A Singapore company has developed a breathalyzer test that can detect COVID-19 within a minute. The breath test, which uses in-built machine learning algorithms, claims to have achieved more than 90 per cent accuracy (with sensitivity of 93 percent and specificity of 95 percent) in a pilot clinical trial involving 180 patients.<\/p>\n
Dr. Jia Zhunan and Mr Du Fang, who developed the breathalyzer, are graduates from National University of Singapore. They founded Breathonix with the support of NUS Graduate Research Innovation Programme, a scheme that encourages the University\u2019s talented graduate students and research staff to establish and run high potential start-ups based on deep technologies.<\/p>\n
The breathalyzer works by detecting Volatile Organic Compounds (VOCs) present in a person\u2019s exhaled breath.<\/p>\n
According to Dr. Jia, Chief Executive Officer of Breathonix, VOCs are consistently produced by various biochemical reactions in human cells. \u201cDifferent diseases cause specific changes to the compounds, resulting in detectable changes in a person\u2019s breath profile. As such, VOCs can be measured as markers for diseases like COVID-19.\u201d<\/p>\n
When one blows into a disposable mouthpiece connected to the breath sampler, a machine learning software analyses the chemical compounds of the exhaled breath that is collected and fed into a mass spectrometer for measurement. The computer then generates the result in less than a minute.<\/p>\n
\u201cThe disposable mouthpiece that our system uses has a one-way valve and a saliva trap, preventing inhalation and any saliva from entering the machine. This makes cross-contamination unlikely,\u201d says Mr Du, Chief Operating Officer of Breathonix.<\/p>\n
The test is easy to administer and doesn\u2019t require trained staff or a lab facility and it could be deployed for mass screening in places with high human traffic like airports.<\/p>\n
The clinical trial for the test is ongoing, and if assessed to be suitable it hopes to get regulatory approval by early next year.<\/p>\n
The cost is estimated to be around $20 per test.<\/p>\n