Respiratory system diseases breath analysis system

Since breath analysis is not invasive, it is a risk-free method that can be sampled repeatedly and has advantages over other diagnostic methods that do not require a professional healthcare worker. In our study, a breath analyzer was developed by using sensors measuring ppb level. Out of 20 non-smoker respiratory patients; The breaths of 13 individuals diagnosed with pneumonia, 3 individuals diagnosed with lung cancer and 4 individuals diagnosed with COPD were analyzed. The breaths of 5 patients with COVID-19, 6 healthy smokers and 7 healthy non-smokers were analyzed. Acetone (C3H6O), Ammonia (NH3), Carbon Monoxide (CO), Nitric Oxide (NO), Ethylene (C2H4), Hydrogen Sulfide (H₂S), Nitrogen Dioxide (NO₂), Carbon Disulfide (CS2), Formaldehyde (CH2O), Ten molecules, including hydrogen (H2), were examined. The results were analyzed with the Origin-Two sample t-test and using the 'Random Forest Algorithm'. As a result of the study, NO2 molecule (p=0.0227), H2S molecule (p=0.03449) and SO2 molecule (p=0.00261) were significantly different from those with respiratory disease. The NO2 molecule (p =0.02337), H2S molecule (p=0.02175) and SO2 molecule (p=0.0056) were significantly different from those with pneumonia. NH3 molecule (p =0.00308), NO2 molecule (p=0.00049), NO molecule (p=0.00792) and H2S molecule (p=0.0128) were significantly different in people with COVID-19 compared to healthy people. When the data of smokers and non-smokers were examined, many molecules were found to be significant. These; CH2O (p=0.00008), CO (p=0.00005), C2H4 (p=0.0025), SO2 (p=0.02), C3H6O (p=0.02) and H2 molecule (p=0.03). Our data has been successfully estimated by Random Forest Algorithm.