Nonrigid registration for tracking incompressible soft tissues with sliding motion

Area of Science:

• Medical Imaging
• Computational Anatomy
• Surgical Technology

Background:

• Respiration causes soft tissue deformation and sliding, impacting minimally invasive abdominal surgery assessment accuracy.
• Nonrigid registration is crucial for eliminating respiratory effects in surgical assessments.
• Incompressibility is a key property of soft tissues during deformation, requiring consideration in registration algorithms.

Purpose of the Study:

• To develop an incompressible nonrigid registration method for tracking deformable soft tissues.
• To specifically address and compensate for sliding motion in soft tissue registration.

Main Methods:

• A novel nonrigid registration framework utilizing stationary velocity fields in the log domain.
• Fourier-based Helmholtz-Hodge decomposition (FHHD) to extract divergence-free and harmonic components.
• Adaptive weighting to maintain deformation incompressibility and compensate for sliding motion.

Main Results:

• The proposed method demonstrated superior performance compared to state-of-the-art techniques in handling large sliding motion and incompressible deformation.
• Achieved high registration accuracy with mean surface distance (MSD) of 0.631 mm, Hausdorff distance (HD) of 6.000 mm, mean corresponding distance (MCD) of 3.555 mm, and Dice similarity coefficient (DSC) of 0.970.
• Demonstrated strong incompressibility with relative volume change (RVC) of 0.006 and Jacobian determinant (J) close to 1.

Conclusions:

• The developed incompressible nonrigid registration method significantly improves accuracy for tissues with large sliding motion.
• This advancement holds potential for enhancing current minimally invasive abdominal surgical procedures.