Voxel-Wise Quantification and Multi-planar Visualization of Plaque Contrast Enhancement in Intracranial Vessel Wall MRI.

Background: Contrast enhancement (CE) of intracranial atherosclerotic plaques is a valuable biomarker for identifying culprit plaques in ischemic stroke, but current assessment methods are limited by qualitative approaches and artifact susceptibility. Purpose: To develop and validate a semi-automated method for visualization and quantification of CE intensity and volume, and evaluate its accuracy and scan-rescan reproducibility. Study type: Retrospective. Population: Two patient groups with intracranial atherosclerotic diseases were included: Group A (n=37, 82 plaques) for developing and validating the CE map against expert review, and Group B (n=11, 23 plaques) for assessing scan-rescan reproducibility. Field Strength/Sequence: 3.0 T, 3D time-of-flight gradient echo sequence and T1-weighted fast spin echo sequences, before and after Gadolinium contrast injection. Assessment: A 3D quantitative CE map was developed that incorporated dedicated multicontrast, multi-planar preprocessing, image intensity normalization, adjustment for enhancement artifacts, and derivation of a signal intensity threshold. Statistical Tests: Receiver operating characteristic curve analysis determined the optimal CE threshold. Sensitivity, specificity, and Cohen's kappa assessed agreement between the quantitative CE map and manual review. Spearman's correlation evaluated CE volume quantification. Intraclass correlation coefficient (ICC) and coefficient of variance (CV) assessed reproducibility. Results: The optimal signal intensity threshold for post-contrast maps was 0.4 plus the median plaque wall intensity on pre-contrast maps. The area under the curve for CE detection was 0.91 (95% CI: 0.87-0.94), with sensitivity and specificity of 0.83 and 0.90. CE volume measurements strongly correlated with expert measurements (Spearman's rho=0.82, p<0.001). Enhancement ratio showed high reproducibility (ICC=0.92, 95% CI: 0.82-0.96), and CE detection demonstrated robust agreement (kappa=0.82, 95% CI: 0.55-1). Data Conclusion: The proposed CE map demonstrated good accuracy compared to expert review and high scan-rescan reproducibility, enabling comprehensive assessment of CE presence, intensity, and volume for evaluating plaque vulnerability in intracranial atherosclerotic disease.