Objectives: Missed pancreatic injury carries significant morbidity. Computerized tomography (CT) imaging is useful, but may lack sensitivity to identify pancreatic injury. New-generation CT scanners should improve sensitivity, but this has not been studied. A previous study published in 2002 evaluating the sensitivity for identifying pancreatic injury with single-slice CT scanners yielded a 68% correlation between operative and CT findings. We aim to study the accuracy of modern CT for diagnosis and grading of pancreatic injury.
Methods: All trauma admissions from 2008 and 2012 were retrospectively reviewed. Patients with a pancreatic injury, either on CT or intraoperatively, were included (n = 96). Sensitivity and specificity were calculated using Student's t test.
Results: 48 patients had injuries noted on CT and in the operating room. In this group, 68.8% had CT findings discordant with operative findings. Of these, 78.8% had no injury noted on CT, of which 26.9% required surgical intervention. Seven patients with injury on CT had none identified in the operating room. Based on these results, the sensitivity for CT imaging to identify an injury is 36.4% with a positive predictive value of 68.2%.
Conclusions: Our results indicate that despite advances in CT technology, the sensitivity and specificity for identifying pancreatic injury remains low. Although CT scans remain critical in trauma evaluation, awareness of this diagnostic gap is important. Further analysis is required to determine any impact on patient outcomes.
Reprint & Copyright © 2017 Association of Military Surgeons of the U.S.