Alibaba open-sources AI model detecting cancer, 150 conditions

Damo Academy, the research arm of Chinese tech giant Alibaba, has released an open-source vision-language model capable of identifying nearly 150 abdominal conditions — including malignant tumors — from computed tomography scans, the institute said Friday according to the South China Morning Post. The system, named Damo Radar, analyzes contrast-enhanced CT scans covering 18 abdominal organs to identify a broad range of diseases and abnormalities.

The model achieved an average area under the curve score of 0.913 across 146 clinical findings in nearly 40,000 real-world examinations, approaching the maximum score of 1.0. The research team described it as "the world's first expert-level generalist medical imaging model," adding that it could eventually be expanded to other types of medical imaging.

Outperforms radiologists in clinical trials

In a study involving 26 radiologists from multiple hospitals, Damo Radar outperformed 23 participants on average, the developers noted. With the model's assistance, radiologists also reduced missed diagnoses by 10% and cut diagnosis time by more than 30%, according to the findings.

Clinical deployment and efficiency gains

Alibaba's Damo Academy developed the tool to support clinical workflows and improve diagnostic accuracy across diverse medical settings. The open-source release allows integration into existing radiology departments to assist physicians interpreting complex abdominal scans.