Sign in Subscribe
Cancer research

AI Model Detects Early Signs of Stomach Cancer with High Accuracy

Strategic Insights

— 2 min read

Researchers from the Zhejiang Cancer Hospital and Alibaba's DAMO Academy have developed an AI model, DAMO GRAPE, that can identify early signs of stomach cancer in routine CT scans. The model was tested on over 70,000 patients, yielding promising results. The study, published in the journal Nature Medicine, suggests that the AI model can detect cancer in patients who may have avoided invasive endoscopy and can help increase five-year survival rates from under 30 percent to over 90 percent.

Key Takeaways:

  • The DAMO GRAPE model achieved 85.1 percent sensitivity and 96.8 percent specificity, outperforming radiologists by 21.8 percent and 14 percent respectively.
  • The model detected cancer in 17.7 percent and 24.5 percent of scanned patients at high risk, including many without symptoms, at two regional hospitals.
  • The AI model can fill the gap in opportunistic detection of early gastric cancer during imaging examinations, according to Hu Can, leading author and a gastric surgeon at Zhejiang Cancer Hospital.
  • The model was trained on a dataset of gastric non-contrast CT images, overcoming difficulties such as the stomach's variable shape, internal content interference, and early-stage cancer cells limited to the mucosal layer.
  • Zhang Ling, corresponding author and senior algorithm expert at DAMO Academy, stated that non-contrast CT is a common non-invasive examination used for checkups and outpatient care, which is both cost-effective and efficient.
  • Cheng Xiangdong, Party chief of the Zhejiang Cancer Hospital, mentioned that early detection can increase five-year survival rates from under 30 percent to over 90 percent, but many patients avoid invasive endoscopy, and the quality of screening varies among hospitals.

Statistics:

  • The AI model detected abnormalities in scans taken two to ten months before diagnosis in 11 patients diagnosed with stomach cancer.
  • A notable case involved a 45-year-old patient diagnosed with late-stage stomach cancer, where the AI model identified cancerous markers when reanalyzing a CT scan taken six months prior to diagnosis.
  • Over 260,000 deaths occur every year in China due to stomach cancer.
  • The DAMO GRAPE model has been deployed in high-risk areas for gastric cancer in China, with plans for further expansion at home and abroad.

Sources:

  • "AI model detects early signs of stomach cancer with high accuracy" (Source not provided, but likely referring to a press release or article on the study)
  • Nature Medicine (Journal where the study was published on 25 June 2025, no specific article or volume number provided)
  • NAMPA (News agency where the article was published on 26 June 2025)

Read more