Comparison of computer-aided quantitative measurement and physician visual assessment in the evaluation of intracranial atherosclerotic stenosis: a vessel wall magnetic resonance imaging study.

Abstract

Background: Intracranial atherosclerotic stenosis is a leading cause of ischemic stroke in China. Accurate assessment of intracranial atherosclerotic stenosis through imaging techniques is crucial for guiding therapeutic interventions and prognostic stratification. Vessel wall magnetic resonance imaging (VWMRI) has emerged as a reliable method for evaluating intracranial arterial vessels. With the advancement of technology, computer-aided quantitative measurement (CAQM) is increasingly used in imaging assessment. This study aimed to compare physician visual assessment (PVA) with CAQM in the VWMRI evaluation of intracranial atherosclerotic stenosis.

Methods: This retrospective cross-sectional study consecutively enrolled patients diagnosed with intracranial atherosclerotic stenosis through imaging examinations at the Fourth Affiliated Hospital of China Medical University from December 2018 to December 2023. Clinical data were collected for analysis. Two radiologists independently and separately conducted CAQM and PVA on the VWMRI images of intracranial atherosclerotic stenosis patients. The imaging features evaluated encompassed stenosis severity, vessel wall remodeling, vessel wall thickening patterns, fibrous cap characteristics, lipid core ratio, and plaque enhancement degree. The study further assessed the discrepancies and concordance between the assessment results obtained from the two methods using paired sample t-tests, Wilcoxon signed-rank tests, and Cohen's kappa coefficient analysis.

Results: This study enrolled a total of 589 patients. The PVA time was shorter than CAQM (12.02±3.63 vs. 20.48±6.50 min). However, compared with digital subtraction angiography, the CAQM had a better area under the curve (0.88) than the PVA (0.80) in assessing luminal stenosis degree. The proportions of vessel wall remodeling (227/38.5%) and plaque surface irregularity (127/21.6%) evaluated by PVA were both lower than those by CAQM (438/74.4%, 171/29.0%). Meanwhile, no statistically significant differences were found in the patterns of wall thickening (P=0.12/0.39) and the proportion of plaque lipid core (P=0.65 and P=0.27), with good agreement between the two methods (K=0.67/0.85, K=0.97/0.94). While there were no statistical differences in the assessment of plaque enhancement degree in specific arteries (middle cerebral artery and basilar artery) (n=77/36, P=0.08/0.21), an overall statistical difference was observed (n=113, P=0.03). Additionally, there was poor agreement in assessing plaque enhancement degree, with Cohen's kappa values of 0.13 (-0.05 to 0.32) and 0.16 (-0.06 to 0.39).

Conclusions: This study revealed disparities between PVA and CAQM in the evaluation of intracranial atherosclerotic stenosis of VWMRI. CAQM is recommended for assessing stenosis degree, vessel wall remodeling, and fibrous cap characteristics. However, PVA is suggested to assess wall thickening patterns and lipid core ratio to expedite diagnosis. Further research is needed to validate CAQM's superiority in evaluating plaque enhancement degrees.