Published on Jan. 15, 2026

Artificial intelligence making progress in the fight against cancer is an aspiration shared all around the world. It is still in its early stages, but the first encouraging tales are now emerging from the Affiliated People’s Hospital of Ningbo University in China.

Qiu Sijun, a retired bricklayer aged 57, had gone for a routine checkup for diabetes. Three days later, he received a call from a doctor he hadn't met before. A tumor had been detected in his pancreas, he was told, and if he could please come in.

He was lucky. Dr. Zhu Kelei, head of the hospital's pancreatic department, had found the tumor very early. The doctor was able to operate and the patient has already returned to tend to his garden.

The early detection happened during the clinical trial of a new AI tool called DAMO PANDA, where PANDA stands for "pancreatic cancer detection". It was developed at DAMO Academy, a research institute affiliated with Alibaba, originally founded by Jack Ma, the People’s Daily reported (in Chinese) .

The Chinese newspaper quoted an earlier article written by Vivian Wang in the New York Times. The reporter had gone to talk to the patient and the doctor, delivering a thoroughly researched piece that raised questions about the dangers of "false positives" in early detection, but also acknowledged this progress in the practical use of AI in a Chinese hospital.

The innovation here lies in what the New York Times calls the first systematic use of a large AI model to analyse non-contrast computed tomography (CT) scans. It can be applied to several different major types of cancers, including lung, colorectal, liver, gastric, breast, and esophageal.

Even in China, where the costs for checkups are not as steep as in the West, the major benefit of this new method is its cost-effectiveness. Traditional methods such as ultrasound, CT scans, gastroscopy and colonoscopy would cost at least RMB 3,000 (around US$ 430) per patient to screen for such a selection of cancers, the financial paper Huaerjie Jianwen writes.

By contrast, the "One-Scan, Multi-Screening" approach based on non-contrast CT plus AI costs less than RMB 200 (around US$ 29), the paper estimates. This makes "large-scale screening of asymptomatic populations a realistic possibility", the Chinese journalist suggests.

Crucially, CT scans with contrast agents for cancer screening examinations are not only too expensive, but also carry health risks through the chemicals administered. This made the researchers at DAMO Academy look for a way to increase the detection rates of non-contrast CT scans.

To train their AI model, the researchers asked a radiologist to manually annotate the contrast CTs of 2,000 known pancreatic patients with the locations of their lesions. The engineers then "algorithmically mapped the highlighted lesions onto the same patients' non-contrast CTs", writes Vivian Wang in the New York Times.

Since the clinical trial with the "AI Panda" began, 180.000 abdominal or chest CTs have. been analysed, resulting in the detection of about two dozen cases of pancreatic cancer, including ductal adenocarcinoma and neuroendocrine tumors.

A few of them were early stage and the operations were successful. "I think you can 100 percent say AI saved their lives", says Dr. Zhu.

At a recent medical congress, the inventors of the new tool described the case of a 75-year-old woman who also underwent a routine CT scan. Her tumor markers had been only slightly elevated. DAMO PANDA identified a 76 percent probability of pancreatic cancer.

A cancer was then found during a follow-up examination with contrast-enhanced imaging in the woman's pancreatic head. When operated, the malignant growth was revealed to measure only one millimetre, representing an "exceptionally early-stage detection" of cancer, the researchers said.

Pancreatic cancer is known to be very painful in the final stages and also notoriously hard to detect. It has a five-year survival rate of less than 10 percent. It has a relatively low prevalence, however, with only 11 cases per 100,000 people.

Effective tools for wider screenings among the population require high specificity to avoid large numbers of false positives, which can bring psychological suffering and other risks to affected patients.

The researchers at the Alibaba-affiliated institute claim that their DAMO PANDA qualifies. “Among 1000 tests of healthy individuals, only one would be a false positive,” one of them said during the "AI for Good Global Summit" in October.

GE Healthcare and the DAMO Academy have signed an agreement to jointly promote the new AI tool in the United States, by integrating the new Chinese AI screening technology with hardware developed by GE Healthcare, reported the tech portal Xinlang Keji.

The DAMO PANDA has already received the U.S. Food and Drug Administration's (FDA) so-called Breakthrough Device Designation (BDD), Xinlang Keji wrote. This gives the developers hope for fast approvals to commercialise the new device co-developed with GE Healthcare.

“We hope that these innovative medical AI technologies can benefit all of humanity,” said Zhang Ling, Senior Technical Expert and Lead of Multi-Cancer Screening Technology at Alibaba DAMO Academy Medical AI Lab, during the conference.

After applying the new AI model to gastric cancer screening with “DAMO GRAPE,” as well as to “DAMO iAORTA” for detecting acute aortic syndrome (AAS), the researchers are now negotiating collaborations in Singapore, Japan, Saudi Arabia, New Zealand, and Antigua and Barbuda, they said during the summit.