In China, A.I. Finds Pancreatic Cancer That Doctors May Miss

In China, A.I. Finds Pancreatic Cancer That Doctors May Miss

https://www.nytimes.com/2026/01/02/world/asia/china-ai-cancer-pancreatic.html

Publish Date: 2026-01-02 00:00:00

Source Domain: www.nytimes.com

• Early Detection of Pancreatic Cancer: A retired bricklayer, Qiu Sijun, was diagnosed with pancreatic cancer after routine check-ups, due to early detection facilitated by an A.I. tool at the Affiliated People’s Hospital of Ningbo University.

• Development of PANDA Tool: The PANDA tool, developed by Alibaba researchers, uses artificial intelligence to detect pancreatic cancer in noncontrast CT scans, which traditionally produce less detailed images and are less prone to use for widespread screening.

• Effectiveness of the Tool: PANDA has helped in detecting approximately two dozen cases of pancreatic cancer, including 14 at an early stage. The tool identified 93 percent of pancreatic lesions in tests conducted with over 20,000 noncontrast CTs.

• Challenges and Potential Risks: The tool has some limitations, such as high false-positive rates, which could lead to unnecessary follow-up tests and distress. There are concerns about the feasibility and reliability of noncontrast CTs in routine screenings.

• Regulatory Approval and Clinical Trials: PANDA received expedited review status from the U.S. FDA and is undergoing clinical trials in China, including in rural Yunnan Province.

• Skepticism and Further Research: Some experts remain skeptical of the tool’s effectiveness, especially for experienced specialists. Further real-world data is needed to gauge its benefits against the risks of false positives and unnecessary testing.

• Operational Challenges: The Ningbo hospital faces difficulties associated with the scale of data processing and follow-up due to limited staff, reflecting broader issues in integrating A.I. tools in healthcare settings.

• Patient Experience: Qiu Sijun expressed relief after successful removal of his tumor, highlighting the potential benefits of early detection despite the complexities and public trust issues in widespread medical applications.