Projective biomechanical depth matching for soft tissue registration in laparoscopic surgery

Area of Science:

• Medical Imaging
• Surgical Robotics
• Computational Biomechanics

Background:

• Accurate registration of preoperative data with intraoperative sensor data is crucial for computer-assisted surgery systems.
• Image-based registration in laparoscopic surgery is challenging due to soft tissue deformations.

Purpose of the Study:

• To present a novel biomechanical approach for soft tissue registration.
• To align preoperative CT data with intraoperative stereo endoscopic image data for enhanced surgical guidance.

Main Methods:

• A two-step registration process involving 3D surface mosaic alignment and biomechanical model updating.
• Utilizing stereo endoscopic images to reconstruct partial surfaces for initial organ alignment.
• Projecting the biomechanical model onto new surfaces to derive boundary conditions for model updates.

Main Results:

• The method was evaluated using in silico (human liver model) and in vivo (porcine data).
• Quantitative in silico results demonstrated stable biomechanical model behavior.
• Achieved a root-mean-square deviation of volume vertices under 3 mm with adjusted parameters.

Conclusions:

• The developed approach offers a fully automatic, featureless, non-rigid registration method.
• The technique shows capability in managing dynamic tissue deformations during surgery.
• Further research on human tissue behavior is recommended for clinical application.