A device the size of a microwave oven could soon be able to detect multiple health problems in seconds, from COVID-19 to cancer to concussions, a Polytechnique Montréal researcher believes.

Best of all, samples don’t have to be shipped to a lab, which can cause delays in getting results and are sometimes treated with reagents that can be harmful to the environment and difficult to obtain in times of high demand.

This means that the analysis could be performed by virtually anyone, virtually anywhere, and the results would be available very quickly.

“We could imagine, for example, a hockey team: before going on the ice, everyone could take the test relatively quickly," said Professor Frédéric Leblond, who shared his findings with The Canadian Press. “ In half an hour, it’s settled, everyone passed, everyone got their result.”

In a nutshell, Raman spectrography -- named after the Indian physicist who won the Nobel Prize in Physics in 1930 for the discovery of the ‘Raman effect’ -- uses a laser beam to determine the composition of a sample, whether it’s saliva, blood or another source of cells.

In the case of COVID-19, Leblond and his colleagues at the CHUM research centre found that a single drop of dried saliva would be enough for Raman spectroscopy and machine learning to determine, in just a few moments, whether the patient is infected with SARS-CoV-2.

The researchers, led by postdoctoral fellow Katherine Ember, analyzed 37 saliva samples from patients with COVID-19, in addition to 513 samples from healthy patients, to train an automated learning tool to discern samples from infected or healthy individuals.

The resulting images are useless to human eyes, but they contain a signature of COVID-19 that the artificial intelligence is able to detect -- although the researchers admit they don’t know what the tool detects, exactly.

Regardless, the tool successfully identifies positive cases with a success rate ranging from 79% to 84% and negative samples with a rate of 64% to 75%, depending on whether the sample came from a male or female.

Improvements made since the work was completed and reported in the Journal of Biomedical Optics have increased the detection of positive cases to 95% and negative cases to 80%.

Professor Leblond admits, however, that his discovery may come a little late in the fight against COVID-19, at a time when the pandemic may be running out of steam and other methods of detecting the virus are widely available.

“But what’s interesting is that our method is not specific to COVID," he added. “We can repurpose this to detect other things like influenza, we could imagine detecting Lyme disease, we can think about early cancer detection tests.”

“It’s like taking a very difficult example to start with,’ Professor Leblond said. It really opens the door to many things.”

Leblond won the Quebec Science Discovery of the Year award in 2017 with his surgeon collaborator, Kevin Petrecca, for his use of Raman spectroscopy as a tool to distinguish between cancerous and healthy tissue during surgery.

— This report was first published in French by The Canadian Press on Feb. 9, 2022.