When a brain tumor is removed via surgery, not everything can necessarily be removed. What’s left has a name: residual tumor. And this is explained by the fact that we do not always see it during surgery, especially in the case of diffuse glioma. But a new technique could help define what’s left, very quickly, as revealed a study published in the journal Nature.
To study what could not be removed during brain surgery, two techniques are available: a fluorescent agent that highlights tumor tissue, or an MRI. But the first is not applicable to all tumors, and the second is not always available.
So researchers looked at the cases of 220 patients suffering from high- and low-grade diffuse gliomas. They used artificial intelligence called FastGlioma, a diagnostic system that was supposed to calculate how much tumor remained in the patients’ brains. Its accuracy is 92%. Conventional methods for determining whether there is tumor remaining in the tissue have a margin of error of 25%. There, the margin of error increased to 3.8%.
Extremely fast detection time
The AI used was trained through 11,000 surgical samples and 4 million unique microscopic fields of view. In 10 seconds, we can have an image of the tumor residue in the brain. To have a better quality image, the time is estimated at 100 seconds.
One of the study’s authors, Todd Hollo, a neurosurgeon at the University of Michigan Health explained what this meant: “we can detect tumor infiltration within seconds with extremely high accuracy, which could inform surgeons if additional resection is needed during an operation“.
Source: www.topsante.com
