Alibaba’s AI Tool for Pancreatic Cancer Detection Gains FDA Breakthrough Status
- tech360.tv
- 2 hours ago
- 2 min read
Alibaba Group Holding’s research division, Damo Academy, has received a breakthrough device designation from the US Food and Drug Administration for its AI-powered pancreatic cancer detection model, Damo Panda.
The designation allows for a faster review and approval process, marking a significant step for the Chinese tech giant’s expansion into global healthcare.
Damo Panda, first introduced in Nov. 2023 in the journal Nature Medicine, is designed to detect early-stage pancreatic cancer in asymptomatic patients using deep learning.
The model was trained on abdominal non-contrast CT scans from 3,208 pancreatic cancer patients. According to Damo researchers, it demonstrated 34.1% higher sensitivity than radiologists in identifying the disease.
Alibaba has already trialled the tool in China, screening 40,000 people at a hospital in Ningbo. The model identified six early-stage cases, including two that routine exams had missed.
Damo Academy, founded in 2017, focuses on AI and RISC-V chip architecture. It has developed a series of processors under the XuanTie brand, including a server-grade CPU announced in Feb.
Alibaba plans to promote Damo Panda both in China and internationally through partnerships with companies such as Ankon Medical Technologies.
AI has increasingly influenced medical imaging, with studies showing its ability to match or exceed human experts in image analysis and enhancement.
Other Chinese tech firms are also investing in AI healthcare tools. In Feb., Huawei Technologies launched a large AI pathology model with Shanghai’s Ruijin Hospital to assist in cancer diagnostics.
Huawei has also formed a new team dedicated to the medical and healthcare sector, according to Chinese media reports in March.
Alibaba’s Damo Panda AI model receives FDA breakthrough device designation
The tool detects early-stage pancreatic cancer with 34.1% higher sensitivity than radiologists
Trials in China screened 40,000 people and identified six early-stage cases
Source: SCMP