Assessment of hepatic steatosis based on virtual non-contrast computed tomography: Initial experiences with a photon counting scanner approved for clinical use.

Purpose: The present study evaluates the diagnostic accuracy of virtual non-contrast (VNC) images acquired with the first clinical photon counting CT scanner for the assessment of hepatic steatosis.

Method: 140 patients were retrospectively enrolled in this study. The following CT indices were calculated according to the attenuation on VNC and true non-contrast (TNC) images: Liver attenuation CT (L), difference between liver and spleen attenuation CT (L-S) and the ratio between liver and spleen attenuation CT (L/S).

Results: Considering previously reported cut-off values, the sensitivity, specificity, positive and negative predictive value (PPV, NPV) for the detection of hepatic steatosis on VNC images was 94%, 87%, 30% and 99.5% for CT (L), 52%, 99%, 81% and 96% for CT (L-S) and 67%, 98%, 74% and 94% for CT (L/S). When adjusting the cut-off values according to the present study, the sensitivity, specificity, PPV and NPV for the detection of hepatic steatosis on VNC images was 94%, 92%, 41% and 99.6% for CT (L), 96%, 90%, 46% and 99.6% for CT (L-S) and 95%, 99.6%, 42% and 99.6% for CT (L/S).

Conclusions: In conclusion, the spectral datasets acquired with the first clinical PCCT scanner enable the reconstruction of VNC images that are reliable for the image-based assessment of hepatic steatosis. By adjusting the cut-off values all tested CT indices perform well and offer an excellent sensitivity as well as specificity.