Alibaba DAMO Academy developed the liver cancer diagnostic AI model DAMO LiON in partnership with Shengjing Hospital and other institutions. The model was tested in a two-month real-world trial involving over 10,000 patients and identified 15 previously missed malignant tumors, most around 1 cm in size. A related paper was published in Nature Medicine. DAMO LiON uses an improved network architecture to accurately detect small liver lesions obscured by conditions like cirrhosis and complex anatomy, offering higher accuracy than radiologists. When used alongside human doctors, it cuts reading time by 27% and raises sensitivity by 11.5%, helping junior clinicians reach senior levels of performance. In practice, if the AI’s initial diagnosis differs from initial human assessments, it is reviewed by senior doctors, with multidisciplinary reviews triggered when needed. The model correctly identified a liver metastasis in a 65-year-old bladder cancer patient whose initial tests showed only normal tumor markers and calcification, altering his chemotherapy plan. Since 2017, DAMO Academy has expanded its medical AI capabilities from screening to diagnosis to help more patients receive timely accurate treatment.