scorecardresearch Skip to main content

How your breath could help doctors diagnose illness

A patient used a breathalyzer for fungal infections as Dr. Xinewi Yu helped monitor his breathing. Jessica Rinaldi/Globe Staff

In a sunny eighth-floor office in the Dana Farber Cancer Center, Mark Rolon removed his face mask and began breathing into a tube that looked like a cross between a breathalyzer and a peace pipe. The 46-year-old was in good spirits, which says a lot, given the nearly two-year medical odyssey he’d endured: chemotherapy treatments for a rare blood disorder, a bone marrow transplant, then pneumonia and broken ribs from coughing.

“I was a mess,” he said.

In December of last year, his doctors approached him about participating in a study. A Brigham and Women’s Hospital doctor, together with scientists at Draper Lab, was developing a device that might eventually get patients like him on the right treatment program, helping them get better, faster. Would he want to take part? Rolon was game, particularly when he learned all he needed to do was breathe into a tube for four minutes.

Breathing is the most basic human function: If we’re doing it right, we’re technically alive. Long before doctors began collecting urine and blood samples, or peering into our bodies with X-rays or CT scans, the earliest practitioners of medicine relied on breath to infer basic information about a person’s health. Doctors in ancient Greece would ask their patients to exhale and if a specific odor emerged (perhaps a fishy smell) it would help them determine the cause of the disease. The idea of this inner breath, known as the pneuma, was seen as a kind of life force that sustained the vital organs and consciousness.


“It was the oldest form of diagnostic, and the sensor was the human nose,” Dr. Joachim Pleil, a breath researcher at the Environmental Protection Agency, said of the early scientists. “Hippocrates would teach his medical students that the first thing that they could do was smell the guy’s breath to indicate different states of disease.”


As modern medicine evolved, reliance on breath diminished, and today there are only a few diagnostic breath tests that have FDA approval for clinical practice. But scientists are beginning to revisit the ancient methods with new tools, resulting in a swath of breath-related research that spans the gamut from measuring “cell breath” in bodily fluids to using breath biomarkers to create personalized patient treatment programs. Researchers believe that breath-related tests could someday help diagnose certain forms of cancer.

A team of MIT and Harvard students recently won MIT’s $100,000 entrepreneurship competition for developing a smartphone-enabled breath test for lung cancer. The company, Astraeus Technologies, is a direct nod to its medical forebears — it’s named after the Greek god of wind.

The surge in interest led Pleil, the editor of the Journal of Breath Research, to publish a paper this past March outlining a path by which researchers would go about “expanding the scope” of their work.

The Dana Farber study in which Rolon took part was the work of Dr. Sophia Koo, who specializes in infectious diseases at Brigham and Women’s hospital. In 2007, early in her residency, Koo had noticed that many of her cancer and transplant patients were succumbing to pneumonia, which is notoriously difficult to diagnose and treat. Because the lung tissue is so delicate, it’s tough to probe, so doctors often have a hard time determining whether a patient has a viral or bacterial infection — or worse, fungal pneumonia, which can be particularly fatal.


On an typical fall day, we inhale as many as 200 airborne fungal spores, which a healthy person’s white blood cells can easily disarm. But in a patient like Rolon, whose immune system was decimated, these inhaled spores could quickly occupy the lungs and lead to fungal pneumonia.

If doctors suspect pneumonia, they will often start a patient on course of antibiotics. If there’s no response, they’ve ruled out a bacterial infection. If it turns out to be fungal, however, precious time has been lost, 11 days on average. Even if treated, fungal pneumonia can be fatal in 30 percent of cases.

Koo hypothesized that she might be able to detect the presence of these foreign organisms if she could take a snapshot of her patient’s breath. With every exhale, we release thousands of airborne compounds that are a result of our bodies’ inner workings, like the carbon dioxide that we process through our lungs or the gases our stomachs release in the digestion process. If a fungus had taken up residence in a patient’s lung, she reasoned, it too would be giving off compounds that would be chemically distinct from that of humans.

“It’s almost like this alien inside your lung that has this totally different metabolism,” she said. If she could isolate them through a four-minute-long breath test, she could diagnose fungal pneumonia and immediately start a patient on the intensive six-month treatment program.

Koo rigged up a device that resembled a breathalyzer, which is typically used to determine blood alcohol levels. Her theory proved fruitful, and Koo began to see patterns emerge almost immediately that matched the metabolic markers of the fungus in her patients’ breath, confirming that they had developed fungal pneumonia. For the past several years, she’s continued to map these patterns as patients undergo treatment.


In January, Koo’s research took a big step forward when she linked up with scientists at Draper Lab who had developed the microAnalyzer, a shoebox-size device that detects vapor particles in the air as tiny as a few parts per trillion. The Draper team had been using it in military settings for more than a decade, and it now helps protect the astronauts on the International Space Station from toxic gas leaks. But they were searching for uses in a medical setting.

Tim Postlethwaite, a program manager in Draper’s biomedical solutions department, said he sees great promise in the partnership. Breath tests are noninvasive, meaning they can be done almost anywhere. And because the microAnalyzers are relatively small, they can easily be put in a health clinic, or even transported between locations.

“This could be a path toward a point-of-care diagnostic test,” Postlethwaite said.

In some cases, breath testing can find evidence of a disorder faster than a traditional blood test, said Anil Modak, a research director who is developing breath diagnostics to detect pulmonary infections, liver diseases, and ensure drug compliance at Cambridge Isotope Laboratories in Tewksbury. He said breath testing can allow for a more personalized look at how we metabolize medications.


“We need to personalize medicine, we have to stop practicing medicine with trial and error,” he said. “This area of breath testing is vast.”

Last May, Koo won a $50,000 health and technology grant from the Brigham to extend her work to detect bacterial and viral pneumonias in patients, and she is working with Draper to start the process to gain FDA approval for her breathalyzer tool. Koo says the ability to have a noninvasive breath test for respiratory disorders is a tremendous step forward for clinical care, one that can have a lasting impact in the developing world, where pneumonia is one of the leading causes of death in children under age 5.

“People have been looking at breath analysis ever since Hippocrates,” she said. “It took awhile to build the technology to understand it.”

Janelle Nanos can be reached at Follow her on Twitter @janellenanos.