Single-Phase Surrogate for True Non-Contrast Liver CT Using 190-keV Monoenergetic Imaging on Photon-Counting CT: Implications for Hepatic Steatosis Assessment.
RATIONALE AND OBJECTIVES To evaluate whether 190-keV virtual monoenergetic imaging (VMI) using photon-counting detector CT (PCD-CT) can serve as a surrogate for true non-contrast (TNC) in hepatic steatosis quantification and to compare its attenuation agreement with virtual non-contrast (VNC). METHODS Forty-nine patients who underwent abdominal PCD-CT and MRI-proton density fat fraction (PDFF) within 2 months were retrospectively analyzed. VNC and 190-keV VMI were reconstructed from equilibrium-phase scans. Liver attenuation was measured on TNC, VNC, and 190-keV VMI in four hepatic segments. Pearson correlation and Bland-Altman analyses assessed agreement with TNC. Receiver operating characteristic analysis evaluated diagnostic performance for steatosis (MRI-PDFF ≥ 5%). Statistical analyses were performed using MedCalc (version 23.4) and GraphPad Prism (version 10.6.1). Statistical significance was defined as p < 0.05. RESULTS The study included 49 patients (30 men, 19 women; mean age, 66.7 ± 11.2 years). In whole-liver analysis, 190-keV VMI correlated strongly with TNC (r=0.921, p<0.001), comparable to VNC (r=0.925). In the anterior segment, VNC showed a larger positive bias relative to TNC (+5.5 HU), whereas 190-keV VMI showed a smaller bias (+2.2 HU) that more accurately reproduced the regional attenuation pattern. For detecting steatosis in the anterior segment, a significant difference in area under the curve was observed between TNC and VNC (0.921 vs 0.817, p=0.018), whereas no significant difference was found between TNC and 190-keV VMI (0.921 vs 0.869, p=0.149). CONCLUSION 190-keV VMI may serve as a practical surrogate for TNC in hepatic steatosis quantification, demonstrating reduced segment-dependent bias compared with VNC in artifact-prone regions.