TOPLINE:

Dual-energy CT demonstrated non-inferior diagnostic performance with an increased sensitivity and specificity compared to conventional triphasic CT in the detection of gastrointestinal bleeding and reduced radiation exposure by 20.4% through the elimination of true non-contrast imaging, according to a retrospective study.

METHODOLOGY:

• Researchers conducted a retrospective single-centre diagnostic accuracy study including 100 patients (mean age, 70.3 years; 34 women) who underwent triphasic abdominal CT for gastrointestinal bleeding at University Hospital Basel, Basel, Switzerland, between September 2015 and June 2024.
• For the analysis, 50 patients with acute gastrointestinal bleeding were age and sex matched with 50 control individuals without radiologic suspicion of the bleeding.
• Five independent readers, including three fellowship-trained abdominal radiologists (6-9 years of experience) and two residents in their second and third year of training, respectively, assessed conventional and dual-energy CT images.
• Diagnostic performance was evaluated on the basis of sensitivity, specificity, reading time, and confidence levels, with the reference standard determined through a consensus review and endoscopic correlation in select cases.

TAKEAWAY:

• Conventional triphasic CT demonstrated a sensitivity of 91.6% (95% CI, 87.5%-94.7%) and a specificity of 94.4% (95% CI, 90.8%-96.9%) in detecting gastrointestinal bleeding.
• Dual-energy CT showed non-inferior performance, achieving a 3% non-inferiority margin with a sensitivity of 94.4% (95% CI, 90.8%-96.9%) and a specificity of 96.0% (95% CI, 92.8%-98.1%). Overall diagnostic accuracy also increased significantly for dual-energy CT compared to conventional triphasic CT (95.2% vs 93.0%; P < .01).
• Median reader diagnostic confidence was significantly higher with dual-energy CT than with conventional triphasic CT (5 vs 4), exhibiting superiority and therefore non-inferiority (P < .01); however, the mean reading time decreased from 96 to 94 seconds and was non-inferior with the prespecified 3-second margin (P < .01).
• The total dose-length product (radiation dose) was reduced by 20.4% through the omission of true non-contrast images.

IN PRACTICE:

"When available, dual-source DECT [dual-energy CT] should be used to replace true non-contrast images in triphasic CT protocols for suspected upper or lower GI [gastrointestinal] bleeding," the authors wrote. They added that "it increases diagnostic confidence while maintaining diagnostic accuracy without increasing reading time" and "patients may benefit from lower radiation doses."

SOURCE:

This study was led by Moritz Oberparleiter, Department of Radiology, University Hospital Basel, University of Basel, Basel, Switzerland. It was published online on December 17, 2025, in European Radiology.

LIMITATIONS:

The case-control design of this study prevented the reliable estimation of positive and negative predictive values. The reference standard was primarily based on imaging data rather than endoscopy to avoid verification bias. All scans were performed on a dual-source dual-energy CT platform, potentially limiting the generalisability to other dual-energy CT technologies. Additionally, readers may have identified bleeding in incorrect locations compared to the reference standard; however, this error rate was expected to be low given the comparable false-positive rates between protocols.

DISCLOSURES:

Open access funding was provided by the University of Basel. One author reported serving as a member of the scientific editorial board of European Radiology.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.