Lung cancer is the leading cause of cancer death around the world, and lung cancer screening remains challenging. This study aimed to develop a breath test for the detection of lung cancer using a chemical sensor array and a machine learning technique. We conducted a prospective study to enroll lung cancer cases and non-tumour controls between 2016 and 2018 and analysed alveolar air samples using carbon nanotube sensor arrays. A total of 117 cases and 199 controls were enrolled in the study of which 72 subjects were excluded due to having cancer at another site, benign lung tumours, metastatic lung cancer, carcinoma in situ, minimally invasive adenocarcinoma, received chemotherapy or other diseases. Subjects enrolled in 2016 and 2017 were used for the model derivation and internal validation. The model was externally validated in subjects recruited in 2018. The diagnostic accuracy was assessed using the pathological reports as the reference standard. In the external validation, the areas under the receiver operating characteristic curve (AUCs) were 0.91 (95% CI = 0.79–1.00) by linear discriminant analysis and 0.90 (95% CI = 0.80–0.99) by the supportive vector machine technique. The combination of the sensor array technique and machine learning can detect lung cancer with high accuracy.

(1) Background: Cooking and burning incense are important sources of indoor air pollutants. No studies have provided biological evidence of air pollutants in the lungs to support this association. Analysis of pleural fluid may be used to measure the internal exposure dose of air pollutants in the lung. The objective of this study was to provide biological evidence of indoor air pollutants and estimate their risk of lung cancer. (2) Methods: We analyzed 14 common air pollutants in the pleural fluid of 39 cases of lung adenocarcinoma and 40 nonmalignant controls by gas chromatography-mass spectrometry. (3) Results: When we excluded the current smokers and adjusted for age, the adjusted odds ratios (ORs) were 2.22 (95% confidence interval CI = 0.77–6.44) for habitual cooking at home and 3.05 (95% CI = 1.06–8.84) for indoor incense burning. In females, the adjusted ORs were 5.39 (95% CI = 1.11–26.20) for habitual cooking at home and 6.01 (95% CI = 1.14–31.66) for indoor incense burning. In pleural fluid, the most important exposure biomarkers for lung cancer were naphthalene, ethylbenzene, and o-xylene. (4) Conclusions: Habitual cooking and indoor incense burning increased the risk of lung adenocarcinoma.

For malignant pleural effusions, pleural fluid cytology is a diagnostic method, but sensitivity is low. The pleural fluid contains metabolites directly released from cancer cells. The objective of this study was to diagnose lung cancer with malignant pleural effusion using the volatilomic profiling method. We recruited lung cancer patients with malignant pleural effusion and patients with nonmalignant diseases with pleural effusion as controls. We analyzed the headspace air of the pleural effusion by gas chromatography-mass spectrometry. We used partial least squares discriminant analysis (PLS-DA) to identify metabolites and the support vector machine (SVM) to establish the prediction model. We split data into a training set (80%) and a testing set (20%) to validate the accuracy. A total of 68 subjects were included in the final analysis. The PLS-DA showed high discrimination with an R2 of 0.95 and Q2 of 0.58. The accuracy of the SVM in the test set was 0.93 (95% CI 0.66, 0.998), the sensitivity was 83%, the specificity was 100%, and kappa was 0.85, and the area under the receiver operating characteristic curve was 0.96 (95% CI 0.86, 1.00). Volatile metabolites of pleural effusion might be used in patients with cytology-negative pleural effusion to rule out malignancy.

Lung cancer is the leading cause of cancer death. For malignant pleural effusions, pleural fluid cytology is a diagnostic method, but sensitivity is low. Many patients need to undergo invasive diagnostic tests such as thoracoscopic pleural biopsy. Pleural space is an enclosed microenvironment, and the pleural fluid contains metabolites directly released from cancer cells. The objective of this study was to diagnose lung cancer with malignant pleural effusion using the volatilomic profiling method. We recruited lung cancer patients with malignant pleural effusion and patients with nonmalignant diseases with pleural effusion as controls. We analyzed the headspace air of the pleural effusion by gas chromatography-mass spectrometry. We used partial least squares discriminant analysis (PLS-DA) to identify metabolites and the support vector machine (SVM) to establish the prediction model. We split data into a training set (80%) and a testing set (20%) to validate the accuracy. A total of 68 subjects were included in the final analysis. The PLS-DA showed high discrimination with an R2 of 0.95 and Q2 of 0.58. The accuracy of the SVM in the test set was 0.93 (95% CI: 0.66, 0.998), and kappa was 0.85, and the area under the receiver operating characteristic curve was 0.96 (95% CI: 0.86, 1.00). Pathway analysis revealed disturbances in pyruvate metabolism, the tricarboxylic acid, glycolysis, and lysine degradation. The volatile metabolites identified from malignant pleural effusion of lung cancer were primarily methylated alkanes. The pleural effusion contained volatile metabolites that have high accuracy in diagnosing lung cancer with malignant pleural effusion.

IntroductionLung cancer is the leading cause of cancer death in the world. The challenge of screening for early stage lung cancer is still unresolved. The exploration of metabolites in breathe using sensor array technique may become a powerful screening tool to solve the problem.MethodsWe conducted a prospective study to enrol cases of lung cancer and controls who received surgery for gall bladder stone, hernia, hemorrhoid resection, and thoracoscopic surgery in the same hospital between July 2016 and June 2017. The alveolar air of subjects were collected under the guidance of mainstream carbon dioxide analyzer. An electronic nose composed of 32 carbon nanotubes sensors was used to measure the VOCs of the alveolar air. The diagnostic accuracy was analysed by linear discriminant analysis (LDA) using the pathological reports as the reference standard.ResultsAfter excluding 2 subjects with technical problems in sampling, 12 subjects with cancers in other sites, benign lung tumour, or metastatic lung cancer, 5 subjects received chemotherapy, 5 subjects with diabetes, 2 subjects with asthma, and 2 subjects with chronic obstructive pulmonary disease, a total of 17 cases and 105 controls were used in the final analysis. We randomly split the data into 80% for model building (training set) and 20% for validation (test set). By LDA, the accuracy, sensitivity, specificity, false positive rate, false negative rate, and ROC-AUC were 96.9%, 75.0%, 100.0%, 0%, 25%, and 0.98 (95% CI: 0.96 to 1.00) in the training set, and 84.0%, 80.0%, 85.0%, 15.0%, 20.0% and 0.84 (95% CI: 0.62 to 1.00) in the test set.ConclusionThe use of sensor array technique to explore the metabolites in breathe may become a powerful tool in the screening for lung cancer. Standardised procedures to eliminate confounding factors are warranted before clinical application.