Alibaba DAMO Academy unveils AI model for non-invasive esophageal cancer screening

Alibaba DAMO Academy has developed DAMO EAGLE, an artificial intelligence model designed to screen for esophageal cancer using non-invasive plain CT scans. Developed in collaboration with institutions including Sichuan Provincial Tumor Hospital and Sun Yat-sen University Cancer Center, the model can identify early-stage and precancerous lesions without the need for endoscopy or contrast agents.

Research published in Nature Medicine indicates the model has been validated on over 80,000 cases across three countries. In opportunistic screening scenarios, DAMO EAGLE demonstrated a sensitivity of 90% for esophageal cancer and 52.5% for precancerous lesions, while maintaining a specificity of 99.2%.

Because esophageal cancer is often diagnosed at advanced stages, early detection is critical for survival rates. The model is designed to integrate into existing lung cancer screening workflows, potentially allowing clinicians to identify esophageal cancer risks during routine low-dose CT scans without increasing patient burden.