A general framework in single and multi-modality registration for lung imaging analysis using statistical prior...

Area of Science:

• Medical Imaging
• Pulmonology
• Computational Anatomy

Background:

• Chronic Obstructive Pulmonary Disease (COPD) analysis requires precise anatomical and functional lung data.
• Current imaging techniques may lack the resolution or integration needed for comprehensive COPD assessment.
• Multi-slice computed tomography (MSCT) and single photon emission computed tomography (SPECT) offer complementary information.

Purpose of the Study:

• To develop and validate a novel, high-performance non-rigid registration algorithm for MSCT/SPECT fusion.
• To accurately map lung lobe information from MSCT onto functional SPECT imaging.
• To enhance the guidance of lung volume reduction surgery for COPD patients.

Main Methods:

• A multi-channel technique for simultaneous registration of MSCT to ventilation and perfusion SPECT scans.
• Implementation of a novel, parameter-efficient function to optimize registration performance.
• Validation using both clinical lung imaging datasets and synthetic data.

Main Results:

• The proposed MSCT/SPECT registration method significantly improves registration accuracy and stability.
• Accuracy improvements reached up to 23%, and stability improvements reached up to 54%.
• The multi-channel approach demonstrated superior registration accuracy compared to single-channel methods.

Conclusions:

• Accurate MSCT/SPECT alignment enables effective fusion of lung lobes onto functional imaging.
• This fusion facilitates lobar lung activity analysis for improved COPD diagnosis and treatment planning.
• The developed algorithm provides a powerful tool for advanced COPD management.