Carrefour de l'actualité

More accurate prostate cancer diagnoses with Raman spectroscopy

August 18, 2020 - Source : NEWS

Diagnosing the different subtypes and grades of prostate cancer could become more accurate with the use of Raman microspectroscopy, an approach developed by a team from Polytechnique Montréal and the Centre de recherche du Centre hospitalier de l’Université de Montréal (Université de Montréal hospital research centre, or CRCHUM). The results of their work conducted on human tissues were published on Friday, August 14, 2020, in the prestigious web journal PLOS Medicine.

Prostate cancer is the most common cancer directly affecting men in Canada, accounting for 20% of all new cancer cases identified in this group. The Canadian Cancer Society says some 23,000 men will be diagnosed with prostate cancer this year and approximately 4,200 will die from the disease during this same period.

As with other cancers, prostate cancer comes in different forms, so-called “subtypes” which also advance through several histological grades. Identifying each of these subtypes is a daily challenge for pathologists. Occasionally misdiagnosis occurs and as a result, patients do not have access to the optimal treatment for their type of cancer.

Frédéric Leblond et Dre Dominique Trudel. (Photo : CHUM)

Frédéric Leblond, professor of engineering physics at Polytechnique Montréal, and Dr. Dominique Trudel, pathologist at the CHUM. Both are also researchers at the CRCHUM. (Photo: CHUM)

Raman spectroscopy for support

This is why Frédéric Leblond, full professor in the Department of Engineering Physics at Polytechnique Montréal, his CRCHUM colleague Dr. Dominique Trudel and their postdoctoral fellow Andrée-Anne Grosset have been working to find out how Raman spectroscopy could help the day-to-day work of pathologists who examine prostate biopsy samples under the microscope.

The team has already obtained preliminary results suggesting that Raman spectroscopy can distinguish cancerous tissue from healthy tissue in lung and ovarian cancers. Recently, it has also begun developing an intraoperative method based on Raman spectroscopy to guide conservative surgery for breast cancer. Professor Leblond and Dr. Kevin Petrecca, affiliated with McGill University, had done the same in the past for brain cancer. That work earned them Québec Science magazine's Discovery of the Year Award for 2017.

As part of the study published in PLOS Medicine, researchers microscopically analyzed samples collected between 1993 and 2013 from 483 men with different forms of prostate cancer.

“We first produced a series of Raman spectra with a portion of these samples, and then compared the results with those obtained by a method traditionally used to identify cancer-related biomarkers,” says Dr. Leblond.

He continues: “A machine-learning algorithm was then trained with all these results to discriminate between cancerous and healthy tissue, as well as a subtype of cancer, intraductal carcinoma of the prostate or IDC-P, an aggressive form of cancer that must be taken into account in the patient’s treatment plan.”

Then it was time to check whether the algorithm had learned from experience. The team presented it with a series of prostate cancer samples it had never seen before. Result: the algorithm's analysis of Raman spectra differentiated cancerous from healthy tissue in about 87 per cent of cases, nearly 9 times out of 10. It also differentiated IDC-P cancers from other prostate cancers with an 85% to 95% accuracy, depending on the sample group.

Raman microspectroscopy has the potential to become one of the tools available to pathologists to identify prostate cancer type and grade.

These results are very promising, according to Professor Leblond. “It's a good proof of concept,” he notes, adding that tools based on Raman microspectroscopy may be of potential interest to hospitals, as this approach is both cheaper and faster than immunohistochemistry, the method most often used in the laboratory to identify cancer subtypes.

More work needs to be done before this can happen, however, as he says the current approach is not yet ready for clinical deployment. “In this study, we were taking several measurements at a specific point in an image to get a result. We’re now developing a more practical method that will enable full imaging of the samples.”

Such an approach, he says, will allow for more accurate stratification of patients and the delivery of treatment tailored to their condition. The team he shares with Dr. Trudel at the CRCHUM is already working on it.

With financial support from the CRCHUM, IVADO, TransMedTech Institute, Mitacs, Institut du cancer de Montréal, Prostate Cancer Canada, Fonds de Recherche du Québec – Santé, Canada First Research Excellence Fund, Ontario Institute for Cancer Research, National Cancer Institute (National Health Institutes) and the Natural Sciences and Engineering Research Council of Canada (NSERC).

Learn more

Article published in PLOS Medicine journal
Professor Frédéric Leblond’s expertise
Department of Engineering Physics website
Centre hospitalier de l’Université de Montréal (CHUM) website (In French)
Université de Montréal hospital research centre (CRCHUM) website
TransMedTech Institute website
Institute for Data Valorization (IVADO) website
Institut du cancer de Montréal website
Mitacs website

Suggested Reading

February 9, 2022

COVID-19: A new detection strategy developed at Polytechnique Montréal

November 11, 2020

Detecting the SARS-CoV-2 virus in saliva: A clinical study starts