Article Text

Download PDFPDF

114 Detecting ischaemia in flow limiting multi-vessel disease – is 3d perfusion cmr where the money lies?
  1. Joy Shome1,
  2. Mr Kerem.C Tezcan2,
  3. Sohaib Nazir3,
  4. Adriana Villa3,
  5. David Snell3,
  6. Kamran Baig4,
  7. Simon Redwood5,
  8. Brian Clapp6,
  9. Antonis Pavlidis6,
  10. Reza Razavi3,
  11. Tevfik.F Ismail3,
  12. Divaka Perera5,
  13. Sebastian Kozerke2,
  14. Sven Plien3
  1. 1King’s College London, St. Thomas’ Hospital
  2. 2Institute for Biomedical Engineering, University and ETH Zurich
  3. 3Division of Imaging Sciences and Biomedical Engineering, The Rayne Institute, King’s College London, St. Thomas Hospital
  4. 4Department of Cardiothoracic Surgery, St. Thomas’ Hospital
  5. 5Cardiovascular Division, The Rayne Institute, King’s College London, St. Thomas’ Hospital
  6. 6Department of Cardiology, St. Thomas’ Hospital


Introduction Myocardial perfusion cardiovascular magnetic resonance (CMR) is a highly accurate non-invasive imaging modality in the diagnosis of coronary artery disease (CAD). High-resolution (hi-res) two-dimensional (2D) perfusion CMR can better detect sub-endocardial ischaemia although with the disadvantage of lesser myocardial coverage. Three-dimensional (3D) perfusion provides whole heart coverage but has a comparatively inferior resolution. The superiority of fractional flow reserve (FFR) over visual angiographic assessment in determining functional significance of a coronary stenosis is now well established. We studied the diagnostic agreement between hi-res 2D and 3D perfusion CMR in patients with significant multi-vessel flow limiting CAD as confirmed by FFR on a per patient and per vessel basis.

Methods Patients with suspected stable CAD referred for invasive coronary angiography as part of their routine clinical care were prospectively recruited. Prior to revascularisation (if performed) all patients underwent both hi-res 2D adenosine vasodilator stress and 3D adenosine stress perfusion scans during the same sitting. Visually, coronary stenoses less than 50% were deemed not to be flow limiting, whereas those 80% or more were considered as flow limiting. For stenoses between 50%–80% FFR study was performed, with FFR of 0.8 or less considered functionally significant. Blinded, independent, qualitative visual analysis by two experienced readers was performed to confirm existence of true perfusion defects on both 2D and 3D perfusion CMR datasets. Vascular territories in relation to CMR perfusion defects were assigned as per AHA 16 segment classification.

Results Prevalence of CAD was 62%. Of the 29 patients studied, 6 (21%) had single-vessel disease, 8 (27%) had two-vessel disease (2VD), 4 (14%) had three-vessel disease (3VD), and 11 (38%) had no significant CAD. In view of the small sample size qualitative analysis was undertaken to determine concordance between hi-res 2D and 3D perfusion CMR. Prevalence of perfusion defects relating to the three vascular territories have been detailed in Table 1.

Conclusions Our study shows an excellent agreement for both modalities in detecting FFR positive CAD on a per patient basis. On a per vessel basis, between the two, agreement in the LAD territory appears to be best. Discrepancy appears to be most in the circumflex territory. 3D perfusion CMR appears to detect circumflex ischaemia more accurately possibly due to better coverage of the basal left ventricle. (Ref Fig 1 and Fig 2) The lateral wall is often the thinnest making it more difficult to detect perfusion abnormalities – better coverage with 3D may be beneficial here. However, in the circumflex vascular territory we also observe that, in the absence of flow limiting disease in the circumflex artery (comparatively with the LAD and RCA) perfusion defects appear to be more prevalent, reflecting a limitation of the AHA classification.

Abstract 114 Figure 1 3D perfusion images showing perfusion defects in LAD (see blue arrows) and circumflex territory (see red arrows)

Abstract 114 Table 1

Abstract 114 Figure 2 High resolution 2D perfusion images in the same patient showing perfusion defects in the LAD territory only (see blue arrows)

  • perfusion
  • magnetic
  • resonance

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.