Newswise — With every respiration, people release over 1,000 discrete compounds, generating a distinct chemical impression or "exhalation impression" packed with indications regarding the body's internal state.
For many years, researchers have attempted to utilize that knowledge, relying on canines, rodents, and even insects to physically detect cancer, diabetes, tuberculosis, and other illnesses.
CU Boulder and NIST scientists have achieved a significant breakthrough in the endeavor to diagnose ailments by analyzing exhaled breath. They have announced that a novel laser-based breathalyzer, empowered by artificial intelligence (AI), can identify COVID-19 instantly with exceptional precision.
On April 5, the findings were released in the Journal of Breath Research.
Qizhong Liang, the lead author and a PhD candidate in JILA and the Department of Physics at CU Boulder, stated, "Our outcomes showcase the potential of breath examination as a substitute, fast, and non-invasive technique for COVID-19 testing, while also emphasizing its incredible potential for detecting various conditions and illnesses." JILA is a joint endeavor between NIST and CU Boulder.
The team, consisting of physicists, biochemists, and biologists from multiple disciplines, is currently directing their attention towards various other illnesses, anticipating that the "frequency comb breathalyzer," created utilizing Nobel Prize-winning technology from CU, may transform medical diagnostics.
Jun Ye, the senior author, a JILA fellow, and an adjunct professor of physics at CU Boulder, commented that "There is an actual, predicted future in which you could have your breath assessed together with your height and weight when you visit the doctor... Alternatively, you could exhale into a mouthpiece incorporated into your phone and receive real-time health updates." "The possibilities are boundless," he added.
A COVID-born collaboration
As early as 2008, Ye's laboratory had announced that frequency comb spectroscopy, which essentially utilizes laser light to differentiate one molecule from another, could potentially identify disease biomarkers in human breath.
At that time, the technology lacked sensitivity, and more significantly, the capacity to associate specific molecules with disease states, which prevented them from testing it for diagnosing illnesses.
Nevertheless, Ye's team has subsequently increased sensitivity by a factor of a thousand, making it possible to detect trace molecules at the parts-per-trillion level. They have also harnessed the power of AI.
Liang stated that "Molecules exhibit an increase or decrease in concentration when related to specific health conditions. Machine learning algorithms examine this data, recognize patterns, and establish criteria that we can employ to anticipate a diagnosis."
Due to the rapid spread of SARS-CoV-2 across the country and the growing frustration surrounding the prolonged response times of existing tests, it was time to evaluate the system on individuals. Ye, being a physicist, had never worked with human subjects before, so he enlisted assistance from CU's BioFrontiers Institute, an interdisciplinary center for biomedical research that was spearheading the COVID testing initiative on campus.
The National Science Foundation and the National Institutes of Health funded the research.
Non-invasive, fast, chemical-free
From May 2021 to January 2022, the research team collected breath samples from 170 CU Boulder students who had taken a polymerase chain reaction (PCR) test in the preceding 48 hours by submitting a saliva or a nasal sample.
Half of the participants had tested positive, while the other half had tested negative. For safety reasons, the volunteer participants arrived at an outdoor campus parking lot, provided a breath sample into a collection bag, and then left it for a laboratory technician who was waiting at a safe distance.
Overall, the process took less than one hour from collection to result.
In comparison to the gold standard COVID test, PCR, the results from the breathalyzer corresponded 85% of the time. An accuracy level of 80% or higher is considered "excellent" for medical diagnostics.
The researchers believe that the accuracy would have likely been higher if the breath and saliva/nasal swab samples had been collected simultaneously.
Unlike a nasal swab, the breathalyzer is non-invasive and does not cause discomfort to the user. Additionally, unlike a saliva sample, users are not required to refrain from eating, drinking, or smoking before using it. The breathalyzer does not require costly chemicals to break down the sample, and it could potentially be used on individuals who are not conscious.
But there is still much to be learned, said Ye.
Jun Ye said, "With one breath, we can collect a large amount of data points from an individual, but the challenge is to understand the correlations between the molecules and specific medical conditions."
Building a smaller breathalyzer
Currently, the "breathalyzer" is comprised of a complicated system of lasers and mirrors that takes up the space of a banquet table.
The breathalyzer system involves a tube that pipes in the breath sample, which is then analyzed using lasers and mirrors. The lasers emit mid-infrared light at various frequencies, and the mirrors reflect the light back and forth through the molecules multiple times. This process enables the light to travel approximately 1.5 miles before being detected and analyzed.
The breath analysis system employs a pipe that delivers the exhaled air sample, which undergoes analysis using lasers and mirrors. The lasers emit mid-infrared light at diverse frequencies, and the mirrors reflect the light repeatedly through the molecules. This procedure allows the light to travel roughly 1.5 miles before detection and analysis.
There are ongoing efforts to shrink such systems to a chip size, enabling Liang's vision of "on-the-go self-health monitoring in real-time." However, the possibilities do not stop there.
Molecular biologist and co-author Leslie Leinwand, who is also the chief scientific officer for BioFrontiers and a co-author of the study, envisions the potential of finding a signature in breath that can detect pancreatic cancer before any symptoms appear, which would be a remarkable achievement.
Other researchers are currently working on developing a Human Breath Atlas, which would chart every molecule present in exhaled breath and link them to health outcomes. Liang aims to contribute to this effort by collecting a larger number of breath samples.
The team is also working with pediatric and respiratory specialists at the CU Anschutz Medical Campus to investigate how the breathalyzer can be used not only to diagnose diseases but also to help scientists gain a better understanding of them. This could provide insights into immune responses, nutritional deficiencies, and other factors that may contribute to or worsen illness.
Ye compared the development of the breathalyzer to the evolution of dogs, which have a remarkable sensitivity to smell due to thousands of years of evolution. According to him, the breathalyzer is just at the beginning of its training, and the more it is taught, the more intelligent it will become.
RSS