Robert G. Hayter, BS ; James T. Rhea, MD ; Andrew Small, MD ; Faranak S. Tafazoli, MD ; Robert A. Novelline, MD
1 From the Department of Radiology, Massachusetts General Hospital and Harvard Medical School, 107 Avenue Louis Pasteur, Room 185, Boston, MA 02115. Received September 3, 2004; revision requested November 3; revision received March 11, 2005; accepted April 15; final version accepted July 27.
PURPOSE: To retrospectively review the authors’ experience with multi–detector row computed tomography (CT) for detection of aortic dissection in the emergency setting.
MATERIALS AND METHODS: The investigation was institutional review board approved, did not require informed patient consent, and was HIPAA compliant. In 373 clinical evaluations in the emergency setting, 365 patients suspected of having aortic dissection and/or other aortic disorders underwent multidetector CT. Criteria for acute aortic disorder were confirmed by using surgical and pathologic diagnoses or findings at clinical follow-up and any subsequent imaging as the reference standard. Positive cases were characterized according to type of disorder interpreted. Resulting sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated by using two-way contingency tables. All cases found to be negative for acute aortic disorders were grouped according to alternative CT findings.
RESULTS: Sixty-seven (18.0%) of the 373 cases were interpreted as positive for acute aortic disorder. One hundred twelve acute aortic disorders were identified in these 67 cases: 23 acute aortic dissections, 14 acute aortic intramural hematomas, 20 acute penetrating aortic ulcers, 44 new or enlarging aortic aneurysms, and 11 acute aortic ruptures. Three hundred five (81.8%) cases were interpreted as negative for acute aortic disorder. In 48 negative cases, multidetector CT depicted alternative findings that accounted for the clinical presentation. Of these, three included both acute aortic disorders and alternative findings, and 45 included only alternative findings. One (0.3%) case was indeterminate for acute aortic disorder. Overall, 112 findings were interpreted as positive for acute aortic disorder, an alternative finding, or both at CT. No interpretations were false-positive, one was false-negative, 67 were true-positive, and 304 were true-negative. Sensitivity, specificity, PPV, NPV, and accuracy were 99% (67 of 68), 100% (304 of 304), 100% (67 of 67), 99.7% (304 of 305), and 99.5% (371 of 373), respectively.
CONCLUSION: The positivity rate for acute aortic dissection or other acute aortic disorder in 373 cases examined at multi–detector row CT was 18.0%.