- Poster presentation
- Open Access
Learning-based super-resolution technique significantly improves detection of coronary artery stenoses on 1.5T whole-heart coronary MRA
https://doi.org/10.1186/1532-429X-16-S1-P218
© Ishida et al.; licensee BioMed Central Ltd. 2014
- Published: 16 January 2014
Keywords
- Diagnostic Performance
- Suspected Coronary Artery Disease
- Coronary Artery Stenos
- Lower Image Quality
- Interpretation Time
Background
Example of conventional whole-heart coronary MRA (A) and high-resolution whole-heart coronary MRA with super-resolution technique (B) in the same subject.
Methods
Forty-six patients with suspected coronary artery disease (CAD) underwent X-ray coronary angiography and whole-heart coronary MRA with a 1.5T MR scanner and 32 channel coils. Image quality was assessed using 5-point scale (1 = not visible, 2 = poor, 3 = moderate, 4 = good, 5 = excellent) for conventional coronary MRA. High-resolution coronary MRA was generated by using the SR technique. Three observers independently rated the confidence level of the presence of stenosis in each coronary segment with a continuous scale from 0 to 1 by using a sliding SLAB MIP method. Receiver operating characteristic (ROC) analysis was employed to evaluate the diagnostic performance of high-resolution coronary MRA generated by SR technique in comparison with coronary MRA without SR technique in detecting luminal narrowing of >50% on X-ray angiography.
Results
Comparison of ROC curves for the average performance of the three observers in detection of coronary artery stenoses on high-resolution coronary MRA generated by SR technique and conventional coronary MRA. With super-resolution technique, the average AUC was significantly improved from 0.792 to 0.840 (p = .020)
Conclusions
High-resolution whole-heart coronary MRA generated by the SR technique allows for more accurate detection of coronary stenoses with reduced interpretation time as compared to conventional whole-heart coronary MRA.
Funding
n/a.
Authors’ Affiliations
Copyright
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.