Monoenergetic Energy Selection for Fat–Liver Differentiation In Photon-Counting CT
Abstract
Purpose
To quantify energy- and size-dependent HU shifts in PCCT monoenergetic imaging and their impact on fat–liver separability.
Methods
A multi-energy phantom with soft-tissue insert modules was scanned on a Siemens NAEOTOM Alpha PCCT in two size configurations (20 cm and 40 cm). Virtual monoenergetic images (VMI) from 40–190 keV were reconstructed with fixed iterative reconstruction strength and constant reconstruction parameters. Mean HU were measured with consistent ROIs in adipose, liver, and background. Fat–liver separability was ΔHU = HU_Liver − HU_Adipose; size dependence was ΔHU_size = HU_40cm − HU_20cm.
Results
At 40 keV, adipose was −150.1 HU (20 cm) vs −122.5 HU (40 cm) and liver was 71.0 HU vs 80.5 HU, yielding ΔHU of 221.1 vs 203.0 HU and positive size bias (ΔHU_size) of +27.7 HU for adipose and +9.5 HU for liver; background bias was +14.7 HU. At 70 keV, ΔHU was 134.9 vs 129.6 HU with reduced size sensitivity (≤~5 HU for adipose/liver/background). At 190 keV, adipose was −37.9 HU vs −48.2 HU and liver was 63.5 HU vs 54.0 HU, yielding ΔHU of 101.4 vs 102.2 HU and a sign reversal in size bias (adipose −10.3 HU; liver −9.5 HU).
Conclusion
Fat–liver separability decreases with increasing keV, while size-related HU bias is greatest at low keV and can reverse at high keV. Mid-range energies (~70 keV) maintain substantial separability while minimizing size bias, supporting robust energy selection and QC expectations for fat-related workflows.