The analysis of the inflammation in exhaled air is a challenge in respiratory diseases

Aim: To assess the feasibility of electronic nose (e-nose) of Volatile Organic Compounds to classify respiratory diseases and establish protocol of use in paediatrics.

Methods: We analysed exhaled air samples from 67 patients (14 healthy, 34 asthmatic, 19 other diseases) who attended the outpatient clinics(age mean 11.43 years). We prepared samples according to the standardized protocol. We used the Cyranose 320®e-nose. For each patient, 3 samples in different types of bags were recruited (transparent plastic and two types of tedlar). We used machine learning models (random forest) to generate an algorithm to classify children according to disease.

Results: There are statistically significant differences in 30 of the 32 sensors between Plastic and Tedlar bags. Cronbach?s alpha was below 0.5 in all cases, so the consistency of the measurements is low. A comparative analysis was performed to distinguish between the healthy and those with asthma, with statistically significant differences in 27 of the 32 sensors. Also, in 31/32 sensors between asthma vs other diseases significant differences were found(p<0.05).

The algorithm predicted asthma with high sensitivity and accuracy(over 90%). The negative and positive predicted values were over 90%. There are not statically significant in the sensibility and accuracy according to time analysis (30??,60??,120??,and?120??) and type of bags used.

Conclusions: The e-nose seems a useful tool for screening asthma in paediatrics, previous studies were only focused among adult population. However, standard technique for e-nose should be developed to facility their use.