Justin E Vranic1, Thien J Huynh2, Peter Fata2, Jason Barber3, Robert H Bonow3, Michael R Levitt4, Nancy Temkin3, Ryan Morton3, Cory Kelly5, Dean K Shibata2, Mahmud Mossa-Basha2. 1. University of Washington Department of Radiology, United States. Electronic address: justedv@uw.edu. 2. University of Washington Department of Radiology, United States. 3. University of Washington Department of Neurological Surgery, Seattle, WA, United States. 4. University of Washington Department of Radiology, United States; University of Washington Department of Neurological Surgery, Seattle, WA, United States; University of Washington Department of Mechanical Engineering, Seattle, WA, United States; University of Washington Stroke and Applied Neuroscience Center, Seattle, WA, United States. 5. University of Washington Department of Neurological Surgery, Seattle, WA, United States; University of Washington Stroke and Applied Neuroscience Center, Seattle, WA, United States.
Abstract
BACKGROUND AND PURPOSE: Blunt cerebrovascular injury (BCVI) is associated with a significant risk of ischemic stroke when left untreated. Cross-sectional imaging is vital to early BCVI diagnosis and treatment; however, conventional luminal vessel imaging is limited in its ability to evaluate for vessel wall pathology. The purpose of this study is to evaluate the ability of vessel wall magnetic resonance imaging (VWI) to detect and evaluate BCVI in acutely injured trauma patients relative to neck computed tomographic angiography (CTA). MATERIALS AND METHODS: Trauma patients with suspected BCVI on initial neck CTA were prospectively recruited for VWI evaluation. Two neuroradiologists blinded to patient clinical history and CTA findings evaluated each artery independently on VWI and noted the presence and grade of BCVI. These results were subsequently compared to neck CTA findings relative to expert clinical consensus review. Interrater reliability of VWI for detecting BCVI was evaluated using a weighted Cohen κ-statistic. RESULTS: Ten trauma patients (40 cervical arteries) were prospectively evaluated using both CTA and VWI. Out of 18 vascular lesions identified as suspicious for BCVI on CTA, six lesions were determined to represent true BCVI by expert consensus review. There was almost perfect agreement between VWI and expert consensus regarding the presence and grade of BCVI (κ=0.82). This agreement increased when considering only low grade BCVI. There was only fair agreement between CTA and expert clinical consensus (κ=0.36). This agreement decreased when considering only low grade BCVI. CONCLUSIONS: VWI can potentially accurately identify and evaluate BCVI in acutely injured trauma patients with excellent inter-rater reliability.
BACKGROUND AND PURPOSE:Blunt cerebrovascular injury (BCVI) is associated with a significant risk of ischemic stroke when left untreated. Cross-sectional imaging is vital to early BCVI diagnosis and treatment; however, conventional luminal vessel imaging is limited in its ability to evaluate for vessel wall pathology. The purpose of this study is to evaluate the ability of vessel wall magnetic resonance imaging (VWI) to detect and evaluate BCVI in acutely injured traumapatients relative to neck computed tomographic angiography (CTA). MATERIALS AND METHODS:Traumapatients with suspected BCVI on initial neck CTA were prospectively recruited for VWI evaluation. Two neuroradiologists blinded to patient clinical history and CTA findings evaluated each artery independently on VWI and noted the presence and grade of BCVI. These results were subsequently compared to neck CTA findings relative to expert clinical consensus review. Interrater reliability of VWI for detecting BCVI was evaluated using a weighted Cohen κ-statistic. RESULTS: Ten traumapatients (40 cervical arteries) were prospectively evaluated using both CTA and VWI. Out of 18 vascular lesions identified as suspicious for BCVI on CTA, six lesions were determined to represent true BCVI by expert consensus review. There was almost perfect agreement between VWI and expert consensus regarding the presence and grade of BCVI (κ=0.82). This agreement increased when considering only low grade BCVI. There was only fair agreement between CTA and expert clinical consensus (κ=0.36). This agreement decreased when considering only low grade BCVI. CONCLUSIONS: VWI can potentially accurately identify and evaluate BCVI in acutely injured traumapatients with excellent inter-rater reliability.