Incorporating Tissue Excision in Deformable Image Registration: A Modified Demons Algorithm for Cone-Beam CT-Guided...

Area of Science:

• Medical Imaging
• Image Registration
• Surgical Navigation

Background:

• Intraoperative imaging, such as cone-beam CT (CBCT), is crucial for guiding head and neck surgery.
• Existing deformable registration methods struggle with anatomical changes due to tissue excision during surgery.
• Accurate registration is essential for real-time surgical adjustments and improved patient outcomes.

Purpose of the Study:

• To investigate the feasibility of fast, accurate deformable registration for intraoperative images with surgical excisions.
• To develop and evaluate a modified deformable registration algorithm capable of handling excised tissue.
• To enhance the precision of CBCT-guided head and neck surgical procedures.

Main Methods:

• Developed a novel deformable registration approach by incorporating an extra dimension to manage excised voxels.
• Utilized a modified Demons deformable registration algorithm to address limitations with excised tissue.
• Employed cadaveric CBCT images from a prototype C-arm to simulate surgical deformation and excision.

Main Results:

• The modified algorithm successfully accounted for tissue excision during the registration process.
• Registration performance was quantified, demonstrating accurate handling of anatomical changes.
• The method avoided spurious deformations around excised tissue areas, ensuring registration integrity.

Conclusions:

• The proposed deformable registration method effectively handles excised tissue in intraoperative imaging.
• This advancement is well-suited for improving CBCT-guided head and neck surgical procedures.
• The technique offers enhanced accuracy and reliability for real-time surgical navigation.