AI Tool Detects Cancerous Brain Tumor During Surgery in 10 Seconds

This happens because cancerous tissue remains after the initial surgery, and it continues to grow, sometimes at a faster rate than the original tumor. Residual tumors not only result in a lower quality of life and premature death for patients but also contribute to the burden on healthcare systems, which are projected to handle 45 million annual surgical procedures by 2030. Now, an artificial intelligence (AI)-based diagnostic system has been developed to detect cancerous tissue that might otherwise go unnoticed during brain tumor surgery. This technology allows neurosurgeons to remove the cancerous tissue while the patient is still under anesthesia or treat it afterward with targeted therapies.

In a new study, led by UC San Francisco (San Francisco, CA, USA) and University of Michigan (Ann Arbor, MI, USA), researchers demonstrated how an AI-powered diagnostic tool aids neurosurgeons in identifying hidden cancer that has spread nearby. This technique holds the potential to delay the recurrence of high-grade tumors and may even prevent recurrence in lower-grade tumors. The tool, called FastGlioma, is open-source and patented by UCSF, but it has not yet been approved by the Food and Drug Administration. FastGlioma combines AI’s predictive capabilities with stimulated Raman histology (SRH), an imaging technology that allows fresh tissue samples to be visualized at the bedside within one to two minutes. This rapid process bypasses the time-consuming procedures typically required in pathology labs for processing and interpreting tumor cells.

The AI system was trained using a dataset of over 11,000 tumor specimens and 4 million microscopic images, allowing it to accurately classify images and distinguish between tumor and healthy tissue. Neurosurgeons can receive diagnostic results within 10 seconds, enabling them to continue surgery if necessary. In the study published in Nature, neurosurgeons examined tumor samples from 220 patients with high-grade and low-grade diffuse gliomas, the most common type of adult brain tumor. The study found that 3.8% of patients who used FastGlioma had remaining high-risk tissue, compared to 24% of patients who did not use the tool. The study suggests that similar AI techniques could be tested in surgeries for other cancers, including breast, lung, prostate, and head and neck cancers.

“FastGlioma has the potential to change the field of neurosurgery by immediately improving comprehensive management of patients with glioma,” said senior author Todd Hollon, MD, of the Department of Neurosurgery at University of Michigan. “The technology works faster and more accurately than current standards of care methods for tumor detection and could be generalized to other pediatric and adult brain tumor diagnoses.”