Alibaba DAMO's LiON liver cancer model beats radiologists, catches missed metastases
jiqizhixin · x · 2026-08-30
Alibaba DAMO Academy released DAMO LiON, a liver cancer detection model:
- Improved architecture captures both lesion–whole-liver relationships and the local texture/boundary details critical for small tumors;
- Handles hard cases — fatty liver, cirrhosis, post-surgical livers — and excels at detecting liver metastases often overlooked by radiologists;
- Identifies malignant tumors more accurately than radiologists: AI assistance cut reading time by 27%, raised sensitivity for malignant tumors by 11.5%, and brought junior doctors to senior-level performance;
- Deployed in a hospital for two months, it caught 15 previously missed liver metastases across 10,000+ patients, changing their treatment plans.
More from Apps
- Google's Co-Scientist Accelerates Discovery in Materials Science and Biology — SRSchmidgall · 2026-08-30
- Opinion: Omarchy is Far From Ready, Linux Desktop Year Not Here Yet — steren · 2026-08-30
- User switches to GLM 5.3 Flash for docs, sees zero quality drop and lower cost than GPT — abacaj · 2026-08-30
- Case Study: Using AI to Plan a $1M ARR Strategy for a Blogger — huangyun_122 · 2026-08-30
- Cardgo: An AI Card Context Manager for Cross-Model Usage — dieutrinhngoc · 2026-08-30
- Hailuo MiniMax Design passes third prototype test: Gaudí-style e-commerce UI — Hailuo_AI · 2026-08-30