A routine medical checkup turned life-saving for a retired bricklayer in China, thanks to an artificial intelligence tool now being tested in hospitals. This development highlights a global race, with implications for countries like India, to apply AI to some of medicine's most challenging problems.
The Life-Saving Call After a Routine Checkup
Qiu Sijun, a 57-year-old from eastern China, went for a standard diabetes checkup. Three days later, he received an unexpected call from Dr. Zhu Kelei, the head of the pancreatic department at the Affiliated People's Hospital of Ningbo University. The doctor asked him to return for a follow-up. "I knew it couldn't be anything good," Qiu recalled. His fears were partially confirmed: he had pancreatic cancer. However, there was a silver lining. The tumour was detected at an early stage, allowing Dr. Zhu to successfully remove it.
This early detection was made possible not by human eyes alone, but by a new AI-powered tool the hospital was trialling. The system, named PANDA (pancreatic cancer detection with artificial intelligence), analysed Qiu's routine noncontrast CT scan and flagged the abnormality before any symptoms appeared. Pancreatic cancer is notoriously deadly, with a five-year survival rate of only about 10%, primarily because it is so difficult to catch early.
How PANDA Works and Its Early Success
Conventional methods for confirming pancreatic cancer, like contrast CT scans, involve high radiation, making them unsuitable for widespread screening. Lower-radiation noncontrast CTs are safer but produce less clear images, often causing radiologists to miss early signs. The PANDA tool, developed by researchers affiliated with Chinese tech giant Alibaba, was specifically trained to identify pancreatic cancer in these noncontrast CT scans.
Since its clinical trial began at Dr. Zhu's hospital in November 2024, PANDA has analysed over 180,000 abdominal or chest CT scans. It has assisted doctors in detecting approximately two dozen pancreatic cancer cases, with 14 of those being in the early stage. Significantly, all these patients had come with general complaints like bloating and had not initially consulted a pancreatic specialist. "I think you can 100% say AI saved their lives," stated Dr. Zhu, noting that several scans had not raised alarms until the AI flagged them.
Potential and Caution from the Global Medical Community
The tool has gained significant recognition. In April, the U.S. Food and Drug Administration (USFDA) granted PANDA "breakthrough device" status, expediting its review for the market. It is also undergoing several clinical trials in China. However, researchers urge caution, emphasizing that more real-world data is needed to prove the tool's ability to detect enough early cases to justify the risks of false positives and unnecessary follow-up tests.
International experts have expressed both interest and skepticism. Dr. Ajit Goenka, a radiologist at the Mayo Clinic, warned about the psychological and financial toll of false alarms. He noted that hundreds in Ningbo might have faced the terror of a potential cancer diagnosis and endured invasive testing only to find they were healthy. Dr. Diane Simeone, a pancreatic surgeon at UC San Diego, suggested the tool might be more useful for training junior doctors than for seasoned specialists but acknowledged its value as a safety net in hospitals with a shortage of experts.
Dr. Zhu's team maintains a critical review process, manually checking all scans the AI marks as high-risk. He admits the model is not yet a replacement for a specialist, as it can sometimes highlight pancreatitis and cannot determine a tumour's origin. The future of such AI tools in India and worldwide will depend on robust trials, managing false positives, and integrating them as supportive aids in the diagnostic pathway, potentially making early detection of silent killers like pancreatic cancer a more common reality.
