Diagnostics (Oct 2022)

Dual-Energy CT for Accurate Discrimination of Intraperitoneal Hematoma and Intestinal Structures

  • Moritz T. Winkelmann,
  • Florian Hagen,
  • Kerstin Artzner,
  • Malte N. Bongers,
  • Christoph Artzner

DOI
https://doi.org/10.3390/diagnostics12102542
Journal volume & issue
Vol. 12, no. 10
p. 2542

Abstract

Read online

The purpose of this study was to evaluate the potential of dual-energy CT (DECT) with virtual unenhanced imaging (VNC) and iodine maps (IM) to differentiate between intraperitoneal hematomas (IH) and bowel structures (BS) compared to linearly blended DECT (DE-LB) images (equivalent to single-energy CT). This retrospective study included the DECT of 30 patients (mean age: 64.5 ± 15.1 years, 19 men) with intraperitoneal hematomas and 30 negative controls. VNC, IM, and DE-LB were calculated. Imaging follow-up and surgical reports were used as references. Three readers assessed diagnostic performance and confidence in distinguishing IH and BS for DE-LB, VNC, and IM. Diagnostic confidence was assessed on a five-point Likert scale. The mean values of VNC, IM, and DE-LB were compared with nonparametric tests. Diagnostic accuracy was assessed by calculating receiver operating characteristics (ROC). The results are reported as medians with interquartile ranges. Subjective image analysis showed higher diagnostic performance (sensitivity: 96.7–100% vs. 88.2–96.7%; specificity: 100% vs. 96.7–100%; p p p ≤ 0.0001). VNC analysis revealed a significantly higher attenuation of hematomas (50.5 HU; IQR [44.4, 59.4]) than BS (26.6 HU; IQR [22.8, 32.4]; p ≤ 0.0001). DE-LB revealed no significant differences between hematomas (60.5 HU, IQR [52.7, 63.9]) and BS (63.9 HU, IQR [58.0, 68.8]; p > 0.05). ROC analysis revealed the highest AUC values and sensitivity for IM (AUC = 100%; threshold by Youden-Index ≤ 19 HU) and VNC (0.93; ≥34.1 HU) compared to DE-LB (0.64; ≤63.8; p < 0.001). DECT is suitable for accurate discrimination between IH and BS by calculating iodine maps and VNC images.

Keywords