Moncton-based Picomole and Emergence client has developed a first-of-its-kind screening tool that makes lung cancer detection as simple as breathing into a tube.
“The fundamental part of our objective is to develop a screening tool for cancer and other diseases and just have, at the patients end of it, a simple breath test,” said Picomole CEO Stephen Graham, in an interview with Huddle.
“Then, when we talk about 10 years from now, you’d be able to go into a pharmacy or your doctor’s office or healthcare clinic and give a breath sample and you’d be able to get screened for a number of different diseases with one breath test. That’s our big goal.”
Picomole’s new technology has three components. First is the breath sampler, which is the part the patient sees. It’s about the size of a microwave and collects the breath samples in slim, stainless steel canisters.
The second component is a spectrometer that processes and measures the amount of light absorbed by organic compounds found in the breath samples. This provides unique digital breath fingerprints and hundreds of biomarkers are provided for each sample collected.
The third component is the machine learning software that analyzes the digital output from the spectrometer to identify the presence of disease. Under the supervision of University of New Brunswick professor Erik Scheme at the Institute of Biomedical Engineering and with funding from Mitacs, MSc student Robyn Larracy has been working with the company to identify patterns in the breath data that correspond to diseases and then training computer algorithms to recognize them. Graham says Larracy’s involvement in the project has been crucial.
“We were able to get funding through Mitacs to bring Robyn in. You read everywhere that AI and machine learning is exploding in every different corner of industry,” he said. “These machine learning people, they don’t grow on trees. They’re very specialized and are in high demand. So our association with UNB has been great.”
So far, the technology has proven to identify lung cancer patterns with an 85 percent accuracy rate. Looking ahead, Picomole plans to expand its breath-based screening tool so it can detect other diseases, including breast cancer and Covid-19.