胸部 CT 扫描前影像有助于深度学习模型预测筛查性肺结节恶性风险
Prior CT Improves Deep Learning for Malignancy Risk Estimation of Screening-detected Pulmonary Nodules.
机构信息
From the Department of Medical Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, the Netherlands (K.V.V., T.A.A., E.T.S., B.v.G., M.P., C.J.); Robotics and Control Laboratory, The University of British Columbia, Vancouver, Canada (T.A.A.); Department of Medicine, Section of Pulmonary Medicine, Herlev-Gentofte Hospital, Hellerup, Denmark (Z.S.); Department of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark (Z.S.); Section of Radiology, Department of Medicine and Surgery, University of Parma, Parma, Italy (M.S., N.S.); Department of Thoracic Surgery, IRCCS Istituto Nazionale Tumori, Milan, Italy (M.S., U.P.); and Department of Radiology, University Medical Center Groningen, Groningen, the Netherlands (M.P.).
出版信息
Radiology. 2023 Aug;308(2):e223308. doi: 10.1148/radiol.223308.
Background Prior chest CT provides valuable temporal information (eg, changes in nodule size or appearance) to accurately estimate malignancy risk. Purpose To develop a deep learning (DL) algorithm that uses a current and prior low-dose CT examination to estimate 3-year malignancy risk of pulmonary nodules. Materials and Methods In this retrospective study, the algorithm was trained using National Lung Screening Trial data (collected from 2002 to 2004), wherein patients were imaged at most 2 years apart, and evaluated with two external test sets from the Danish Lung Cancer Screening Trial (DLCST) and the Multicentric Italian Lung Detection Trial (MILD), collected in 2004-2010 and 2005-2014, respectively. Performance was evaluated using area under the receiver operating characteristic curve (AUC) on cancer-enriched subsets with size-matched benign nodules imaged 1 and 2 years apart from DLCST and MILD, respectively. The algorithm was compared with a validated DL algorithm that only processed a single CT examination and the Pan-Canadian Early Lung Cancer Detection Study (PanCan) model. Results The training set included 10 508 nodules (422 malignant) in 4902 trial participants (mean age, 64 years ± 5 [SD]; 2778 men). The size-matched external test sets included 129 nodules (43 malignant) and 126 nodules (42 malignant). The algorithm achieved AUCs of 0.91 (95% CI: 0.85, 0.97) and 0.94 (95% CI: 0.89, 0.98). It significantly outperformed the DL algorithm that only processed a single CT examination (AUC, 0.85 [95% CI: 0.78, 0.92; = .002]; and AUC, 0.89 [95% CI: 0.84, 0.95; = .01]) and the PanCan model (AUC, 0.64 [95% CI: 0.53, 0.74; < .001]; and AUC, 0.63 [95% CI: 0.52, 0.74; < .001]). Conclusion A DL algorithm using current and prior low-dose CT examinations was more effective at estimating 3-year malignancy risk of pulmonary nodules than established models that only use a single CT examination. Clinical trial registration nos. NCT00047385, NCT00496977, NCT02837809 © RSNA, 2023 See also the editorial by Horst and Nishino in this issue.
背景 先前的胸部 CT 提供了有价值的时间信息（例如，结节大小或外观的变化），可准确估计恶性肿瘤风险。目的 开发一种深度学习（DL）算法，该算法使用当前和先前的低剂量 CT 检查来估计肺结节的 3 年恶性肿瘤风险。材料与方法 在这项回顾性研究中，使用国家肺癌筛查试验（NLST）的数据（收集于 2002 年至 2004 年）对算法进行了训练，其中患者的两次成像时间间隔最长为 2 年，并通过丹麦肺癌筛查试验（DLCST）和多中心意大利肺癌检测试验（MILD）的两个外部测试集进行了评估，分别收集于 2004-2010 年和 2005-2014 年。使用癌症富集亚组的受试者工作特征曲线（ROC）下面积（AUC）评估性能，这些亚组的大小匹配良性结节分别来自 DLCST 和 MILD，两次成像时间间隔为 1 年和 2 年。该算法与经过验证的仅处理单次 CT 检查的 DL 算法和 Pan-Canadian Early Lung Cancer Detection Study（PanCan）模型进行了比较。结果 训练集包括来自 4902 名试验参与者的 10508 个结节（422 个恶性）（平均年龄，64 岁±5[SD]；2778 名男性）。大小匹配的外部测试集包括 129 个结节（43 个恶性）和 126 个结节（42 个恶性）。该算法的 AUC 为 0.91（95%CI：0.85，0.97）和 0.94（95%CI：0.89，0.98）。它显著优于仅处理单次 CT 检查的 DL 算法（AUC，0.85[95%CI：0.78，0.92； =.002]；和 AUC，0.89[95%CI：0.84，0.95； =.01]）和 PanCan 模型（AUC，0.64[95%CI：0.53，0.74； <.001]；和 AUC，0.63[95%CI：0.52，0.74； <.001]）。结论 一种使用当前和先前低剂量 CT 检查的 DL 算法在估计肺结节 3 年恶性肿瘤风险方面比仅使用单次 CT 检查的既定模型更有效。临床试验注册号 NCT00047385、NCT00496977、NCT02837809 ©RSNA，2023 另见本期 Horst 和 Nishino 的社论。
Zotero 插件