HANGZHOU, Jan. 14 (Xinhua) -- When a hospital in eastern China phoned a patient about a critical tumor discovered not by a doctor but by artificial intelligence (AI), he nearly hung up, almost convinced it was a scam.
His surgeon, Shen Yiyu, credited the AI with saving his life, highlighting a quiet revolution in early disease detection that is now spreading from China's mega‑labs to its most remote county clinics.
The patient, surnamed Fang, who is from the city of Jiaxing in east China's Zhejiang Province, visited a local hospital in early 2025 due to a persistent cough. A routine chest CT appeared normal, and he went home thinking little of it. Days later, the hospital called back to inform him that an AI system reviewing his scan had detected a suspicious shadow on his pancreas, an organ that had not even been the focus of the original exam.
"I almost dismissed the call as a scam," Fang said, a decision that could have proved fatal. Further tests confirmed the existence of a pancreatic neuroendocrine tumor, the same rare cancer that afflicted Steve Jobs. Because the anomaly had been detected early by the AI, surgeons were able to remove it successfully in a minimally invasive operation last April.
"Without that AI alert, I would have missed my chance," Fang said.
Fang's case is not an isolated one. At the Affiliated People's Hospital of Ningbo University, also in Zhejiang, a 57‑year‑old retiree was similarly diagnosed by the same AI project during a scan for stomach discomfort.
Since the system went live there in late 2024, it has reviewed over 180,000 abdominal and chest CTs, uncovering more than 20 pancreatic cancers -- 14 of them early‑stage cases and many that had never triggered any alert before going through the AI system.
"For some of these patients, I can say with certainty that AI saved them," said Zhu Kelei, vice president of the hospital.
The technology, named DAMO PANDA, has been developed by Alibaba's DAMO Academy to take on one of medicine's most lethal adversaries, namely pancreatic cancer.
In November 2023, a paper in Nature Medicine showed that DAMO PANDA had retrospectively identified 31 missed lesions, including two early-stage cancers later cured by surgery, from over 20,000 patient records.
In April 2025, the model earned the U.S. Food and Drug Administration's "breakthrough device" designation, a first for any medical AI developed by a Chinese tech company, fast‑tracking its path to global clinical use.
A MULTIPLYING EFFECT
The growing integration of AI into China's healthcare system is providing its citizens with accelerated drug development, enhanced diagnostic and treatment precision, and more effective rehabilitation therapies.
At a drug research facility of SSY Group Limited in Shijiazhuan, capital city of north China's Hebei Province, an AI-powered virtual screening system runs around the clock, scanning a molecular database.
"With intelligent screening, we can identify more than 100 new compounds from a library of 300 million in just over a month. Previously, this would have taken two to three years," said Sun Lijie of the company.
With the help of AI, residents in remote and less-developed areas, in particular, are able to enjoy better and swifter treatment at a much lower cost.
DAMO PANDA, notably, has been deployed not only in urban tertiary hospitals but also in remote county facilities in Zhejiang, southwest China's Yunnan Province, Hunan Province in central China and Shanxi Province in north China, bringing cutting-edge screening closer to patients in underserved areas.
"This approach enhances the chances of early detection, without placing any additional financial or physical burdens on patients," Zhu said.
At a township health center in Boye County in Hebei, an AI system flagged a red alert, indicating a doctor's prescription as "unsuitable with contraindications" and simultaneously gave a viable alternative. The system, developed by Chinese AI firm iFLYTEK, has been used by local primary care doctors for two years to improve diagnostic accuracy.
"The core mission of AI is to empower doctors," said Dong Bin, deputy general manager of iFLYTEK Brank Marketing Center and director of Xunfei Healthcare Brand Marketing Department. The system, powered by the Spark Medical Large Language Model, now serves over 250,000 primary care doctors across more than 77,000 grassroots institutions nationwide, providing over 1.1 billion AI-assisted suggestions.
"Human-machine collaboration is the new healthcare model," Dong noted. He explained that AI can elevate a mid-level doctor's capabilities and free top-tier doctors from repetitive tasks, allowing them to focus on complex patient care.
What makes the AI system transformative is its simplicity and ease of access.
Take DAMO PANDA for instance. For ordinary patients, this AI model offers a particularly suitable approach for opportunistic screening compared to specialized methods like blood tests or contrast-enhanced CT scans.
Opportunistic screening refers to the AI-powered analysis of routine non-contrast CT scans taken when patients visit hospitals for other symptoms or undergo regular check-ups, enabling the detection of hidden risks.
"The AI system can find the needle in the haystack and provide valuable information for doctors," explained Zhang Ling, a senior DAMO Academy algorithm expert.
Such impressive progress is supported by China's national drive in boosting AI-powered healthcare. In 2025, China's National Health Commission and other relevant departments issued a set of guidelines to promote the application of AI tools to improve the quality of the country's healthcare service. ■
