This prospective study assessed whether virtual non-contrast (VNC) images generated from dual-layer spectral CT can reliably replace true non-contrast (TNC) acquisition in abdominal imaging.
A total of 100 patients undergoing triphasic abdominal CT were evaluated. VNC datasets reconstructed from both arterial (VNC_A) and portal venous (VNC_P) phases were compared with TNC images in terms of attenuation accuracy, image quality, lesion characterization, calcification assessment, and possible radiation dose reduction. VNC_A generally showed better agreement with TNC attenuation values than VNC_P, while lesion size remained stable across all image sets. However, both VNC techniques consistently underestimated attenuation values in abdominal organs and focal lesions, and it significantly reduced calcification density. Calcification diameter was also slightly underestimated on arterial-phase VNC images. Although VNC reconstructions demonstrated similar or even higher objective image quality, with improved signal-to-noise and contrast-to-noise ratios, radiologists rated their subjective image quality significantly lower because of a blotchier image texture. Most VNC examinations were considered a reasonable approximation of TNC, but only 12% of portal venous VNC datasets were judged sufficiently accurate to fully replace dedicated non-contrast imaging.
Replacing the non-contrast phase with VNC images would reduce radiation exposure by approximately one-third in the evaluated protocol. Nevertheless, the authors conclude that current dual-layer spectral CT VNC reconstructions cannot yet serve as a reliable substitute for true non-contrast abdominal CT in routine clinical practice, particularly when accurate attenuation measurements or calcification assessment are required.