Optical Technology Detects Cancer With High Accuracy During Surgery

An intraoperative, multimodal optical cancer detection system can detect brain cancer in real time during surgery, with high degrees of accuracy, sensitivity, and specificity, according to a new study.

“With brain cancer, near-perfect detection is important so that we can remove as much cancer as possible, without removing healthy tissue,” said Kevin Petrecca, MD, PhD, associate professor of neurology and neurosurgery at McGill University in Montreal. “Residual cancer postsurgery is associated with decreased time to recurrence and lower survival.”

The researchers published their results today in Cancer Research.

There is a need for the development of highly sensitive and specific cancer detection instruments that can be seamlessly integrated into clinical practice. The new technology holds promise to impact a wide range of surgical and noninvasive interventional oncology procedures by improving cancer detection capabilities, said Petrecca.

Raman spectroscopy alone can achieve 90% detection accuracy. Petrecca and colleagues developed an integrated cancer detection system that combines Raman spectroscopy with intrinsic fluorescence spectroscopy and diffuse reflectance spectroscopy.

The optical detection system consists of a handheld probe coupled with an analytics platform on a mobile cart. The probe contains miniaturized spectroscopy technology, and has a tip the same size surgeons use to remove brain cancers. “The probe can scan a 500 micron-diameter area of tissue. Since a cell is roughly 10 microns across, the tool is able to detect down to small numbers of cancer cells,” said Petrecca.

The researchers developed the platform’s machine-learning algorithm using archived tissue samples of brain cancer and normal cells from brain surgeries classified by standard pathology.

In a validation study, they investigated the use of the optical system among 15 patients with grade II–IV gliomas and metastatic brain tumors from primary lung cancer, colon cancer, or melanoma who were undergoing open cranium surgery. They interrogated 10 to 15 sites in each patient, 161 sites in total, taking optical readings of normal and tumor tissue regions at each site for blinded post-analysis. The goal was to determine if cancer cells were present, and then to compare the data to corresponding spectral data.

Image analysis demonstrated improvements, measured as area under the curve. The technology demonstrated 97% accuracy, 100% sensitivity, and 93% specificity, which approaches the threshold for complete resection, the researchers stated. They were able to detect cancer with virtually the same degree of accuracy, sensitivity, and specificity among all cancer types investigated.

“Our findings are novel since optical techniques are not standard in any surgeries at present,” said Petrecca. “The results also indicate a strong potential for this technology to be adapted to a wide range of surgical and detection applications, including laparoscopic and robotic surgeries, and colonoscopy,” and thereby increase the diagnostic accuracy of these procedures and treatments.

In the future, the technology will be tested in clinical trials designed across a larger patient population and for a variety of cancer types, looking at residual cancer following surgical intervention, diagnostic accuracy, progression-free survival, and overall survival.