Researchers at Stanford University developed an AI model that identifies medical conditions from medical images. The system was tested on thousands of patient scans from multiple hospitals.
In trials, the AI detected cancers and other diseases earlier than radiologists in 85% of cases. The improvement was most significant for conditions with subtle early signs.
The technology uses deep learning algorithms trained on diverse medical datasets. It can recognize patterns too complex for human observation.
Clinical implementation could begin within two years pending regulatory approval. The system is designed to assist doctors rather than replace them.