Optimization of Virtual Monoenergetic Imaging In Photon-Counting CT: Material Decomposition Evaluation with Phantom and Biological Tissue.
Abstract
Purpose
To optimize virtual monoenergetic imaging (VMI) in spectral photon-counting CT (SPCCT) through material decomposition, aiming to improve soft-tissue contrast and image quality.
Methods
A QRM phantom containing water, iodine (10 and 15 mg/mL), hydroxyapatite (100 and 200 mg/cm³), adipose, and muscle tissue equivalents was scanned using a five-bin SPCCT protocol (7–118 keV). Two- and three-material decomposition approaches were evaluated to select optimal energy-bin and material combinations for VMI generation. Image quality was assessed using the contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), relative error (RE), and a clinician score (CS). The optimal two-material configuration was further validated on an ex vivo beef steak containing iodine inserts.
Results
For two-material decomposition, the combination of iodine 15 + water (Bins 2 and 4) produced a 40 keV VMI with CS = 9.5, CNR = 1.51, and SNR = 6.45. Three-material decomposition yielded comparable results, with the highest CNR (1.71) not coinciding with the highest CS. Application to beef tissue demonstrated low RE (0.7–6.9%) and high CNR/SNR, confirming the transferability of optimized parameters from phantom material to biological tissues.
Conclusion
Optimized SPCCT-derived VMIs can enhance soft-tissue contrast while maintaining low noise. The two-material decomposition approach (iodine 15 + water) shows robust performance and applicability to realistic tissue, highlighting its potential for improved diagnostic imaging in clinical CT.