Detection of lung cancer in exhaled breath with an electronic nose using support vector machine analysis

Abstract

Lung cancer is one of the most common malignancies and has a low 5-year survival rate. There are no cheap, simple and widely available screening methods for the early diagnostics of lung cancer. The aim of this study was to determine whether analysis of exhaled breath with an artificial olfactory sensor using support vector analysis can differentiate patients with lung cancer from healthy individuals and patients with other lung diseases, regardless of the stage of lung cancer and the most common comorbidities. Patients with histologically or cytologically verified lung cancer, healthy volunteers and patients with other lung diseases (e.g. chronic obstructive pulmonary disease (COPD), asthma, pneumonia, pulmonary embolism, benign lung tumors) were enrolled in the study. Breath sample collection and analysis with a Cyranose 320 sensor device was performed and data were further analyzed using a support vector machine (SVM). The SVM correctly differentiated between cancer patients and healthy volunteers in 98.8% of cases. The cancer versus non-cancer group patients (healthy volunteers and patients with other lung diseases) were classified correctly by SVM in 87.3% of cases. In the mixed diagnosis groups (only cancer, only COPD, cancer + COPD and control) all 79 out of 79 patients were predicted correctly in the cancer + COPD group, with the rate of correct prognosis in other patient groups being lower. Exhaled breath analysis by electronic nose using a SVM is able to discriminate patients with lung cancer from healthy subjects and mixed groups of patients with different lung diseases. It can also provide a certain level of discrimination between lung cancer patients, lung cancer patients with concomitant COPD, COPD alone and a healthy control group.