A morphological indicator for aortic dissection: fitting circle of the thoracic aorta

Abstract

Abstract Background This study aims to identify a morphological indicator of aortic dissection (AD) based on the geometrical characteristics of the thoracic aorta. Methods We evaluated computed tomographic angiograms of 63 samples with AD (22 with type A AD, 41 with type B AD) and 71 healthy samples. Via centerline extraction and spatial transformation, the spatial entanglement of the aorta was minimized, and the expanded 2D aortic morphology was obtained. The 2D morphology of the thoracic aorta was fit to a circle. The applicability of the fitting circle method for identifying aortic dissection was verified by multivariable logistic regression analysis. Results Via the 3D coordinate transformation algorithm, the optimal aortic view was obtained. On this view, the geometrical characteristics of the thoracic aortas of the healthy controls were similar to a portion of a circle (sum of residuals: 3502.45 ± 2566.71, variance: 86.23 ± 56.60), while that of AD samples had poorer similarity to the circle (sum of residuals: 5404.78 ± 3891.69, variance: 129.90 ± 90.09). This difference was significant (p < 0.001). A logistic regression model showed that increased deformation of the thoracic aorta was a significant indicator of aortic dissection (odds ratio: 1.35, p = 0.034). Conclusions The morphology of the healthy thoracic aorta could be fit to a circle, while that of the dissected aorta had poorer similarity to the circle. The statistics of the circle are an effective indicator of aortic deformation in AD. Trial registration This study is registered in the Chinese Clinical Trial Registry (ChiCTR2000029219).

Keywords

• Aortic dissection
• Computed tomographic angiography
• Geometrical analysis
• Diagnostic imaging
• Algorithm