Assessment of mediastinal widening associated with traumatic rupture of the aorta

In order to best determine the reliability and usefulness of widening of the mediastinum (WMED) and other radiographic abnormalities in the selection of trauma patients for aortography to detect traumatic rupture of the aorta (TRA), we designed a blind study in which a panel of radiologists and surgeons reviewed 149 chest films of trauma victims who subsequently underwent aortography to rule out TRA. Sixteen patients had TRA. Panelists identified mediastinal widening (WMED) in 83 of 93 observations on films in cases of TRA (89%). There was a significant association between WMED and TRA found both for the panel as a whole and for each panelist individually (p = 0.0000), making this an extremely reliable sign both in terms of detectability and in signalling the need for aortography. Significant associations with TRA were also found overall for six other radiographic abnormalities but none of these was reliable for all panelists or was as sensitive as WMED in the detection of TRA. Despite the reliability of these signs, panelists making decisions based on the chest film alone failed to recommend aortography in seven of 93 instances of TRA. Direct mediastinal measurements varied by at least 2 cm among panelists in one half of the cases of TRA, and 25% of these (4/16) had at least two measurements of mediastinal width of 7 cm or less. Of all the radiographic signs associated with TRA, widening of the mediastinum is the most reliable, but in this study all observers would not have detected all cases of ruptured aorta using radiographic signs alone. Clinical judgment and consideration of the forces involved in the injury must continue to play an important role in the selection of patients for aortography.