Spectral CT in patients with acute thoracoabdominal bleeding-a safe technique to improve diagnostic confidence and reduce dose?

Medicine (Baltimore). 2019 Jun;98(25):e16101. doi: 10.1097/MD.0000000000016101.

Abstract

Computed tomography (CT) protocols for the detection of bleeding sources often include unenhanced CT series to distinguish contrast agent extravasation from calcification. This study evaluates whether virtual non-contrast images (VNC) can safely replace real non-contrast images (RNC) in the search for acute thoracoabdominal bleeding and whether monoenergetic imaging can improve the detection of the bleeding source.The 32 patients with active bleeding in spectral CT angiography (SCT) were retrospectively analyzed. RNC and SCT series were acquired including VNC and monoenergetic images at 40, 70, and 140 keV. CT numbers were measured in regions of interest (ROIs) in different organs and in the bleeding jet for quantitative image analysis (contrast-to-noise ratios [CNR] and signal-to-noise ratio [SNR]). Additionally, 2 radiologists rated detectability of the bleeding source in the different CT series. Wilcoxon rank test for related samples was used.VNC series suppressed iodine sufficiently but not completely (CT number of aorta: RNC: 33.3±12.3, VNC: 44.8 ± 9.5, P = .01; bleeding jet: RNC: 43.1 ± 16.9, VNC: 56.3 ± 16.7, P = .02). VNC showed significantly higher signal-to-noise ratios than RNC for all regions investigated. Contrast-to-noise ratios in the bleeding jet were significantly higher in 40 keV images than in standard 140 keV images. The 40 keV images were also assigned the best subjective ratings for bleeding source detection.VNC can safely replace RNC in a CT protocol used to search for bleeding sources, thereby reducing radiation exposure by 30%. Low-keV series may enhance diagnostic confidence in the detection of bleeding sources.

MeSH terms

  • Aged
  • Female
  • Hemorrhage / diagnosis*
  • Hemorrhage / diagnostic imaging
  • Humans
  • Male
  • Middle Aged
  • Radiation Exposure / classification
  • Radiographic Image Interpretation, Computer-Assisted / methods
  • Retrospective Studies
  • Single Photon Emission Computed Tomography Computed Tomography / methods*
  • Single Photon Emission Computed Tomography Computed Tomography / standards