Researchers from the University of Michigan and the University of California, San Francisco published a breakthrough research result in the journal Nature - the FastGlioma artificial intelligence model. This model can quickly judge the residual cancerous tumor in brain tumor surgery within 10 seconds, significantly improving surgical efficiency and accuracy, and is expected to completely change the process of neurosurgery. This innovation combines micro-optical imaging and AI basic models, using more than 11,000 surgical samples and 4 million microscope images for pre-training. Its high-resolution images are obtained by the self-developed stimulated Raman tissue imaging technology developed by the University of Michigan. . The development of this model will have a profound impact on the precise treatment of brain tumors.
American scientists recently published major research results in the journal Nature: FastGlioma, an artificial intelligence model jointly developed by the University of Michigan and the University of California, San Francisco, can quickly determine the residual cancer in brain tumor surgery within 10 seconds, and is a neurosurgery. Bring a revolutionary breakthrough.
This innovation combines micro-optical imaging with basic AI models. The research team used more than 11,000 surgical samples and 4 million microscopic images for pre-training, and used the stimulated Raman tissue imaging technology independently developed by the University of Michigan to obtain high-resolution images.
FastGlioma's outstanding advantages are reflected in its excellent detection capabilities. In practical applications, the model has a high-risk tumor residual missed rate of only 3.8%, which is far better than the 25% missed rate of traditional image and fluorescence-guided surgery. Even in "fast mode", its average accuracy rate can still reach 92%.
Research shows that FastGlioma can also reduce its dependence on traditional methods such as radiographic imaging, contrast enhancement or fluorescent labeling. This breakthrough technology not only helps surgeons make quick decisions during surgery, but also promotes application in other types of brain tumor diagnosis.
It is worth noting that complete resection of brain tumors has always been a major challenge facing neurosurgery, and some residual tumors are difficult to distinguish from healthy brain tissue. The emergence of FastGlioma provides new solutions to this clinical problem, marking another important step in artificial intelligence in the field of precision medicine.
The successful development of the FastGlioma model not only brought revolutionary changes to brain tumor surgery, but also set a new benchmark for the application of artificial intelligence in the medical field, indicating that precision medicine will be more efficient and accurate in the future. Its low omission rate and high accuracy rate will significantly improve patient outcomes and improve survival. In the future, we look forward to FastGlioma being applied in a wider range of clinical practice and benefiting more patients.