A quantitative analysis of biomechanical lung model consistency using 5DCT datasets.

Area of Science:

• Pulmonary Biomechanics
• Medical Imaging Analysis
• Computational Physiology

Background:

• Lung biomechanical models are crucial for understanding lung function.
• Tissue elasticity distribution is a key parameter in these models.
• Estimating human lung elasticity lacks ground truth, necessitating model consistency checks.

Purpose of the Study:

• To propose and validate a consistency check for lung biomechanical estimation techniques.
• To assess the stability of lung dynamics over short time intervals (2-3 minutes).
• To verify the reliability of elasticity distribution estimations in lung models.

Main Methods:

• Acquired two sets of 12 fast helical free breathing computed tomography (FHFBCT) scans for 10 subjects.
• Generated five-dimensional CT (5DCT) breathing motion models from each scan set.
• Created finite element biomechanical lung models, optimized elasticity values, and compared resulting distributions.

Main Results:

• 90% of lung voxels showed less than 2 kPa Young's modulus difference between estimations (mean 0.6 kPa).
• 97% of lung voxels exhibited less than 2 mm displacement difference (mean 0.48 mm).
• Consistent elasticity histograms were observed for parenchymal tissues across estimations.

Conclusions:

• Demonstrated consistent estimation of biomechanical lung models using motion-model based imaging datasets.
• Validated the stability and reliability of the proposed consistency check method.
• Confirmed that models created from scans acquired at different breaths yield stable results.