A team of international researchers recently unveiled a nano array that can identify the chemical signatures of 17 different diseases, possibly bringing us closer to the day when doctors might be able to use a medical tricorder a la Star Trek to instantly diagnose a patient’s conditions.
Though it isn’t exactly a new idea – Hippocrates wrote about the correlation between breath odors and disease back in 400 B.C. and traditional Chinese medicine has long seen halitosis as an indication of an unbalanced qi – using breath tests to diagnose and monitor bodily disorders and disease is a research field that has been gaining momentum in recent years. And for good reason too. It would be the ultimate diagnostic test – potentially inexpensive and painless (not to mention a godsend for anyone with a fear of needles), and it would be able to deliver results fairly quickly too.
That said, in order for this to happen, breathalyzers need to be able to identify more than one disease at any given time. The technologies developed to date have a limited scope and are designed to detect only one kind of disease, such as a particular type of cancer or diabetes. And while there have been attempts to identifya wider scope of ailments, there has been no real breakthrough at distinguishing different diseases in a breath sample – till now.
Key to these new findings is the identification of the different odor signatures that each disease has, and the answer to this lies in our breath. Every time we exhale, we expel nitrogen, carbon dioxide, oxygen as well as a host of volatile organic compounds (VOCs), the composition and concentration of which changes depending on our state of health.
“Just as each of us has a unique fingerprint that distinguishes us from others, each disease has a chemical signature that distinguishes it from other diseases and from a normal state of health,” explains lead researcher Hossam Haick, an Israeli nanotech expert whose name is synonymous with disease-detecting sensors. “These odor signatures are what enables us to identify the diseases using the technology that we developed.”
To do this, the researchers used an artificially intelligent nano array of carbon nanotubes and gold particles developed by Haick to detect the individual components in breath samples collected from 1,404 patients spanning five countries: the United States, Israel, France, Latvia and China. Eight hundred and thirteen of them had one of 17 different diseases, including kidney cancer or Parkinson’s disease. The rest were healthy.
Two breath samples were collected from each patient: one for chemical mapping via mass spectrometry, and the other for clinical diagnosis using the AI nano array
Using mass spectrometry to identify the breath components associated with the diseases, they found that each disease produces a unique volatile chemical breathprint, based on differing amounts of 13 components.
The data was then entered into a computer system, which identified disease with an accuracy of 86 percent. Since factors such as gender, age, smoking habits and geographic location can affect the accuracy of the results, the team also examined the effects of these elements on the breath samples and found they did not impair the array’s sensitivity. Equally important was the fact that the presence of one disease did not prevent the detection of others – a prerequisite for developing a portable diagnostic device to detect various diseases in a noninvasive and inexpensive manner.
Likening the AI to the way a dog recognizes a scent, Haick says the system can be taught to associate a breathprint with a particular disease. “We let it smell a given disease but instead of a nose we use chemical sensors, and instead of the brain we use the algorithms. Then in the future, it can recognize the disease as a dog might recognize a scent,” he told Smithsonian.com.
Such a system would do away with the need for costly and invasive procedures such as biopsies, and could also be used to detect early-stage diseases, thus increasing one’s chances of survival.
“Breath is an excellent raw material for diagnosis,” says Haick. “It is available without the need for invasive and unpleasant procedures, it’s not dangerous, and you can sample it again and again if necessary.”
In addition, its application extends beyond medical diagnostics and as long as the AI receives the appropriate training, it can be used to detect everything from food spoilage to explosives.
Haick explains the team’s work in the video below.
The study was published in ACS Nano.