Assessing eligibility for lung cancer screening: Parsimonious multi-country ensemble machine learning models for lung cancer prediction
Assessing eligibility for lung cancer screening: Parsimonious multi-country ensemble machine learning models for lung cancer prediction
复制标题
DOI：
10.1101/2023.01.27.23284974
复制
发表时间：
2023-01
期刊：
影响因子：
--
通讯作者：
Thomas Callender MBChB MSc;Fergus Imrie DPhil;Bogdan Cebere MSc;N. Md;Neal Navani Mbbs;M. Schaar;S. M. J. Mbbs
中科院分区：
文献类型：
--
作者：
Thomas Callender MBChB MSc;Fergus Imrie DPhil;Bogdan Cebere MSc;N. Md;Neal Navani Mbbs;M. Schaar;S. M. J. Mbbs
Ensemble机器学习可以支持高度简约的预测模型的开发，这些模型可以保持更复杂模型的性能，同时最大限度地提高简单性和通用性，支持个性化筛选的广泛采用。在这项工作中，我们的目标是开发和验证集成机器学习模型，以确定基于风险的肺癌筛查的资格。方法为了建立模型，我们使用了来自英国生物库前瞻性队列中216，714名曾经吸烟者和美国国家肺筛查随机对照试验对照组中26，616名高危曾经吸烟者的数据。我们在胸部X线摄影组的49，593名参与者和美国前列腺、肺、结直肠和卵巢筛查试验（PLCO）的所有80，659名吸烟参与者中外部验证了我们的模型。开发模型来预测从基线开始的五年内两种结局的风险：肺癌诊断和肺癌死亡。我们评估了模型区分（受试者工作曲线下面积，AUC），校准（校准曲线和预期/观察比值），整体性能（Brier评分）和决策曲线分析的净效益。结果使用三个变量--年龄、吸烟持续时间和吸烟包年--预测肺癌死亡率（UCL-D）和发病率（UCL-I）的模型，尽管只需要四分之一的预测因子，但在鉴别力、总体性能和净效益方面与目前使用的对照品达到或超过了同等水平。在PLCO试验的外部验证中，UCL-D的AUC为0.803（95% CI：0.783-0.824），并经过良好校准，预期/观察（E/O）比值为1.05（95% CI：0.95-1.19）。UCL-I的AUC为0.787（95% CI：0.771-0.802），E/O比为1.0（0.92-1.07）。在5年风险阈值为0.68%和1.17%时，UCL-D和UCL-I的敏感性分别为85.5%和83.9%，分别比USPSTF-2021标准高7.9%和6.2%。我们提出了简约的集成机器学习模型来预测曾经吸烟者的肺癌风险，展示了一种新的方法，可以简化在多种环境中实施基于风险的肺癌筛查。
Background Ensemble machine learning could support the development of highly parsimonious prediction models that maintain the performance of more complex models whilst maximising simplicity and generalisability, supporting the widespread adoption of personalised screening. In this work, we aimed to develop and validate ensemble machine learning models to determine eligibility for risk-based lung cancer screening. Methods For model development, we used data from 216,714 ever-smokers in the UK Biobank prospective cohort and 26,616 high-risk ever-smokers in the control arm of the US National Lung Screening randomised controlled trial. We externally validated our models amongst the 49,593 participants in the chest radiography arm and amongst all 80,659 ever-smoking participants in the US Prostate, Lung, Colorectal and Ovarian Screening Trial (PLCO). Models were developed to predict the risk of two outcomes within five years from baseline: diagnosis of lung cancer, and death from lung cancer. We assessed model discrimination (area under the receiver operating curve, AUC), calibration (calibration curves and expected/observed ratio), overall performance (Brier scores), and net benefit with decision curve analysis. Results Models predicting lung cancer death (UCL-D) and incidence (UCL-I) using three variables - age, smoking duration, and pack-years - achieved or exceeded parity in discrimination, overall performance, and net benefit with comparators currently in use, despite requiring only one-quarter of the predictors. In external validation in the PLCO trial, UCL-D had an AUC of 0.803 (95% CI: 0.783-0.824) and was well calibrated with an expected/observed (E/O) ratio of 1.05 (95% CI: 0.95-1.19). UCL-I had an AUC of 0.787 (95% CI: 0.771-0.802), an E/O ratio of 1.0 (0.92-1.07). The sensitivity of UCL-D was 85.5% and UCL-I was 83.9%, at 5-year risk thresholds of 0.68% and 1.17%, respectively 7.9% and 6.2% higher than the USPSTF-2021 criteria at the same specificity. Conclusions We present parsimonious ensemble machine learning models to predict the risk of lung cancer in ever-smokers, demonstrating a novel approach that could simplify the implementation of risk-based lung cancer screening in multiple settings.