基于欧洲筛查数据对肺结节恶性风险分层深度学习算法的外部测试
本研究利用丹麦、意大利多中心及荷兰-比利时三项欧洲肺癌筛查试验的基线CT,外部测试了在NLST数据上训练的肺结节恶性风险深度学习算法,并与PanCan模型比较。汇总队列4146例参与者中,DL算法在1年、2年及全筛查期诊断癌症的AUC分别为0.98、0.96、0.94;在5-15mm不确定结节亚组中DL显著优于PanCan(AUC 0.95、0.94、0.90 vs 0.91、0.88、0.86,均P<.05),且在1年诊断癌症100%敏感度下将68.1%良性病例判为低风险(PanCan为47.4%)。
要点
- 在5-15mm不确定结节中,DL算法对1年、2年及全筛查期诊断癌症的AUC分别为0.95、0.94、0.90,均显著优于PanCan模型。
- 在1年诊断癌症100%敏感度下,DL将68.1%良性病例判为低风险,PanCan为47.4%,假阳性相对减少39.4%。
- 在癌症与良性结节1:2大小匹配亚组中,DL的AUC为0.79,PanCan为0.60(P<.01)。
核心结论
- DL算法在多个欧洲筛查数据集上优于PanCan模型,对不确定结节展现出更优的恶性预测能力并大幅减少假阳性分类。
研究设计与样本
- 回顾性研究,使用丹麦肺癌筛查试验、意大利多中心肺检测试验及荷兰-比利时肺癌筛查试验的基线CT外部测试在NLST数据上训练的DL算法。
诊断标准与关键数值
- 汇总队列(4146例参与者,7614个良性结节和180个恶性结节)中,DL算法对1年、2年及全筛查期诊断癌症的AUC分别为0.98、0.96、0.94,PanCan模型分别为0.98、0.94、0.93(P分别为.19、.02、.46)。
- 亚组A(5-15mm不确定结节,129个恶性、2086个良性)中,DL对1年、2年及全筛查期诊断癌症的AUC分别为0.95、0.94、0.90,PanCan分别为0.91、0.88、0.86(均P<.05)。
- 对1年诊断癌症达100%敏感度时,DL将68.1%良性病例判为低风险,PanCan为47.4%,假阳性相对减少39.4%。
- 亚组B(180个恶性与360个良性结节,1:2大小匹配)中,DL的AUC为0.79,PanCan为0.60(P<.01)。
局限性
- 原文未报告局限性。
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
Background Low-dose CT screening reduces lung cancer-related deaths but has high rates of false-positive findings. A deep learning (DL) algorithm could improve nodule risk stratification but requires robust external testing. Purpose To externally test a DL algorithm for nodule malignancy risk estimation using pooled data from three large European lung cancer screening trials. Materials and Methods In this retrospective study, a DL algorithm trained on National Lung Screening Trial data was externally tested using baseline CT scans from the Danish Lung Cancer Screening Trial, the Multicentric Italian Lung Detection trial, and the Dutch-Belgian Lung Cancer Screening Trial. Performance was assessed across the pooled cohort and two subsets: subset A, including indeterminate nodules (5-15 mm); and subset B, including cancers size-matched to benign nodules (1:2 ratio). Performance, including the area under the receiver operating characteristic curve (AUC), was compared with the Pan-Canadian Early Detection of Lung Cancer (PanCan) model. Results The pooled cohort included 4146 participants (median age, 58 years; 78% male participants; median smoking history, 38 pack-years) with 7614 benign and 180 malignant nodules. The DL algorithm achieved AUCs of 0.98, 0.96, and 0.94 for cancers diagnosed within 1 year, 2 years, and throughout screening, respectively, compared with 0.98, 0.94, and 0.93 (P = .19, .02, and .46, respectively) for the PanCan model. In subset A (129 malignant and 2086 benign nodules), DL significantly outperformed PanCan across the same cancer diagnosis timeframes (respective AUCs: 0.95, 0.94, and 0.90 vs 0.91, 0.88, and 0.86; all P < .05). At 100% sensitivity for cancers diagnosed within 1 year, DL classified 68.1% of benign cases as low risk versus 47.4% for the PanCan model, a 39.4% relative reduction in false-positive findings. In subset B (180 malignant and 360 benign nodules), the AUC of the DL algorithm versus the PanCan model was 0.79 versus 0.60 (P < .01), respectively. Conclusion The DL algorithm outperformed the PanCan model across multiple European screening datasets, demonstrating superior malignancy prediction while substantially reducing false-positive classifications for indeterminate nodules. © RSNA, 2025 Supplemental material is available for this article.
引用
- Antonissen N, Venkadesh KV, Dinnessen R, et al. External Test of a Deep Learning Algorithm for Pulmonary Nodule Malignancy Risk Stratification Using European Screening Data. Radiology. 2025. PMID: 40956165. doi:10.1148/radiol.250874
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。