CT esophagogram in the emergency setting: typical findings and suggested workflow.

Abstract

Purpose: Esophageal perforation (EP) can be a diagnostic challenge. Computed tomography (CT) and CT esophagography (CTE) are often used to rule out EP in the emergency setting with promising diagnostic performance, but the standard of care remains fluoroscopic esophagography (FE). We assess the diagnostic performance of CT and CTE when interpreted by expert and generalist radiologists and created an imaging workflow guide.

Methods: Retrospective study of patients presenting with suspected EP. Two expert radiologists independently reviewed blinded CT/CTE studies, recorded CT findings, and assigned an esophageal injury grade. We also collected initial (general radiologist) CT findings and interpretation and FE diagnoses. We assessed inter-reader reliability and diagnostic performance.

Results: EP was diagnosed in 46/139 (33%) encounters. The most common CT/CTE findings in EP were esophageal wall thickening (46/46, 100%), pneumomediastinum (42/46, 91%), and mediastinal stranding (39/46, 85%). CT and CTE sensitivity for detecting EP was 89% and 89% for expert radiologists, respectively, and 79% and 82% for general radiologists, compared with 46% for FE. Inter-reader agreement for detecting EP by CT and CTE was kappa 0.35 and 0.42 (both p < .001) between expert and generalist radiologists. We present radiographic images for key CT/CTE findings and a suggested workflow for the evaluation of possible EP.

Conclusion: CT and CTE are more sensitive than FE for EP in the emergency setting. Due to the rarity of EP and current wide variability in imaging interpretation, an imaging workflow and injury grading system based on esophageal and mediastinal CT findings are offered to help guide management.