When it comes to diagnosing tumors, the human eye can only do so much. MRIs and CT scans have gone a long way toward enhancing the eye’s proficiency, and the microscope, which dates back to the 16th century, remains unsurpassed in helping clinicians discern details about surreptitious growths and other abnormalities. But now, a revolution in computer imaging is heralding a new era for treatment and diagnostics.

The key? Machine learning, a fast-growing branch of artificial intelligence. Scientists enter a wealth of tissue data into a computer. Over time, the computer “learns” to distinguish between sick and healthy tissue, detecting patterns invisible to the eye.

"With microscopes, we can’t always tell the difference between a localized tumor and one that’s likely to metastasize. The new image analysis approach does just that," says Arkadiusz Gertych, PhD, assistant professor of Surgery, and Pathology and Laboratory Medicine at Cedars-Sinai. The technology is especially promising for high-grade prostate cancer, and could help distinguish an aggressive tumor from a slow-growing one.

The imaging revolution is not limited to diagnostics. It’s also taking place in the operating room. A system known as the BrightMatter Guide lets neurosurgeons see tiny details inside the brain — minutiae that can’t be detected with the naked eye or an MRI — as they work to remove a tumor. Providing 3-D images in real time, the guide helps surgeons navigate the brain’s tiny neural connections to avoid damaging healthy tissue.

“The advantages can’t be overstated,” says Keith Black, MD, chair of the Cedars-Sinai Department of Neurosurgery, director of the Maxine Dunitz Neurosurgical Institute, and the Ruth and Lawrence Harvey Chair in Neuroscience. "Some 62,000 primary brain tumors and 150,000 metastatic brain tumors are diagnosed every year in the U.S. This new technology will make all the difference for many patients."