AI guidance system for robotic mitral valve procedures

This technology is an AI-based intraoperative support platform that integrates real-time object recognition with 4D patient-specific mitral valve reconstruction to guide robotic mitral valve repair surgery.

Unmet Need: Real-time guidance for complex robotic mitral valve repair

Robotic mitral valve repair is currently performed using surgeon experience, preoperative imaging review, and intraoperative visual judgment as the primary decision-making tools. While robotic systems improve visualization, the procedure remains technically demanding, lacks haptic feedback, and depends heavily on the expertise of a small number of high-volume centers. There is no standardized, real-time intraoperative support system to assist with structure identification or repair strategy. As a result, outcomes can vary significantly, training pathways are prolonged, and broader adoption of this minimally invasive approach remains limited.

The Technology: AI-guided co-pilot for robotic mitral valve repair

This technology is a software platform that guides robotic mitral valve repair by analyzing real-time surgical video and multimodal cardiac imaging using a deep learning-based co-pilot algorithm. This model tracks surgical instruments, cardiac structures, and mitral valve components to guide procedural steps and provide contextual visual overlays. In parallel, multimodal preoperative imaging and intraoperative data are used to generate a time-resolved (4D) reconstruction of the mitral valve that reflects patient-specific anatomy and valve dynamics. Together, these components provide structured anatomical references and procedural guidance informed by both live surgical context and reconstructed mitral valve anatomy.

Applications:

  • Intraoperative guidance for robotic mitral valve repair
  • Standardization of complex mitral valve repair techniques
  • Preoperative planning using patient-specific mitral valve modeling
  • Intraoperative decision support for repair strategy selection
  • Surgical training and simulation for robotic mitral procedures

Advantages:

  • Provides real-time intraoperative visual and procedural guidance
  • Integrates live surgical context with patient-specific 4D anatomy
  • Reduces variability in complex mitral repair procedures
  • Shortens learning curve for robotic mitral surgery
  • Software platform compatible with existing robotic systems
  • Enables standardization across institutions and surgical teams

Lead Inventor:

Arnar Geirsson, M.D.

Related Publications:

Tech Ventures Reference:

Quick Facts:
Tags
Decision support systemHaptic technologyHeart valveMedical imagingMitral valveOutline of object recognitionSurgery
Inventors
Arnar GeirssonMakoto Mori
Manager
Dovina Qu
Departments
Surgery
Divisions
Columbia University Medical Center (CUMC)
Reference Number
CU26049
Release Date
2026-03-20