便携式电子鼻用于肺癌筛查的训练与验证
本研究用五台便携式电子鼻采集60例肺癌患者与107名健康对照的呼气样本,各呼出5分钟,分为训练组建人工神经网络模型、盲法对照组验证模型。区分肺癌与健康对照的诊断准确率83%、敏感度83%、特异度84%、AUC 0.84;盲法组敏感度88%、特异度86%、准确率86%。作者认为该便携电子鼻有区分能力,值得扩大队列验证。
要点
- 便携式电子鼻区分肺癌与健康对照的诊断准确率为83%,盲法验证组为86%。
- 训练组敏感度83%、特异度84%、AUC 0.84;盲法组敏感度88%、特异度86%。
- 属可行性研究,需更大队列及更多早期肿瘤患者验证其稳健性与诊断链价值。
核心结论
- 该可行性研究显示便携式电子鼻可区分肺癌患者与健康对照,结果可能对未来肺癌筛查有重要意义。
研究设计与样本
- 研究使用五台电子鼻设备采集呼气样本,共60例肺癌患者与107名健康对照各通过电子鼻呼气5分钟。
- 患者被分配到训练组用于构建人工神经网络模型,或分配到盲法对照组用于验证该模型。
主要发现
- 区分肺癌患者与健康对照时,诊断准确率为83%,敏感度83%,特异度84%,曲线下面积0.84。
- 盲法组结果可比,敏感度88%,特异度86%,诊断准确率86%。
临床怎么用
- 该便携电子鼻有望作为非侵入性工具用于肺癌筛查,但其在诊断链中的附加价值尚待确定。
局限性
- 作者指出需开展更大队列研究并纳入更多早期肿瘤患者,以提高该非侵入性诊断工具的稳健性。
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)。
原文
INTRODUCTION: Profiling volatile organic compounds in exhaled breath enables the diagnosis of several types of cancer. In this study we investigated whether a portable point-of-care version of an electronic nose (e-nose) (Aeonose, [eNose Company, Zutphen, the Netherlands]) is able to discriminate between patients with lung cancer and healthy controls on the basis of their volatile organic compound pattern. METHODS: In this study, we used five e-nose devices to collect breath samples from patients with lung cancer and healthy controls. A total of 60 patients with lung cancer and 107 controls exhaled through an e-nose for 5 minutes. Patients were assigned either to a training group for building an artificial neural network model or to a blinded control group for validating this model. RESULTS: For differentiating patients with lung cancer from healthy controls, the results showed a diagnostic accuracy of 83% with a sensitivity of 83%, specificity of 84%, and area under the curve of 0.84. Results for the blinded group showed comparable results, with a sensitivity of 88%, specificity of 86%, and diagnostic accuracy of 86%. CONCLUSION: This feasibility study showed that this portable e-nose can properly differentiate between patients with lung cancer and healthy controls. This result could have important implications for future lung cancer screening. Further studies with larger cohorts, including also more participants with early-stage tumors, should be performed to increase the robustness of this noninvasive diagnostic tool and to determine its added value in the diagnostic chain for lung cancer.
引用
- van de Goor R, van Hooren M, Dingemans AM, et al. Training and Validating a Portable Electronic Nose for Lung Cancer Screening. J Thorac Oncol. 2018. PMID: 29425703. doi:10.1016/j.jtho.2018.01.024
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。