Scientists from Germany have developed an artificial intelligence system capable of identifying the type of brain tumor within minutes, by analyzing microscopic images of brain tissue.
The scientists pointed out that current methods for diagnosing brain tumors rely on complex molecular analyses that are considered the "gold standard" for determining the type of disease, but they require special, expensive equipment and may take about two weeks, while the (Hetairos) algorithms they developed can diagnose the disease quickly using available images of tissue samples.
The system was trained using over 11,000 tumor images from 9,606 patients across 11 medical centers worldwide and can distinguish between 102 molecular patterns of central nervous system tumors. The system's developers also indicated that its tumor diagnosis accuracy ranges between 87% and 88%, achieving a level of accuracy comparable to that of experienced healthcare professionals.
During its tests, the (Hetairos) system was able to provide a diagnostic result in approximately 12 minutes after the image was uploaded, while the traditional molecular diagnosis of this type of tumor takes about 12 days.
Researchers believe this technology could be particularly useful for hospitals that do not have access to complex genetic testing. They emphasize that the system is not a replacement for current diagnostic methods, but rather a tool for obtaining rapid information about the tumor and making faster decisions regarding tests and treatment.
