3D Mapping of Vessel Wall Enhancement could Assist in Robust Risk Stratification of Intracranial Aneurysms

Abstract

Vessel Wall Enhancement (VWE) has emerged as a potential tool to aid clinicians in risk stratification of intracranial aneurysms (IAs). However, this is currently graded manually which introduces subjectivity. Herein, we evaluated the inter-user variability of clinicians in grading VWE manually and used an existing pipeline to derive quantitative first order metrics. These metrics were then used to build statistical models for more objective VWE quantification and characterization. We observed that clinicians agree on the presence of VWE in 75% of the cases but only on 54% of the cases for the type of VWE and this agreement decreases in smaller IAs. Through our automated pipeline, we mapped the VWE intensity on to the sac of the IA and computed 10 different first order metrics. We found that 8 of these 10 metrics were significantly different between IAs exhibiting VWE and IAs without VWE. Additionally, we found that statistical models built using these metrics have a good performance in predicting the presence of VWE (AUC=0.94) and the type of VWE (AUC=0.78). This pipeline can be used as a tool for more objective quantification and characterization of VWE.