On November 14, 2024, a significant advancement in brain tumor surgery was reported by researchers from UC San Francisco and the University of Michigan. An artificial intelligence (AI)-based diagnostic system, named FastGlioma, has shown promise in improving survival rates for patients undergoing surgery for brain tumors.
This innovative tool assists neurosurgeons in identifying cancerous tissue that may not be visible during surgery, allowing for more comprehensive removal of tumors while patients are still under anesthesia. The study highlights that when brain tumors recur, survival rates diminish significantly, particularly for high-grade tumors, which often lead to patient mortality within a year.
The research, published in the journal Nature, indicates that the application of FastGlioma resulted in only 3.8% of patients having remaining high-risk tissue post-surgery, compared to 24% in those who did not use the tool. This advancement could delay or even prevent tumor recurrence, offering new hope for patients with various types of cancers.
FastGlioma combines AI technology with stimulated Raman histology (SRH), enabling rapid visualization of fresh tissue samples within minutes. The system has been trained on a dataset of over 11,000 tumor specimens, allowing for precise classification of tissue types. The tool is currently open source and patented by UCSF, although it has yet to receive approval from the Food and Drug Administration.
With the potential for broader applications, similar AI techniques are set to be tested in surgeries for breast, lung, prostate, and head and neck cancers, marking a pivotal shift in the field of neurosurgery.