| Title: | An Experimental Apparatus for E-Nose Breath Analysis in Respiratory Failure Patients |
| Author(s): | Bax C; Robbiani S; Zannin E; Capelli L; Ratti C; Bonetti S; Novelli L; Raimondi F; Di Marco F; Dellaca RL; |
| Address: | "Department of Chemistry, Materials and Chemical Engineering 'Giulio Natta' (DCMC), Politecnico di Milano, 20133 Milano, Italy. TechRes Lab, Department of Electronics Information and Bioengineering (DEIB), Politecnico di Milano, 20133 Milano, Italy. Unit of Pneumology, Azienda Ospedaliera Socio Sanitaria Territoriale Papa Giovanni XXIII, 24127 Bergamo, Italy. Department of Health Sciences, Universita degli Studi di Milano, 20142 Milano, Italy" |
| DOI: | 10.3390/diagnostics12040776 |
| ISSN/ISBN: | 2075-4418 (Print) 2075-4418 (Electronic) 2075-4418 (Linking) |
| Abstract: | "Background: Non-invasive, bedside diagnostic tools are extremely important for tailo ring the management of respiratory failure patients. The use of electronic noses (ENs) for exhaled breath analysis has the potential to provide useful information for phenotyping different respiratory disorders and improving diagnosis, but their application in respiratory failure patients remains a challenge. We developed a novel measurement apparatus for analysing exhaled breath in such patients. Methods: The breath sampling apparatus uses hospital medical air and oxygen pipeline systems to control the fraction of inspired oxygen and prevent contamination of exhaled gas from ambient Volatile Organic Compounds (VOCs) It is designed to minimise the dead space and respiratory load imposed on patients. Breath odour fingerprints were assessed using a commercial EN with custom MOX sensors. We carried out a feasibility study on 33 SARS-CoV-2 patients (25 with respiratory failure and 8 asymptomatic) and 22 controls to gather data on tolerability and for a preliminary assessment of sensitivity and specificity. The most significant features for the discrimination between breath-odour fingerprints from respiratory failure patients and controls were identified using the Boruta algorithm and then implemented in the development of a support vector machine (SVM) classification model. Results: The novel sampling system was well-tolerated by all patients. The SVM differentiated between respiratory failure patients and controls with an accuracy of 0.81 (area under the ROC curve) and a sensitivity and specificity of 0.920 and 0.682, respectively. The selected features were significantly different in SARS-CoV-2 patients with respiratory failure versus controls and asymptomatic SARS-CoV-2 patients (p < 0.001 and 0.046, respectively). Conclusions: the developed system is suitable for the collection of exhaled breath samples from respiratory failure patients. Our preliminary results suggest that breath-odour fingerprints may be sensitive markers of lung disease severity and aetiology" |
| Keywords: | Covid-19 breath analysis diagnosis electronic nose odour analysis; |
| Notes: | "PubMed-not-MEDLINEBax, Carmen Robbiani, Stefano Zannin, Emanuela Capelli, Laura Ratti, Christian Bonetti, Simone Novelli, Luca Raimondi, Federico Di Marco, Fabiano Dellaca, Raffaele L eng Switzerland 2022/04/24 Diagnostics (Basel). 2022 Mar 22; 12(4):776. doi: 10.3390/diagnostics12040776" |
