Researchers from the University of Michigan in the United States, working alongside colleagues from New York University and the University of California, San Francisco, have trained a neural network that can forecast mutations in brain cancer cells in under 90 seconds. This advancement has been reported in the scientific literature and discussed by experts in the field. The project centers on a neural model named DeepGlioma, which analyzes downloaded images of tumor samples obtained during surgical procedures to gather critical genetic information quickly.
The system processes pathology images to identify genetic alterations present in tumor cells with remarkable speed. According to the developers, DeepGlioma achieves a high level of accuracy in detecting malignant mutations, with reported performance surpassing 90 percent on typical test sets. The researchers emphasize that rapid detection is especially valuable for guiding intraoperative decisions and planning targeted therapies, potentially shortening the time between surgery and treatment while maintaining diagnostic reliability.
In related progress, the University of Gothenburg previously announced that artificial intelligence has been used to analyze data captured by microscopes, illustrating a broader push to apply AI to histology and pathology. These efforts collectively reflect a growing trend toward real-time or near-real-time analysis of tissue samples, enabling clinicians to interpret complex cellular features with the aid of advanced computational tools. The focus remains on improving speed, accuracy, and consistency in cancer diagnostics, while continuing to validate these approaches across diverse patient populations and clinical settings.