Lu Michael T, Raghu Vineet K, Mayrhofer Thomas, Aerts Hugo J W L, Hoffmann Udo
Massachusetts General Hospital Cardiovascular Imaging Research Center, Brigham and Women's Hospital Program for Artificial Intelligence in Medicine, and Harvard Medical School, Boston, Massachusetts (M.T.L., V.K.R., U.H.).
Harvard Medical School, Boston, Massachusetts,and Stralsund University of Applied Sciences, Stralsund, Germany (T.M.).
Ann Intern Med. 2020 Nov 3;173(9):704-713. doi: 10.7326/M20-1868. Epub 2020 Sep 1.
Lung cancer screening with chest computed tomography (CT) reduces lung cancer death. Centers for Medicare & Medicaid Services (CMS) eligibility criteria for lung cancer screening with CT require detailed smoking information and miss many incident lung cancers. An automated deep-learning approach based on chest radiograph images may identify more smokers at high risk for lung cancer who could benefit from screening with CT.
To develop and validate a convolutional neural network (CXR-LC) that predicts long-term incident lung cancer using data commonly available in the electronic medical record (EMR) (chest radiograph, age, sex, and whether currently smoking).
Risk prediction study.
U.S. lung cancer screening trials.
The CXR-LC model was developed in the PLCO (Prostate, Lung, Colorectal, and Ovarian) Cancer Screening Trial ( = 41 856). The final CXR-LC model was validated in additional PLCO smokers ( = 5615, 12-year follow-up) and NLST (National Lung Screening Trial) heavy smokers ( = 5493, 6-year follow-up). Results are reported for validation data sets only.
Up to 12-year lung cancer incidence predicted by CXR-LC.
The CXR-LC model had better discrimination (area under the receiver-operating characteristic curve [AUC]) for incident lung cancer than CMS eligibility (PLCO AUC, 0.755 vs. 0.634; < 0.001). The CXR-LC model's performance was similar to that of PLCO, a state-of-the-art risk score with 11 inputs, in both the PLCO data set (CXR-LC AUC of 0.755 vs. PLCO AUC of 0.751) and the NLST data set (0.659 vs. 0.650). When compared in equal-sized screening populations, CXR-LC was more sensitive than CMS eligibility in the PLCO data set (74.9% vs. 63.8%; = 0.012) and missed 30.7% fewer incident lung cancers. On decision curve analysis, CXR-LC had higher net benefit than CMS eligibility and similar benefit to PLCO.
Validation in lung cancer screening trials and not a clinical setting.
The CXR-LC model identified smokers at high risk for incident lung cancer, beyond CMS eligibility and using information commonly available in the EMR.
None.
胸部计算机断层扫描（CT）肺癌筛查可降低肺癌死亡率。医疗保险和医疗补助服务中心（CMS）的CT肺癌筛查资格标准要求详细的吸烟信息，且会遗漏许多新发肺癌病例。基于胸部X光片图像的自动化深度学习方法可能会识别出更多可从CT筛查中受益的肺癌高危吸烟者。
开发并验证一种卷积神经网络（CXR-LC），该网络使用电子病历（EMR）中常用的数据（胸部X光片、年龄、性别以及当前是否吸烟）来预测长期新发肺癌。
风险预测研究。
美国肺癌筛查试验。
CXR-LC模型在前列腺、肺癌、结直肠癌和卵巢癌筛查试验（PLCO）（n = 41856）中开发。最终的CXR-LC模型在另外的PLCO吸烟者（n = 5615，12年随访）和国家肺癌筛查试验（NLST）重度吸烟者（n = 5493，6年随访）中进行验证。结果仅报告验证数据集的情况。
CXR-LC预测的长达12年的肺癌发病率。
CXR-LC模型对新发肺癌的区分能力（受试者操作特征曲线下面积[AUC]）优于CMS资格标准（PLCO的AUC为0.755对0.634；P < 0.001）。在PLCO数据集（CXR-LC的AUC为0.755对PLCO的AUC为0.751）和NLST数据集（0.659对0.650）中，CXR-LC模型的性能与具有11个输入的先进风险评分PLCO相似。在同等规模的筛查人群中进行比较时，CXR-LC在PLCO数据集中比CMS资格标准更敏感（74.9%对63.8%；P = 0.012），遗漏的新发肺癌病例少30.7%。在决策曲线分析中，CXR-LC的净效益高于CMS资格标准，与PLCO的效益相似。
在肺癌筛查试验而非临床环境中进行验证。
CXR-LC模型识别出了超出CMS资格标准且使用EMR中常用信息的肺癌高危吸烟者。
无。
