Room: Room 202
Purpose: To investigate the effects of kVp, reconstruction kernel/filter, and reconstructed image thickness on water equivalent diameter (WED) in computed tomography (CT).
Methods: An anthropomorphic lung phantom (Kyoto Kagaku N1) was scanned at various energies (70-140 kVp) and reconstructed with a smooth (B31f) and sharp (B80f) filter at two image thicknesses (1 and 5 mm). Scan parameters were based on the AAPM Adult Routine Chest CT Protocols V2.1 for the Siemens Flash. The phantom was segmented and the WED was calculated for each axial image using the method described in AAPM Report 220. Additionally, patient raw data from another thoracic quality improvement project acquired with the institutional protocol were reconstructed using these four kernel/thickness combinations for comparison.
Results: A comparison of individual phantom voxel HU values showed relatively large local variations due to kVp and reconstruction parameters. However, differences in the calculated WED due to kVp, reconstruction kernel, and image thickness were relatively small. Differences were less than 10 mm for the phantom, which had WED in the range of 206-262 mm. Using 120 kV, B31f WED values for each image location as the reference when comparing other kV and kernel combinations, the maximum differences through the volume were 2.3% and 3.3% for 5 mm and 1 mm thicknesses, respectively. Patient data showed similar small changes in WED (<5 mm) due to reconstruction kernel and thickness.
Conclusion: While kVp, reconstruction kernel, and image thickness can often produce substantial variations in local HU values, the global effect on WED is relatively small. For calculations of size-specific dose estimates (SSDE) and patient dose, WED should be relatively consistent under changes in kVp, kernel, and thickness.
Funding Support, Disclosures, and Conflict of Interest: The Ohio State University and Qaelum have a research agreement to evaluate and develop the dose management solution DOSE.