Japan's First Surgical-Specialized Generative AI 'Surgical VLM' Developed to Act as a 'Brain' in the Operating Room

Direava has developed Japan's first surgical-specialized generative AI, 'Surgical VLM', which understands and discusses surgical situations in real-time. Clinical validation at Keio University Hospital demonstrated high anatomical accuracy and clinical utility. The company aims to commercialize the technology in 2026 to enhance surgical education and patient safety.
researchNQ 72/100出典:PR Times

📋 Article Processing Timeline

  • 📰 Published: March 28, 2026 at 16:04
  • 🔍 Collected: March 28, 2026 at 21:59 (5h 54m after Published)
  • 🤖 AI Analyzed: April 15, 2026 at 06:41 (416h 41m after Collected)
Under the GENIAC project, Direava Co., Ltd. has developed 'Surgical VLM,' a generative AI model specifically for surgery that can understand and engage in dialogues about operative procedures in real-time. Unlike previous tools focusing solely on anatomical identification, this AI functions like a surgeon's 'brain' to suggest procedural steps. Validated on February 20, 2026, at Keio University Hospital during gastric cancer surgeries, the AI achieved high scores for anatomical accuracy (84.7%) and clinical utility (82.9%). Direava plans to commercialize the technology within 2026, aiming to improve surgical training and quality of care.

FAQ

What does this AI do?

It understands surgical imagery in real-time and engages in dialogue with surgeons regarding next steps and potential risks.

When will it be commercialized?

The company is preparing for commercialization within 2026.