Evaluation of virtual non-contrast detector-based spectral CT images in comparison to true unenhanced images in 20 rabbits

Manon Mikic ^1* Philipp Lietz ¹ Julie-Ann Dierig ² Sebastian Meller ¹ Michael Pees ² Kristina Merhof ¹

• ¹ Clinic for Small Animals, University of Veterinary Medicine Hannover, Hanover, Germany
• ² Department of Small Mammal, Reptile and Avian Medicine and Surgery, University of Veterinary Medicine Hannover, Hanover, Germany

Introduction: Spectral detector Computed Tomography (SDCT) enables generation of virtual noncontrast (VNC) images derived from a post-contrast scan, as previously investigated in healthy dogs. This technique is especially promising for awake scanned rabbits where motion between the pre-and the post-contrast scans makes comparison challenging. This study aimed to determine the reliability of VNC images for assessing abdominal organs in 20 rabbits by evaluating their qualitative and quantitative parameters compared to true unenhanced (TUE) images. Our hypothesis were that the VNC series would be comparable to TUE series and that the quality of the VNC images would be equal to or even superior to the native images. Methods: Attenuation values of VNC and TUE series were assessed using a two one-sided t-test (TOST) and the signal-to-noise ratio was calculated for each ROI in the different series. Additionally, a qualitative assessment of the VNC images relative to TUE images was performed in consensus by a board-certified radiologist and a second year diagnostic imaging resident by evaluating the iodine subtraction, image quality and noise of VNC images based on a 5-point Likert scale. Results: A total of 219 regions of interest (ROIs) where drawn in abdominal organs. 72.1% of the ROIs displayed differences of less than 15 Hounsfield Units (HU) between TUE and VNC images. The differences in attenuation values of TUE and VNC were statistically significant (p < 0.05) for ≤ 5 HU in the spleen, for ≤ 10 HU and ≤ 15 HU additionally in the liver, musculature and renal cortices. These findings support the equivalence between VNC and TUE images. An average score of 4.4 was achieved for iodine subtraction across all patients, which encompasses all organ values, with no individual patient scoring below 4.0. Discussion: VNC images present a promising alternative to TUE images for abdominal organs without pathology in rabbits with the benefit of eliminating motion between the compared series and reducing examination time and radiation exposure by replacing pre-contrast scans. Further research is necessary to demonstrate the applicability of the technique to morphologically diseased organs.