ARRS 2022 Abstracts

RETURN TO ABSTRACT LISTING


1657. Comparing DECT Iodine Overlay and Virtual Non-Contrast Reconstructions to Conventional CT in Evaluation for Active Gastrointestinal Bleeding
Authors * Denotes Presenting Author
  1. Mark Kovacs; Medical University of South Carolina
  2. Megan Mercer *; Medical University of South Carolina
  3. Jeffrey Waltz; Medical University of South Carolina
  4. Clark Phelps; Medical University of South Carolina
  5. Brian Flemming; Medical University of South Carolina
  6. Nicholas Shaheen; Medical University of South Carolina
  7. Andrew Hardie; Medical University of South Carolina
Objective:
This study aims to assess the potential of using the iodine overlay (IO) and virtual non-contrast (VNC) image sets generated as standard reconstructions from dual energy CT (DECT) data alone for rapid assessment of active gastrointestinal bleeding (AGIB).

Materials and Methods:
Thirty subjects (15 with active GI bleed) were retrospectively enrolled having undergone two phase DECT (precontrast [WO]) and arterial phase postcontrast [ART]) with IO/VNC image sets reconstructed from the arterial phase dual energy acquisition. For each subject, the conventional CT images (ART/WO) and DECT IO/VNC image sets were separately assessed by five readers with different levels of experience from a single institution using a single-blinded, crossover design. Statistical endpoints were assessed for diagnosis of AGIB, time to interpret (seconds), and confidence in diagnosis (1 - 5 scale).

Results:
There was no significant difference in the diagnosis of AGIB between the conventional CT and IO/VNC for any reader nor for pooled data. There was a reduced time to interpret with IO/VNC for pooled data, although this was only significant for the cases without AGIB. There was also higher recorded confidence for IO/VNC but this was not statistically significant for pooled data.

Conclusion:
DECT IO and VNC reconstructions can be interpreted alone in evaluation for AGIB with a reduced time to interpret, high confidence, and without loss of diagnostic performance.