Creating A New Test for Lung Cancer

Researcher Fred Hirsch is Developing a Breathalyzer

By Lisa Marshall

(May 2015)  Fred Hirsch can already envision it: An at-risk patient—perhaps a former smoker—goes into a grocery store, sits down at a kiosk and exhales a long slow breath into a plastic tube.

In an instant, a tiny sensor embedded within that device begins to “smell” her breath, identifying a unique chemical signature that hints at a brewing cancer within her lung. Within minutes, the test results appear on a computer screen, suggesting she see a doctor for further screening. The cancer is caught early, and a life is saved.

“I know it sounds like a dream, but I don’t think it is entirely unrealistic someday,” says Hirsch, MD, PhD, a researcher and professor in the CU School of Medicine’s Division of Medical Oncology.

Having lost his own father-in-law to the disease just as Hirsch was finishing up medical school back home in Copenhagen, the physician-scientist has spent the last 25 years working to improve the grim odds of surviving the world’s most deadly cancer. Now, as the lead investigator of a new non-invasive, affordable breathalyzer screening test for lung cancer, he predicts that—if all goes well—the game-changing device could be on the market within 5 years. And someday do-it-yourself versions might even be available. “It could make a huge impact on lung cancer mortality,” he says.

The idea of using smell as a harbinger of disease is nothing new. As far back as 2000 BC, ancient Greek and Chinese practitioners used the scent of stool, urine, skin and breath as a key clue in the diagnosis of disease. A distinctly fishy breath odor has long been associated with liver disease, while the smell of acetone hints at diabetes. A recent review in the journal Sensors links 50 diseases with distinct aromas.

Scientists have also known for years that long before a cancer is detected, subtle metabolic changes begin to occur throughout the body. These changes, and/or a tumor itself, can alter the make-up of the cocktail of chemicals, or volatile organic compounds (VOC), we emit. In one case, reported in The Lancet in 1989, an otherwise healthy dog owner made an appointment to see her dermatologist after her dog began sniffing at a mole on her leg. It turned out to be melanoma. Since then, teams around the globe have shown that animals and insects with keen olfactory systems can distinguish between cancer and non-cancer tissues. “Trained dogs can smell cancer very quickly,” says Hirsch. “But you cannot use that in a clinical practice.”

Enter Hossam Haick, PhD, a chemical engineering professor from the  Technion Institute of Technology in Israel. In an attempt to, as he puts it, “imitate the canine olfactory system” via nanotechnology, Haick in 2011 developed a sensor called a NaNose, which can measure VOCs in breath with unprecedented precision.  

When Israeli pulmonologist Nir Peled, MD, PhD, traveled to Colorado for a fellowship in Hirsch’s lab, he introduced Haick and Hirsch, and a global research project was born. “Professor Hirsch is a brilliant clinician and researcher who knows the unmet need very sharply,” says Peled, now a thoracic oncologist and professor at Tel Aviv University. “It seems to be a very good collaboration.”

Hirsch and Peled have since collected breath samples of hundreds of patients in the United States and Israel for analysis with the NaNose. Thus far, studies reveal the test may not only be able to determine whether someone has lung cancer or not, but also what stage it is.

“The data are only preliminary but are certainly encouraging,” says Gregory Masters, MD, an expert spokesperson for the American Society of Clinical Oncologists (ASCO). It could lead to “more widespread distribution of screening and potentially a less costly and more effective manner for finding lung cancer, the leading cause of cancer death in the United States.”

One study, presented by Peled in May 2014 at the ASCO annual meeting, looked at 358 subjects – 213 with diagnosed lung cancer, and 145 without. The breathalyzer was able to distinguish with more than 80 percent accuracy between those with and without cancer and distinguish with 78 percent accuracy between early and late stage cancer.

Another in-vitro study, looking at the “headspace” or contents of the gas above the cell lines in a lab suggested that different sub-types of lung cancer, fueled by different molecular drivers, or “oncogenes,” may emit different VOC fingerprints which could be detected via breath. In recent years, molecularly targeted drugs have been developed—with great success—to treat such subsets. But the current way of identifying which kind a patient has—via DNA sequencing—can be expensive and time-consuming. “Detecting and monitoring the metabolic signature associated with cancer-specific genetic mutations could be faster and easier than conventional gene-profiling methods,” Peled wrote in the study. “This feature would help to improve drug selection, thereby increasing the clinical benefit for the patients.”

Because the unique molecular signature or “smell” of cancer is believed to retreat as a cancer fades, the technology could also potentially be used to determine whether a treatment is working or not.

Haick and his colleagues recently developed a version that plugs into a home computer’s USB port, opening the door to a do-it-yourself application someday.

But for now, oncologists stress, the most exciting prospect is the test’s potential to catch the disease early, when it is still easily treatable.

“The majority of lung cancer patients present with stage 4 metastatic disease,” says Paul Bunn, MD, a distinguished professor in the Division of Medical Oncology, noting that only about one in five catch it when it can still be addressed with surgery. “Until recently there has been no way of diagnosing it early.”

With 224,000 new cases and 160,000 deaths, lung cancer kills more people each year than colon, breast and pancreatic cancer combined. According to the National Institutes of Health, the five-year survival rate is about 16.6 percent, far worse than for colon cancer (64 percent) or breast cancer (89 percent). More than half of people diagnosed with lung cancer die within one year.

In 2011, a study of 53,000 current or former smokers found that screening by low-dose CT scans was more effective than X-rays in catching lung cancer early, and could reduce mortality by 20 percent. But the study also found a high rate of false positives, with 90 percent of the suspicious nodules ending up to be benign when biopsied. “That can be both psychologically and physiologically stressful,” says Hirsch.

Initially, he believes the breathalyzer could be used as an adjunct to the scan, helping to either reduce unnecessary follow-ups or prompt necessary action. It could not, in and of itself, diagnose cancer. More research must be done before the test is ready for prime time. CU recently began enrolling patients for a 600-patient clinical trial. But Hirsch has high hopes.

“This could totally revolutionize lung cancer screening and diagnosis,” he says. “And the fact is: If you diagnose it at a very early stage, lung cancer is curable.”