Healthcare
Chinese medical AI models win national imaging contest
A Chinese hospital team won a national medical-AI imaging contest with an AI model for kidney-cancer CT diagnosis, while another team placed second with breast-cancer X-ray AI.
A Chinese medical AI team won first prize in the National Healthcare Security AI Image Recognition Contest with a CT diagnostic model for kidney cancer. The model was developed by the Second Affiliated Hospital of Guangxi University of Chinese Medicine using DIAGENS’s iMedImage cross-modal foundation model and iMedLoop development platform. A breast-cancer X-ray diagnostic model from the First Affiliated Hospital of Guangxi Medical University won second prize. The results indicate growing use of foundation models and imaging AI in disease-specific diagnosis. DIAGENS reported that it has delivered 158 model projects with 99 hospitals, including 65 Grade III Class A hospitals. Its projects cover 43 anatomical sites and 61 disease categories. The contest provides a showcase for medical-image recognition systems and their application to hospital-based diagnostic work. Cross-modal models are designed to process different types of medical information, potentially supporting more flexible clinical AI development. The kidney-cancer and breast-cancer models illustrate applications in both CT and X-ray imaging. The development matters because medical imaging remains one of the most active areas for AI deployment, while partnerships between technology providers and hospitals are becoming a common route for creating and implementing specialized models.
Original source: Deutsche Börse