Intraoperative localized constrained registration in navigated bronchoscopy.

Area of Science:

• Medical Imaging
• Respiratory Medicine
• Surgical Navigation

Background:

• Electromagnetic navigated bronchoscopy (EMN) accuracy is challenged by lung anatomy shifts post-CT acquisition.
• Breathing, coughing, and tool interaction can degrade initial rigid registration accuracy during EMN procedures.
• A dynamic, localized registration update is crucial for maintaining navigation precision throughout bronchoscopy.

Purpose of the Study:

• To develop and evaluate an intraoperative local registration method for enhancing navigation accuracy in EMN.
• To address the limitations of initial rigid registration in dynamic bronchoscopic procedures.
• To minimize the impact of anatomical shifts and patient movement on EMN accuracy.

Main Methods:

• An intraoperative local registration method was adapted from existing techniques for initial CT-to-patient registration.
• The method optimizes registration within a specific area of interest, such as near a biopsy site.
• Validation was performed using an experimental breathing phantom with induced respiratory movements and deformations.

Main Results:

• The local registration method demonstrated successful application across all 15 repetitions in the airway phantom.
• Mean registration accuracy significantly improved from 11.8-19.4 mm to 4.0-6.7 mm.
• Accuracy improvements varied slightly across different segments of the airway model.

Conclusions:

• A novel local registration method was developed to update and enhance initial image-to-patient registration during navigated bronchoscopy.
• Successful testing in a breathing phantom setup validates the method's potential.
• Further research is required for method automation and verification in human clinical trials.