Research on Intelligent Electronic Nose Based on Exhaled Breath to Detect Halitosis

Abstract

Introduction Halitosis refers to the offensive or unpleasant gas emitted from the mouth, also known as oral odor. Previous studies have shown that 80-90% of halitosis patients are caused by oral diseases[1]. The patients with thick tongue coating and periodontal disease usually have obvious halitosis and the concentrations of hydrogen sulfide and methyl mercaptan in the breath of the oral cavity were significantly higher than those of the control group. Currently there are many methods for the diagnosis of halitosis in clinic. The commonly used direct measurement techniques include organoleptic measurement, gas chromatography and portable sulfide monitor[2]. Organoleptic measurement, which involves using the human nose to score the intensity of odours emanating from the patient’s mouth, is considered the gold standard for halitosis measurement[3]. Gas chromatography is expensive and complex to operate, which requires professionals to operate. Commercial portable gas chromatography such as Oral Chroma™ (Oral Chroma™, Abimedical, Abilit Corp., Osaka, Japan) can detect the concentration of hydrogen sulfide, methyl mercaptan and dimethyl sulfide, but its price is still relatively expensive. Portable sulfide monitor such as the Halimeter (Interscan Corp., Chatsworth, LA, USA) is specially used to measure the total concentration of sulfide. But it takes a long time to measure and is only suitable for clinical application.Based on the breath of patients with halitosis, this study uses an electrochemical sensor array to construct a small and intelligent electronic nose device for detecting bad breath, which realized the rapid quantitative detection of hydrogen sulfide and methyl mercaptan, and gave the preliminary screening results of halitosis grade and oral diseases. Method Halitosis samples were collected in the dental clinic. All patients had not smoked, drank or ate spicy food before the examination. The dentist first checks the patient's oral health, and then uses organoleptic measurement to evaluate the level of Halitosis. This information will be marked on the sampling bags. Finally, 100 ml exhaled air was collected with Tedlar sampling bag.The volatile sulfur compounds in 81 samples of breath were qualitatively and quantitatively detected using GC-MS(gas chromatography and mass spectrometry) in the laboratory. ROC method was used to establish the threshold model of halitosis grade. In addition, a diagnosis model of orogenic diseases based on linear discriminant analysis, logistic regression and support vector machine was established. On the basis of the measurement result, a small intelligent electronic nose device for halitosis detection was constructed with an electrochemical sensor array, and the calibration was carried out using hydrogen sulfide and methyl mercaptan standard gases. The device can establish a connection with the mobile phone through Bluetooth. The level threshold model of halitosis and the diagnosis model of oral diseases are transplanted into the app on the mobile phone. The diagnosis results can be calculated by substituting the concentration value transmitted by the device into the model. Results and Conclusions The halitosis level threshold model and oral disease diagnosis model show that halitosis can be heard when the total concentration of methyl mercaptan and hydrogen sulfide is greater than 0.2ppm. When the total concentration is greater than 0.4ppm, the patient's halitosis is very obvious. The severity of periodontal disease (including periodontitis and gingivitis) is significantly correlated with the concentration of methyl mercaptan. When the concentration ratio of methyl mercaptan to hydrogen sulfide is greater than 1:1, patients usually have obvious periodontitis. The size of the electronic nose device based on the diagnosis model is 50 mm × 50 mm × 25 mm and the detection range of methyl mercaptan and hydrogen sulfide is 0 to 2 ppm with a detection limit of 0.05 ppm. In addition, the response time and recovery time of the sensor array are both within 15s, which greatly shortens the time compared with the common clinical detection methods. The standard gases of hydrogen sulfide and methyl mercaptan are mixed with air to simulate the breath samples with different odor levels. The gas detection process is controlled by smart phones, and the respective concentrations of hydrogen sulfide and methyl mercaptan can be calculated within 15 seconds. Taking threshold model as the evaluation standard, the accuracy of the detection results is more than 90%, which indicates that the device has the potential to become a household or portable electronic nose for halitosis detection.