Categorization of aortic aneurysm thrombus morphology by magnetic resonance imaging

Background: Magnetic resonance imaging (MRI) has been proposed for qualitative categorization of intraluminal thrombus morphology. We aimed to correlate the qualitative MRI categorization previously described to quantitative measurements of signal intensity and to compare morphological characteristics of intraluminal thrombus specimens to the appearance on magnetic resonance imaging. Methods: Thirty-four patients undergoing open surgery for abdominal aortic aneurysm had a preoperative MRI obtained with a 1.5 T magnet. Qualitative categorization was performed (blinded and in consensus) and correlated to intraluminal thrombus to muscle signal-intensity ratios. Morphology of intraluminal thrombus specimens collected during surgery were compared to the magnetic resonance imaging categories and specimen weight was correlated to thrombus volume measured on preoperative computer tomography angiography. Results: Blinded MRI categorization resulted in agreement in 22 out of 34 intraluminal thrombi (Kappa value 0.3, p = 0.006). Medians (p = 0.004) and distribution (p = 0.002) of signal-intensity ratios varied significantly across the three MRI categories obtained by consensus. Heterogeneous and homogenous specimen appearance corresponded to similar appearances on MRI in 78% and 55% respectively, resulting in an overall Kappa = 0.4 (p = 0.04). Intraluminal thrombus volume and weight correlated well (r_s 0.831, p < 0.001) with a mean difference of 60 g (95% CI 38-80 g), without proportional bias. Conclusion: Qualitative evaluation of intraluminal thrombus morphology based on MRI can be quantified by measuring signal-intensity ratios. Concurrently a fair agreement to blinded qualitative evaluation of thrombus specimens can be obtained. However, the evaluation is impaired by loss of a large proportion of thrombus during sampling.