Direava validates Surgical VLM during live gastric-cancer surgery at Keio University Hospital
Direava completed an operating-room demonstration of its Surgical VLM, showing real-time surgical-context recognition and dialogue during gastric-cancer surgery.
Evidence notes
- The clinical demonstration took place at Keio University Hospital on 20 February 2026.
- Surgical VLM combines intraoperative images with structured language to recognise anatomy, procedure state and relevant next-step context.
- Evaluation scores reached 84.7 percent for anatomical accuracy, 82.9 percent for clinical usefulness and 97.4 percent for language fluency.
Company context
Direava develops vision-language artificial intelligence for surgery. Its software interprets operative video and clinical context to support workflow understanding, documentation and decision assistance, creating an intelligence layer that can be integrated with surgical imaging and robotic platforms.