According to a report in Science Daily, researchers have developed a device that uses artificial intelligence to mark any part of a cancerous brain tumor within 10 seconds after surgery.
The breakthrough technique, called FastGlioma, was presented in a research paper titled “Foundation models for fast, label-free detection of glioma infiltration” in the journal Nature, and is claimed to outperform traditional methods by a wide margin.
Fastglioma has an extraordinary average success rate of 92 percent in detecting and enumerating residual tumor in an operated patient. The new technique missed high-risk residual tumors only 3.8 percent of the time, compared to 25 percent with traditional methods.
According to the team of experts from the University of Michigan and the University of California, San Francisco, such a device could be a cutting-edge innovation for treating brain cancer patients.
how it works
Removing brain tumors has proven to be a difficult matter for neurosurgeons around the world. The brain is an extremely sensitive organ, and surgical removal of a tumor often leaves some cancerous mass behind. This can lead to cancer recurrence in patients, which often results in loss of life.
Traditional methods for detecting residual tumor tissue in neurosurgery use MRI imaging during the surgery procedure, combined with the use of fluorescent agents – a method that has resource limitations as well as specificity constraints. This is because it only works on certain types of tumors.
FastGlioma fills this gap by providing a quicker, more accessible and accurate solution to this serious problem. It is an artificial intelligence system referred to as a foundation model – similar to AI tools based on GPT-4 and DALL·E 3, trained on large datasets that can be used for a variety of tasks, from image classification to text generation. Can be tailored to applications.
The model has been pre-trained on over 11,000 surgical specimens and over 4 million unique microscopic fields for fastglioma. Imaging is performed via stimulated Raman histology – a high-resolution optical imaging technique developed at the University of Michigan.
“This technology works faster and more accurately than current standard of care methods to detect tumors and can be generalized to other pediatric and adult brain tumor diagnoses. “Can serve as a basic model for guidance.” said Todd Holden, MD, a neurosurgeon at the University of Michigan and co-author of the research paper.
There are two modes of fastglioma; One with full resolution images which takes about 100 seconds, and a faster mode which outputs lower resolution images with only 10 seconds output.
“This means we can detect tumor infiltration in seconds with extremely high accuracy, which can inform surgeons whether more resection is needed during an operation,” Hollon said.
The implications of this technology go far beyond brain tumors. Researchers suggest that FastGlioma could be adapted to other types of brain tumors, including pediatric cases such as medulloblastoma and ependymoma and non-glioma tumors such as meningiomas.
