Meeting Banner
Abstract #1170

On the Generalization of Carotid Vessel Wall MRI Risk Factors Across Imaging Centers

Kiyofumi Yamada1, Martine Truijman2, William Kerwin1, Jie Sun1, Yan Song3, Mat Daemen2, Eline Kooi2, Chun Yuan1

1Radiology, University of Washington, Seattle, WA, United States; 2Maastricht University Medical Center, Maastricht, Netherlands; 3Radiology, Beijing Hospital, Beijing, China

The purpose of this investigation was to understand the cross-center similarities and differences of two independent carotid vessel wall imaging centers. Ten subjects were recruited from each of two imaging centers for multicontrast MRI and plaque compositional analysis. We found good association across centers suggesting similar results regarding the association of risk factors with outcome can be expected across centers. On the other hand, considerable biases between absolute values of quantitative parameters were observed. Thus, in order for risk factors to be translated into clinical practice with appropriate risk thresholds, the source of bias must be ascertained and eliminated.

Keywords

ability absence absolute according addition agent agreement appropriate ascertained assessments association best bias biases boundaries calcifications carotid categorical china clinical collecting combine come commercial comparative compatibility components compositional conclude considerable consistent continuous contrast core criteria cross customized dependence dependent differing drawn eliminated employed evaluations expected fibrous findings general generalization generalize greater hand hemorrhage hospital illustrates include independent independently indicating injection interactively intercepts investigate investigation local materials matrix maximal medical moderate must necrotic noted outcome outlined overall overview paired people percentage plaque populations practice predicting presence prospective protocol protocols proving quantitative quantities radiology recruited regarding regression relevant republic review reviewed reviewer reviewers risk rupture scanners significantly similarities similarity simple site software song source statistical status stratification stroke strongly studies subjects suggested symptomatic table thinning thresholds translated typical uncertain understand variable variables various vessel volume wall whereas yuan