Evaluation of cerebral aneurysm wall thickness in experimental aneurysms: Comparison of 3T-MR imaging with direct microscopic measurements

Abstract

Thin aneurysm wall thickness (AWT) is thought to portend an elevated risk of intracranial aneurysm rupture. Magnetic resonance imaging (MRI) is biased by AWT overestimations. Previously, this suspected bias has been qualitatively described but never quantified. We aimed to quantify the overestimation of AWT by MRI when compared to the gold standard of AWT as measured by light microscopy of fresh aneurysm specimens (without any embedding procedure). This analysis should help to define the clinical potential of MRI estimates of AWT. 3-Tesla (3T) MRI (contrast-enhanced T1 Flash sequences; resolution: 0.4 x 0.4 x 1.5 mm(3)) was performed in 13 experimental aneurysms. After MR acquisition, the aneurysms were retrieved, longitudinally sectioned and calibrated micrographs were obtained immediately. AWT at the dome, AWT at the neck and parent vessel wall thickness (PVT) were measured on precisely correlated MR-images and histologic micrographs by blinded independent investigators. Parameters were statistically compared (Wilcoxon test, Spearman's correlation). AWT was assessed and reliably measured using MRI. Interobserver variability was not significant for either method. MR overestimation was only significant below the image resolution threshold: AWT at the dome (0.24 ± 0.06 mm vs. MR 0.30 ± 0.08 mm; p = 0.0078; R = 0.6125), AWT at the neck (0.25 ± 0.07 mm vs. MR 0.29 ± 0.07 mm; p = 0.0469; R = 0.7451), PVT (0.46 ± 0.06 mm vs. MR 0.48 ± 0.06 mm; p = 0.5; R = 0.8568). In this experimental setting, MR overestimations were minimal (mean 0.02 mm) above the image resolution threshold. When AWT is classified in ranges defined by the MR resolution threshold, clinical usage may be beneficial. Further quantitative and comparative experimental and human studies are warranted to confirm these findings.