Rationale: Lung cancer, a major cause of morbidity and mortality throughout the world, is also a frequent comorbidity in COPD. When using exhaled breath profiles by electronic nose in the discrimination between COPD and lung cancer the data have shown definite overlap (Dragonieri Lung Canc. 2009).
Aim: To determine diagnostic accuracy of exhaled breath analysis by SpiroNose for COPD vs. lung cancer and to identify those patients who, based on exhaled breath profiles, are allocated in the overlap area between both diseases.
Methods: This was a cross-sectional study on outpatients with standardized diagnosis of COPD and/or lung cancer. During spirometry, fingerprints from exhaled breath were collected in triplicate by an eNose based on 3 identical sensor arrays at the rear end of a pneumotach. Data-analysis involved signal processing, environment correction and statistics based on principal component analysis, followed by discriminant analysis (Matlab2014).
Results: Exhaled breath data of 29 lung cancer (age 63±9.5yr) and 30 COPD (age 66±8.4) patients were available. Principal component 1 (p<0.01) showed a significant difference between COPD and lung cancer patients with a cross-validation value of 80%. 7 lung cancer patients were classified as COPD. Patient characteristics revealed that 6 of these 7 patients were also diagnosed with COPD in addition to lung cancer.
Conclusion: Exhaled breath analysis by SpiroNose is able to discriminate between COPD and lung cancer. Patients with a double diagnosis are correctly allocated in the overlap area. However, the diagnosis of lung cancer is not unambiguous in patients with comorbid COPD, highlighting the need for optimizing the sensor arrays.