Accuracy of thoracic aortic measurements assessed by contrast enhanced and unenhanced magnetic resonance imaging

Abstract

To assess, whether unenhanced balanced steady-state-free precession sequences provide axial thoracic aortic measurements comparable to contrast enhanced magnetic resonance angiography with good intra- and interobserver agreement. Enhanced and unenhanced sequences of the thoracic aorta in 23 consecutive patients were evaluated. Axial thoracic aortic diameters were measured at predefined levels by two independent readers. Pearson's correlation coefficient and Bland-Altman analysis were used to compare enhanced and unenhanced sequences. t-Test was used to determine possible significant differences between the measurements obtained by enhanced and unenhanced sequences. A p-value of less than .05 indicated statistical significance. Intraclass correlation coefficient and Bland-Altman were used for inter- and intraobserver correlation and agreement. There was no significant difference in diameter measurements between enhanced and unenhanced sequences (ascending aorta, p=0.98; descending aorta, p=0.52). Bland-Altman revealed good agreement between enhanced and unenhanced sequences for ascending (mean bias, -0.01cm; with 95% limits of agreement, ±0.30cm) and descending aortic diameters (mean bias, 0.05cm; with 95% limits of agreement, ±0.30cm). Inter- and intraobserver agreement (mean bias, less than ±0.15; with 95% limits of agreement, less than ±0.42cm for all measurements) as well as correlation (r>0.8 for all measurements) were excellent. Unenhanced balanced steady-state-free precession sequences enable rapid and accurate determination of axial thoracic aortic diameters with excellent inter- and intraobserver agreement, but without the risk of contrast media associated side-effects.