This open-source AI model can detect almost 150 abdominal diseases
What's the story
Alibaba's research division, Damo Academy, has open-sourced an advanced artificial intelligence (AI) model. The system can detect nearly 150 abdominal diseases, including different types of cancer. It does this by analyzing computed tomography (CT) scans. This latest development is part of Alibaba's ongoing efforts to leverage medical AI technology for healthcare improvement.
Model features
Damo Radar is a vision-language model
The AI model, named Damo Radar, is a vision-language model.
It was developed to analyze contrast-enhanced CT scans of 18 abdominal organs and detect a wide range of diseases and abnormalities.
These include malignant tumors, among other things.
The research team at Damo Academy revealed this information on Friday as part of their ongoing work in the field of medical imaging technology.
Model training
Impressive diagnostic accuracy
Damo Radar was trained on a dataset of CT scans and clinical reports.
In nearly 40,000 real-world examinations, it achieved an impressive average area under the curve (AUC) of 0.913 across 146 clinical findings.
An AUC of 1.0 indicates perfect diagnostic accuracy.
The research team believes that this training method could be adapted for other types of medical imaging in the future.
Significance
Revolutionizing medical imaging
The research team has dubbed Damo Radar as "the world's first expert-level generalist medical imaging model."
This title highlights its potential to revolutionize the field of medical imaging.
By accurately detecting a wide range of diseases from CT scans, this AI model could significantly improve diagnostic processes and patient care in healthcare systems worldwide.