Sources of error in the measurement of aortic diameter in computed tomography scans

Abstract

This study was conducted to determine the differences in the diameter of the thoracic aorta when measured from electrocardiographic (ECG)-gated and nongated computed tomography (CT) angiography. Another aim was to define the difference in the aortic diameter when it is measured at peak systole and end diastole in ECG-gated scans. The gated and nongated CT angiograms of 27 patients (mean age, 58 ± 16 standard deviation [SD] years) obtained on a 256-slice multidetector CT scanner were used. The transverse and anteroposterior diameters and the lumen areas were measured at 1, 4, and 8 cm below the origin of the left subclavian artery. There was a significant difference in the aortic measurements of diameter between gated and nongated scans found in samples taken at 1, 4, and 8 cm distal to the left subclavian artery (P < .0001). We found a considerable difference between the systolic and diastolic diameters (P < .0001). The maximum change in diameter between systole and diastole was 2.9 ± 0.9 (SD) mm (14.5%, P < .0001) at 1 cm, 5.4 mm (22.6%; median, 1.7 mm; P < .0001) at 4 cm, and 4.4 mm (16.9%; median, 1.3 mm; P < .0001) at 8 cm. There was a significant difference between the transverse and anteroposterior diameters in systole and diastole at all locations (P < .0001): The maximum change in diameter between transverse and anteroposterior diameters in systole was 5.4 ± 1.1 (SD) mm (15.7%, P < .0001) at 1 cm, 5.8 mm (19%; median, 1.4 mm; P < .0001) at 4 cm, and 5 mm (15%; median, 1.02 mm; P < .0001) at 8 cm. There was also a substantial difference between measuring the transverse diameter directly and deriving it from the lumen area (P < .0001). Our results showed an important difference between systolic and diastolic diameters measurements in ECG-gated scans. The standard protocol for measuring aortic diameters in gated scans of the thoracic aorta uses images at end diastole because the lack of wall motion at this time provides better resolution. This is likely to result in undersizing that, in some instances, may threaten stability and the proper seal of the stent graft. The dimensions of the aorta in a gated CT should be measured at peak systole rather than the conventional end diastole used today. Most medical centers use nongated CT or gated CT scans in end diastole to calculate sizes of endografts. In view of our findings, the latter method could result in potential complications.