文献速递肺结节与肺癌

基于呼出气组学的肺癌分诊工具开发与多中心验证:一项5214人前瞻性研究

Development and Multicenter Validation of a Breathomics-Based Triage Tool for Lung Cancer: A Prospective Study of 5,214 Participants.
PubMed 文献 · Chest · 2026-07-28

本研究在5214例有影像学肺部异常的症状患者中开发并验证了结合挥发性有机化合物特征与临床因素的机器学习呼出气模型。外部验证队列中模型AUC为0.850,预设排除阈值下敏感度93.1%、特异度55.9%、阴性预测值71.0%。该模型有望作为无创分诊工具用于呼吸科门诊,减少不必要的活检。

要点

  • 呼出气组学机器学习模型外部验证AUC达0.850,早期病变AUC为0.849。
  • 预设排除阈值下外部验证敏感度93.1%、特异度55.9%、阴性预测值71.0%。
  • 呼吸科预期使用亚组敏感度91.4%、特异度55.8%、阴性预测值89.0%。

核心结论

  • 呼出气组学预测模型经多中心前瞻性验证,可作为无创分诊工具有助于减少不必要的活检。

研究设计与样本

  • 前瞻性多中心诊断准确性研究,在两家院区纳入5214例有影像学肺部异常的症状患者。
  • 发现队列4669例,地理独立外部验证队列545例。
  • 采用高通量质子转移反应飞行时间质谱分析呼出挥发性有机化合物,机器学习模型整合特定挥发性有机化合物特征与临床因素,锁定后在外部队列盲法验证。

主要发现

  • 整合模型在独立内部测试集AUC为0.891(95%CI 0.864-0.915)。
  • 外部验证AUC为0.850(95%CI 0.805-0.890)。
  • 外部验证队列应用预设排除阈值,敏感度93.1%(95%CI 90.5%-95.4%),总体特异度55.9%,阴性预测值71.0%。
  • 呼吸科预期使用亚组(n=155)敏感度91.4%,特异度55.8%,阴性预测值89.0%。
  • 亚组分析显示早期病变检测AUC为0.849。

关键数值与推荐

  • 外部验证预设排除阈值:敏感度93.1%(95%CI 90.5%-95.4%),特异度55.9%,阴性预测值71.0%。
  • 呼吸科预期使用亚组(n=155):敏感度91.4%,特异度55.8%,阴性预测值89.0%。

临床怎么用

  • 用于有影像学肺部异常的症状患者分诊,高敏感度有助于排除肺癌以减少不必要的侵入性操作。
  • 呼吸科门诊场景下亚组阴性预测值达89.0%,支持其作为分诊工具整合入诊断路径。

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

BACKGROUND: The inherent false-positive rate of CT imaging necessitates efficient, noninvasive triage tools to distinguish lung cancer from benign mimics, thereby streamlining early detection and mitigating unnecessary invasive procedures. RESEARCH QUESTION: Can a machine learning-derived, molecularly resolved breathomics prediction model effectively triage patients with radiologically detected pulmonary abnormalities in a real-world symptomatic cohort? STUDY DESIGN AND METHODS: In this large-scale, prospective, multicenter diagnostic accuracy study, we enrolled 5,214 symptomatic patients with radiological lung abnormalities at two campuses. Participants were allocated into a discovery cohort (n = 4,669) and a geographically independent external validation cohort (n = 545). Exhaled volatile organic compounds were analyzed by high-throughput proton-transfer-reaction time-of-flight mass spectrometry. A machine learning model integrating a specific volatile organic compound signature with clinical factors was developed, locked, and validated blind in the external cohort. RESULTS: The integrated model achieved an area under the receiver operating characteristic curve (AUC) of 0.891 (95% CI, 0.864-0.915) in the independent internal testing set. In the independent external validation, the model maintained robust performance (AUC, 0.850; 95% CI, 0.805-0.890). In this validation cohort, applying the prespecified "rule-out" threshold, the model achieved a sensitivity of 93.1% (95% CI, 90.5%-95.4%), with an overall specificity of 55.9% and negative predictive value of 71.0%. Importantly, in the intended-use pulmonary medicine subgroup (n = 155), the sensitivity was 91.4%, with specificity and negative predictive value reaching 55.8% and 89.0%, respectively. Subgroup analyses confirmed consistent efficacy across diverse clinical scenarios, notably maintaining robust accuracy in detecting early-stage disease (AUC, 0.849). INTERPRETATION: This study represents the largest prospective validation of a mass-spectrometry-based breath test to date. The validated prediction model holds potential to serve as a robust, noninvasive triage tool. Its integration into the diagnostic pathway shows promise for enhancing efficiency and reducing unnecessary biopsies, particularly in respiratory outpatient settings.

来源:Chest,查看原文。版权归原作者所有。

引用

  1. Qin J, Sun M, Li X, et al. Development and Multicenter Validation of a Breathomics-Based Triage Tool for Lung Cancer: A Prospective Study of 5,214 Participants. Chest. 2026. PMID: 42521150. doi:10.1016/j.chest.2026.06.067

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。