China's Alibaba unveils world’s first AI tool for early stomach cancer detection

The AI model, named GRAPE, was developed jointly by Alibaba’s Damo Academy and Zhejiang Cancer Hospital. It leverages deep learning to detect and segment stomach cancer, according to the South China Morning Post.

Stomach cancer causes 260,000 deaths annually in China. Early detection can boost five-year survival rates from under 30 percent to more than 90 percent, yet many patients avoid invasive endoscopy, and the quality of screening varies across hospitals, reported by Xinhua. Cheng Xiangdong, Party chief of Zhejiang Cancer Hospital, said fewer than 30 percent of patients in China follow doctors' advice to undergo an endoscopy.

Doctor conducts CT scan on patient. Photo courtesy of Arizona Center for Cancer Care

The model was tested on over 70,000 patients, producing promising results. The study, published in Nature Medicine on Tuesday, revealed that the AI model achieved 85.1 percent sensitivity and 96.8 percent specificity, outperforming radiologists by 21.8 percent and 14 percent.

Zhang Ling, a senior algorithm expert at DAMO Academy and corresponding author of the study, noted, "This marks the first time non-contrast CT can detect early gastric cancer."

Cheng highlighted the potential for the model to transform gastric cancer screening in China and globally, with early-stage detection rates likely to "increase significantly."

Gastric cancer is the fourth leading cause of cancer-related deaths worldwide, often presenting no symptoms even in advanced stages. Damo added that, in a real-world case, a patient with late-stage disease could have been diagnosed six months earlier via CT scans, but the condition was missed by radiologists.

Cheng stated that the AI model has already been deployed in high-risk areas for gastric cancer in China, with plans for broader expansion both domestically and internationally.