Alibaba’s research division, Damo Academy, has open-sourced an advanced AI model named Damo Radar that can identify nearly 150 abdominal health conditions, including various cancers, by analyzing CT scans. This breakthrough marks a significant advance in AI-assisted medical diagnostics in China.
- Damo Radar identifies 146 abdominal conditions with expert-level accuracy
- Model outperforms many radiologists in clinical studies
- Open-sourced to encourage broader medical AI development
What happened
Alibaba Group’s research arm, Damo Academy, publicly released an AI model called Damo Radar that can analyze contrast-enhanced computed tomography (CT) scans to detect nearly 150 abdominal diseases, including various forms of cancer. The model was trained using a large dataset of paired CT scans and clinical reports and was tested across almost 40,000 real-world cases in collaboration with multiple hospitals and research institutions.
In clinical comparisons involving 26 radiologists, Damo Radar's diagnostic accuracy exceeded that of the majority of participants, scoring an area under the curve (AUC) of 0.913 across 146 clinical findings. The tool also helped radiologists reduce missed diagnoses by 10% and cut the time needed for analysis by over 30%.
Why it matters
This release represents a notable step forward in AI-assisted healthcare, demonstrating the potential for AI to support clinical decision-making by improving accuracy and efficiency in disease detection. By open-sourcing the technology, Alibaba is encouraging wider adoption and development of AI models that could impact diagnostic practices globally.
Given China’s growing focus on integrating technology with healthcare, Damo Radar aligns with national priorities to enhance medical services using AI. It builds on prior efforts by Alibaba's Damo Academy aimed at elevating early cancer detection and tackling other complex medical conditions using advanced machine learning models.
What to watch next
The medical community will likely monitor further clinical deployments of Damo Radar to evaluate real-world impact on healthcare outcomes and workflow improvements. Expansion into other imaging modalities beyond CT scans, as suggested by developers, could broaden the AI’s application in diagnostic radiology.
Additionally, interest will grow in collaborations between AI firms and healthcare providers within China and internationally, leveraging open-source models like Damo Radar. Regulatory acceptance, integration into clinical systems, and ethical deployment will be key factors shaping the next phase of medical AI innovation.