Early research finds that breath analysis may eliminate the need for some diagnostic tests for patients with benign lung disease.
Not all lesions in the lung are cancerous. Histoplasmosis, which is especially common in the Ohio and Mississippi river valleys, is an infection that presents with lesions in the lung. Though it resembles cancer, it is in fact an inflammatory disorder and is rarely life-threatening. What’s more, cancers are often confused with pneumonia. This is the contention of Michael Bousamra II, M.D., an associate professor and thoracic surgeon at the James Graham Brown Cancer Center, University of Louisville.
Patients with benign diseases often come in with positive PET scans, Bousamra told Healthline. “And once a patient has a positive PET scan, the clinician is obligated to prove it’s not cancer.”
Bousamra and his colleagues believe they have found a way to alleviate the physical and financial burdens of invasive testing on patients who have no life-threatening disease. A breath-analysis technique that uses specially coated silicon microchips to collect exhaled breath samples showed a lower false-positive rate than PET scans did in test cases.
According to a press release following Bousamara’s presentation of the study’s results at the American Association for Thoracic Surgery (ATTS) Annual Meeting on April 29, 2014, “Previous work had pinpointed four specific substances, known as carbonyl compounds, in breath samples as elevated cancer markers (ECMs) that distinguish patients with lung cancer from those with benign disease. The carbonyl compounds found in the breath are thought to reflect chemical reactions occurring in malignant lung tumors.”
Cautioning that the breath analyzer is not a screening test, Bousamara said that it is used in conjunction with a CT or CAT scan. “So, we have a CAT scan that shows a spot in the lung. Is it cancer or is it not cancer?”
According to Bousamara, in some cases, the doctor looks at a CT and knows it’s cancer. “No amount of testing would keep you out of that patient’s chest,” he said, explaining that there’s a subset of people where the diagnosis is truly uncertain, and that’s where his group thinks breath analysis can help. In fact, he said, it helps in two ways.
If a breath analysis is strongly positive (that is, if three of the four or all four cancer markers are positive), then it is very likely to be cancer. In that subset of patients, rather than do a CT guided biopsy or a bronchoscopy, the clinician could proceed directly to surgery and take out that nodule, said Bousamra, “because what we call the pre-test probability of it being cancer would be very high.”
Second, the study showed that breath analysis was always positive in big cancerous tumors that were 3 cm or larger. “So if you have a tumor that’s bigger than 3 cm, and the breath is negative, you can probably leave it alone or watch it,” Bousamra explained. He noted that a PET scan is invariably positive on such cases and that a thoracic surgeon can usually tell whether or not it’s likely to be benign. “If we had a negative breath analysis, I think clinicians could be encouraged to not make their next step an invasive biopsy procedure,” said Bousamra.
In early findings for 147 patients, breath analysis was much more accurate (75 percent) than PET (38.7 percent) at identifying those who did not have cancer.
However, at present, breath analysis remains a strictly investigational tool. “We are not making clinical decisions based on it,” Bousamra cautioned. “We are collecting the data and patients are interested in knowing what their breath analysis is. We do tell them, but we also inform them we’ve got more investigating to do.”
The paper Bousamra presented on those early findings at the AATS meeting will be published later this year in AATS’s Journal of Thoracic and Cardiovascular Surgery.
The next step will be for the researchers to accumulate two to three times more patients in order to confirm their original results. They also need to search for other disease processes that may produce the same cancer markers, yet may not actually be cancer. They had already begun to do that during initial testing, Bousamra said. “We’ve looked at patients with cystic fibrosis and pulmonary fibrosis and they did not produce the same carbonyl cancer markers. But, we need to do that more extensively so we don’t get fooled.”
The clinic is enrolling about 10 patients a week, so Bousamra expects they will be able to build out their data base over the next year. Then, if their findings hold up, they will begin the FDA review process.