Neurosurgeons need all the help they can get to remove brain cancer tumors. If they leave cancer cells behind, the tumors can regrow. Finding cancer cells can be particularly difficult with infiltrative cancers such as gliomas, which invade surrounding brain tissue.
Raman spectroscopy could help neurosurgeons find those errant cells. A team led by engineer Frédéric Leblond of Montreal Polytechnique and neurosurgeon Kevin Petrecca of McGill University, also in Montreal, has developed a Raman probe that distinguishes between normal and cancer cells. They showed that their method could find previously undetectable cancer cells in the brains of glioma patients (Sci. Transl. Med. 2015, DOI: 10.1126/scitranslmed.aaa2384).
“It’s very uncomfortable when you’re performing an operation and are not certain if you are removing all the cancer,” because missing some can impact a patient’s survival, Petrecca says. The Raman probe, he says, allows surgeons to spot cancer cells they might have thought were normal.
To use the tool, a surgeon simply holds a fiber-optic probe in contact with the brain tissue to collect a Raman spectrum. The researchers use an algorithm that statistically analyzes the data to differentiate between healthy and cancerous cells.
“Our algorithm doesn’t use just one peak,” Petrecca says. “It covers the entire spectrum using a machine-learning algorithm to classify different tissue types.” What they know at this point is that spectra of cancer cells have more intense peaks in certain spectral regions associated with lipids, nucleic acids, and proteins.
During surgeries on 17 glioma patients, the researchers used the Raman probe in conjunction with magnetic resonance imaging, a conventional method surgeons use to find the borders of brain tumors. “We show that we can detect cancer cells at least 1.5 cm beyond the abnormal signals detected by MRI, for both low- and high-grade gliomas,” Petrecca says.
Other researchers have tested similar Raman methods in animals and biopsy samples from patients, but these are some of the first data collected during surgery.
The work “represents an important step in bringing Raman spectroscopy into medical practice,” says Ji-Xin Cheng, a biomedical engineer at Purdue University who is also developing medical applications of spectroscopy. Unlike other medical imaging techniques, Raman provides information on molecules, which could allow for faster characterization of tissue types.
Henry Brem, a neurosurgeon at Johns Hopkins University, says the work “represents an important new tool in neuro-oncology that can lead to safer and more effective neurosurgery.”
Petrecca says the next step is to run clinical trials to demonstrate that the Raman technique can improve surgery outcomes. He and his collaborators plan to start such a trial soon.