Surgeon interpretation of intravascular ultrasound is insufficient for aortic injury detection during uncontrolled hemorrhage.

Abstract

Noncompressible torso hemorrhage (NCTH) is the leading cause of potentially preventable death in modern combat. Intravascular ultrasound (IVUS) offers real-time, intraluminal visualization that could enable targeted endovascular hemorrhage control, but its effectiveness depends on the operator's ability to correctly interpret IVUS images during active bleeding. Surgeon diagnostic performance in this context is unknown. We conducted a diagnostic performance-based study using IVUS clips derived from a validated swine model of uncontrolled supraceliac aortic hemorrhage. Board-certified vascular surgeons independently reviewed 31 randomized IVUS clips (14 no-injury; 17 true injury) and identified (1) whether an injury was present, (2) the correct circumferential (clock-face) position, and (3) the correct timestamp in the pullback video sequence. "Presence accuracy" reflected the correct identification of injury presence; "composite accuracy" required all three domains. We examined performance across surgeon characteristics, IVUS wire platforms, perceived image quality, and IVUS use frequency. Mixed-effects logistic regression identified predictors of presence accuracy. Seven vascular surgeons provided 217 evaluations. Surgeons achieved a sensitivity of 71.4% and specificity of 43.9%. Presence and composite accuracies were 59.0% and 52.5%, respectively. False positive counts exceeded false negatives (55 vs. 34; p<0.001), but 28.6% of true injuries were missed. Diagnostic accuracy improved with perceived image quality (ρ=0.19-0.42, all p<0.01) and with the 0.018-inch versus 0.035-inch wire platform. In adjusted analysis, true injury presence (aOR=2.41, p=0.004) and surgeon confidence (aOR=1.78, p=0.039) independently predicted accuracy; perceived image quality and wire type did not. Vascular surgeons demonstrated limited accuracy in detecting and localizing aortic injuries using IVUS alone. These findings highlight important limitations of unaided IVUS interpretation during active hemorrhage and underscore the need for researching additional strategies to improve real-time injury localization in endovascular trauma care (J Trauma Acute Care Surg. 2026;00:00-00. Copyright © 2026 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Association for the Surgery of Trauma). Omitted (animal model). Diagnostic performance study.