Open-Source AI Model for Cancer Detection

Alibaba's Damo Academy released an open-source medical AI model that analyzes CT scans to detect cancer and 146 clinical conditions.

Alibaba's Damo Academy released an open-source medical AI model that analyzes contrast-enhanced CT scans of 18 abdominal organs. The model was validated across 40,000 real-world CT scans and achieved an average area under the curve (AUC) of 0.913 across 146 clinical findings. It outperformed most radiologists per a Science journal study and is described as the world's first expert-level generalist medical imaging model. This work enables low-resource regions to detect cancers and nearly 150 conditions without requiring specialized radiologists to interpret scans—a critical reduction in barriers for areas lacking imaging expertise. The model’s open-source nature allows free deployment, though it requires existing CT scans to function.

This advancement moves cancer detection toward greater accessibility in resource-limited settings by shifting the burden from expensive human interpretation to scalable AI analysis. For regions where CT scanners are scarce but basic imaging exists, the model reduces the need for radiologists while maintaining clinical accuracy. The next step is testing the model’s performance on non-abdominal scans and real-world clinical workflows outside controlled validation environments.

Source: SCMP Tech