Room: Track 2
Purpose: To systematically analyze the ability to perform lung elastography from fast helical free-breathing CT (FHFBCT) scans by measuring the consistency of elasticity estimation in paired FHFBCT scans.
Methods: A set of 10 patients were retrospectively studied with 25 FHFBCT scans acquired for each patient. Deformable image registration (DIR) was then performed between an arbitrary reference scan and each of the 24 remaining FHFBCT scans to obtain ground-truth deformation vector fields (DVFs). For each DVF, a well-validated inverse elasticity approach, which employs a lung biomechanical model, was performed to obtain multiple elasticity estimations for each reference scan voxel. The estimated elasticity values for all 24 model iterations were clustered by reference scan voxel location. Global statistics were calculated for parenchymal voxels (-600 to -900 HU). Mean elasticities were then calculated for each voxel, and the percentage of voxels differing from the mean voxel elasticity by greater than 1 kPa were recorded. The percentage of voxels converging successfully in accordance with the ground-truth DVFs was also calculated.
Results: An average model convergence level of 89.3% of voxels was reached among patients, which supports the ability to achieve the ground-truth deformations with the resulting elasticities. The average percentage of voxels whose elasticity values differed from their voxel-specific average elasticity by greater than 1 kPa was 6.85%, which supports the lung elastography methods using a limited number of FHFBCT scans. Furthermore, the average percentage of clustered elasticity values per voxel within 1 kPa of the mean voxel value is 93.48%, which suggests that as few as 4 FHFBCT scans may be adequate for elasticity estimation.
Conclusion: Consistent elasticity results between a reference FHFBCT scan and each of the additional 24 scans demonstrates the feasibility of performing lung elastography with a small number of FHFBCT scans.