Diagnosing Lung And Gastric Cancers Through Exhaled Breath Analysis By Using Electronic Nose Technology Combined With Pattern Recognition Methods

Abstract

Lung cancer (LCa) and gastric cancer (GCa) are two of the most lethal cancers worldwide. Unspecific clinical symptoms and the lack of defined risk factors often delay the diagnosis of the disease, which could high the mortality rate. The aim of the present study is to evaluate the capability of an electronic nose (e-nose) based on metal-oxide semi-conductor sensors combined with pattern recognition methods to discriminate between patients groups with LCa, GCa, and healthy controls (HC). Breath samples were collected from 35 volunteers containing 13 HC, 14 LCa, and 8 GCa patients. The e-nose dataset was treated with principal component analysis (PCA), discriminant function analysis (DFA), and support vector machines (SVM). As result, PCA and DFA have shown good discrimination between data-points of breath samples related to HC, LCa and GCa patients. The SVMs method reached a 100% success rate for the recognition of the analyzed three groups. In the light of these results, we can state that the presented e-nose system demonstrates that an inexpensive and non-invasive approach based on exhaled breath analysis could be considered a reliable screening tool to differentiate between the three studied groups.