COMPUTATIONAL VASCULAR MORPHOMETRY FOR THE ASSESSMENT OF PULMONARY VASCULAR DISEASE BASED ON SCALE-SPACE PARTICLES

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Pulmonary Medicine

Background:

• Pulmonary vascular disease (PVD) significantly impacts patients with chronic obstructive pulmonary disease (COPD).
• Accurate assessment of pulmonary vasculature is crucial for understanding disease progression and patient outcomes.
• Current methods for vascular morphometry can be labor-intensive and subjective.

Purpose of the Study:

• To introduce a fully automatic computational vascular morphometry (CVM) approach for PVD assessment.
• To develop image-based biomarkers for PVD using automated vascular feature extraction.
• To evaluate the CVM approach in a large cohort of COPD patients.

Main Methods:

• Automatic extraction of intraparenchymal lung vasculature using scale-space particles.
• Development of biomarkers based on vessel radii, cross-sectional area, and blood volume.
• Validation using simulated data and analysis of 2,500 CT scans from COPD patients.

Main Results:

• The CVM approach successfully extracted lung vasculature across varying complexities.
• Image-based biomarkers demonstrated the ability to track vascular remodeling in COPD.
• The method showed potential for assessing PVD involvement in patient populations.

Conclusions:

• The automated CVM pipeline offers a robust method for PVD assessment.
• This approach can quantify vascular changes associated with COPD severity.
• The CVM method holds promise for future clinical studies in PVD and COPD research.