NEW YORK – With the launch of a multi-center clinical trial, Canadian biotech firm Breathe BioMedical said that it is developing a breath-based test to help identify breast cancer in patients whose mammography results are difficult to interpret due to dense breast tissue.
The Moncton, New Brunswick-based firm announced late last month it was partnering with researchers at George Washington University Breast Center along with two investigators from the Mayo Clinic on an observational study with up to 1,000 patients that will use infrared spectroscopy and machine learning to identify breath-based volatile organic compounds (VOCs) that are biomarkers of breast cancer. The research will help the company to refine its test as an adjunct to mammography for the detection of early-stage breast cancer in patients with dense breast tissue.
Breathe BioMedical CEO Bill Dawes said early study results show that the company's test would be equally useful in patients with high- or low-density breast tissue, but the firm has focused on the use of the test in women with dense breast tissue because of a need in the market for better methods to stratify those patients by breast cancer risk.
About half of US women over the age of 40 have dense breast tissue. The US Food and Drug Administration began requiring in September that mammography results include notification if a patient's breasts are classified as dense, including a statement that dense tissue makes cancers more difficult to detect on a mammogram and other imaging-based tests may aid the discovery of cancers.
Patients with dense breast tissue also have an increased risk of breast cancer, although research has not found an increased risk of breast cancer death in patients with dense breasts, according to the National Cancer Institute.
In an article published this week in JAMA, a pair of authors from the University of Washington and University of California Los Angeles noted that prior studies have raised concerns about the potential for discordant assessments of dense breast tissue in mammograms. However, digital mammography and digital breast tomosynthesis perform well in detecting breast cancer in women with dense breasts, and prior research has shown that digital breast tomosynthesis could in some patients reduce false-positive results and improve breast cancer detection in comparison with digital mammography.
Breathe BioMedical's breath testing platform includes its tabletop SohnoXB breath sampler that could be used at the point-of-care to collect samples onto sorbent tubes. Dawes said that collecting the breath sample requires about the same amount of effort as a person would use to cool a spoonful of soup.
The sorbent tubes are heated to specified temperatures and transferred into the company's proprietary infrared cavity ring-down spectrometer, which is used to generate a breath spectrograph based on laser light absorption of analytes in the sample, including VOCs. An algorithm is used to identify biomarker-based patterns of breast cancer.
The results of Breathe BioMedical's test are used to help determine which patients would proceed to downstream imaging such as ultrasound or MRI. The authors of the JAMA article wrote that decisions whether to perform supplemental screening with MRI or ultrasonography should be based on comprehensive risk assessment rather than breast density alone.
Meanwhile, some other research teams have reported success in identifying breath-based VOCs that could be used for the detection of breast cancer.
In 2020, researchers from China wrote in the Journal of Clinical Laboratory Analysis that they had identified seven breast cancer-specific VOCs that differed between breast cancer patients and controls as well as between breast and gastric cancer patients. Two years later, scientists in Japan wrote in the Journal of Breath Research that they had found that they could discriminate between breath samples from patients with and without breast cancer through differences in concentrations of the three VOCs 3,7-dimethyl-2,6-octadien-1-ol, ethanolamine, and ethyl nonanoate.
Another research group from Mexico also reported in 2020 in the Journal of Breath Research that an "electronic nose" with polymer-based sensors to detect changes in electrical resistance could be used to identify overall patterns of VOCs that are distinctive for patients with breast cancer.
As for the broader breath testing space, a handful of firms has been developing breath-based tests for cancer, other chronic diseases, and infectious diseases. Cambridge, UK-based Owlstone Medical, for example, has been developing a menu of tests that use VOCs, starting with the early detection of liver disease followed by a test for early detection of lung cancer. However, the firm is developing its breath-based tests for endogenous markers of disease as well as signals that are produced following the consumption or inhalation of VOC probes.
The FDA's Center for Devices and Radiological Health also announced last year it was partnering with the Bill and Melinda Gates Foundation on research into analytical methods that could be used to advance breath-based diagnostic tests, with a goal of bringing efficient and affordable testing options into rural and remote areas.
Dawes said that the company previously studied the use of its technology for non-small cell lung cancer and COVID-19–related pneumonia but it has been focused on breast cancer for the past 18 months. He said that it is much easier to run a clinical trial on the use of the technology for the detection of breast cancer in part because many more patients undergo breast cancer screening compared to NSCLC.
Breathe BioMedical reported in 2023 at the American Society of Clinical Oncology annual meeting that it found in a proof-of-concept study that an early version of its breast cancer test performed with upward of 88.2 percent accuracy, with 92.9 percent sensitivity and 80 percent specificity. He said that the number of VOCs that were used in that analysis is proprietary.
Dawes said that the company collects upward of 800 data points from each patient sample, although he expects that the amount of data will be reduced with refinement of the company's testing model and the identification of a small number of VOCs that form the signal for breast cancer. In the observational study, Dawes said that the company will use deep learning to develop a model, and he thinks that the breast cancer signal could consist of five, 10, or 20 VOCs.
While the firm intends to use machine learning to identify markers of disease, Dawes said the results will not include details on the individual VOCs that are identified with the test.
"In our go-to-market efforts, it's not our intent to unpack the sample and say 'This is what's there, this is why it's there, and this is how it correlates to underlying disease,'" he said. "Our intent is to say 'Our data and our machine learning algorithms have identified the presence of disease.'"
He said that the test is simple, noninvasive, and inexpensive, and he expects that the company will be able to secure reimbursement for the test. He added that the test also could improve testing access for women who are recommended for imaging-based follow-up to mammography but have travel, scheduling, or payment issues, among other barriers.
"We believe that the payors will pay for it," Dawes said. "We believe that the institutions will find capacity for it, and we believe that women will find their way to the tests because they have a cancer signal."
Dawes said the test's cost is expected to be equivalent to the cost of mammography.
The company had filed in June 2024 a preliminary prospectus for an initial public offering with the US Securities and Exchange Commission, but the company said in October it had decided against pursuing the IPO and filed for withdrawal of its registration statement. The firm said that it decided against the IPO due to market conditions and a determination that registering its securities would not be in its best interests.
In its prospectus, the company said that it planned to complete in 2025 verification and validation studies for components of its breath cancer breath test. The firm also said it planned for a pivotal trial in 2026 to support an application in 2027 for US Food and Drug Administration pre-market approval. If the test gains acceptance in the US and Canada, the firm plans to develop a network of labs to analyze breath samples.
The firm noted in the SEC filing that it had 13 patents in the US and three in Canada as well as 64 pending patent applications in the US, Canada, EU, Australia, China, India, and Japan.
The company noted in the SEC filing that it had developed its SohnoXB breath sampler and its spectrometer with funding support from the Atlantic Canadian Opportunities Agency and the National Research Council's Industrial Research Assistance Program. The collection device is registered with the FDA and Health Canada.
Because the company is training its model to identify VOCs that are not specific to dense breast tissue, Dawes thinks that the test eventually could be used as a first-line screening test. He said that the firm is almost exclusively focused on breast cancer testing, although he said that the firm anticipates the use of the technology next to develop of tests for other women's health-related cancers such as ovarian and cervical cancers, with other cancers, viral diseases, and chronic disease tests to follow.
As for the breast cancer test, Dawes said that Breathe BioMedical plans to continue partnering with academic institutions on clinical studies and publishing study results to gain traction in the medical community. He said that the firm has been working for more than a year on its market plans with a focus on securing reimbursement for the test.
Dawes said that the use of the breast cancer test could reduce healthcare spending, and he thinks that the self-insured corporate market will find value for the test through its identification of cancers in early stages when patients have a higher likelihood of survival and the cost of treatment is lower. The current clinical trial will provide data to support those discussions, he said.
"The economics work for this test, which is why we're going to be focusing on the payors as part of our market launch," he said.