Image analysis in time-resolved large field of view 3D MR-angiography at 3T

Alex Frydrychowicz*, Thorsten A. Bley, Zahra Abdullah Zadeh, Andreas Harloff, Jan T. Winterer, Jürgen Hennig, Mathias Langer, Michael Markl

*Corresponding author for this work
14 Citations (Scopus)


Purpose: To qualitatively and quantitatively evaluate the image quality in accelerated time-resolved 3D contrast-enhanced MR angiography (tr-CE-MRA) at 3T. Materials and Methods: In all, 113 MRA were performed in 107 patients on a 3T MR system after written informed consent and approval by the ethics committee. Twenty consecutive thoracic (n = 87) or craniocervical (n = 26) 3D data volumes were acquired. The timeframes with maximum arterial and venous contrast were determined and a total of 663 arterial and venous segments were analyzed by two blinded observers. Diagnostic image quality was graded by applying a 0 (low) to 3 (excellent) scale. Additionally, local signal-to-noise (SNR) and contrast-to-noise ratios (relative CNR) were evaluated. Results: Tr-CE-MRA was successfully performed in all patients. Good to excellent image quality (2.42 ± 0.31) was observed in all individuals with preserved discrimination of arteries (2.43 ± 0.48) and veins (2.20 ± 0.56). Minor image degradation due to artifacts (2.62 ± 0.25) and constantly high vascular signal and contrast were detected. There was a significant superiority of coronal orientation during thoracic MRA (P < 0.05). In 18 cases tr-CE-MRA provided additional information on vascular pathologies. Conclusion: Large field of view tr-CE-MRA enables constantly high-quality thoracic and craniocervical angiographies. In addition, the dynamics of tr-CE-MRA can offer additional information on vascular pathologies.

Original languageEnglish
JournalJournal of Magnetic Resonance Imaging
Issue number5
Pages (from-to)1116-1124
Number of pages9
Publication statusPublished - 01.11.2008


Dive into the research topics of 'Image analysis in time-resolved large field of view 3D MR-angiography at 3T'. Together they form a unique fingerprint.

Cite this