gov Identifier: NCT00047385 © RSNA, 2024 See also the editorial by Yanagawa and Hata in this issue.
胸部 CT 扫描常可发现肺外意外发现，且可能具有临床重要性。
将基于人工智能 (AI) 的多结构分割、冠状动脉钙化 (CAC) 和心外膜脂肪组织与自动特征提取方法和机器学习相结合，以在大型多中心队列中检测肺外异常并预测全因死亡率 (ACM)。
本研究为回顾性分析，纳入 2002 年 8 月至 2007 年 9 月参与国家肺癌筛查试验 (NLST) 的 33 个参与中心的基线胸部 CT 扫描。对每个扫描，使用多结构模型对 32 个结构进行分割。对每个结构，定量分析 15 个临床可解释的放射组学特征。对 CT 扫描上的肺外显著意外发现，应用 NLST 放射科医生报告的四个通用编码来识别。采用集成 AI 模型预测 2 年和 10 年随访时的死亡和肺外显著意外发现，并评估个体结构风险评分。使用受试者工作特征曲线下面积 (AUC) 分析评估模型对 ACM 和肺外显著意外发现的预测性能。采用 Pearson χ 检验和 Kruskal-Wallis 秩和检验进行统计学分析。
共纳入 24401 例参与者（中位年龄，61 岁 [四分位距，57-65 岁]；14468 例男性）。24401 例参与者中，3880 例（16%）报告 4283 例肺外显著意外发现。在 10 年随访期间，24401 例参与者中有 3389 例（14%）死亡。CAC 对预测这 3 个研究终点的重要性最高。10 年 ACM 模型的 AUC 表现最佳（风险阈值以上和以下每年死亡率分别为 2.6%和 0.8%），其次是 2 年 ACM（0.71；风险阈值以上和以下每年死亡率分别为 1.13%和 0.3%）和肺外显著意外发现预测（0.70；概率分别为 25.4%和 9.6%）。
全自动 AI 模型提示胸部 CT 扫描中的肺外结构存在风险，并提供 ACM 预测的解释。
NCT00047385
研究在本期杂志中还包括 Yanagawa 和 Hata 的社论。
