Advertisement

High resolution ultra-fast sparse sampling with iterative reconstruction imaging for left ventricular evaluation: clinical comparison with standard SSFP imaging

  • Christian R Hamilton-Craig
  • Robyn Riley
  • Andrew Trotter
  • Johanne Neill
  • Michael O Zenge
  • Michaela Schmidt
  • Benjamin Schmitt
  • Wendy E Strugnell
Open Access
Oral presentation

Keywords

Iterative Reconstruction Sparse Sampling Excellent Image Quality Coil Sensitivity Iterative Reconstruction Imaging 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Background

Cardiovascular MRI (CMR) has the advantages of high temporal and spatial resolution with excellent endocardial definition, but ECG-gated cine SSFP imaging requires multiple sequential breath-holds and is time-consuming. Sparse sampling with iterative reconstructionallows for ultra-fast acquisitions of SSFP cine images. The accuracy of this technique compared to standard SSFP imaging requires further validation in the clinical setting.

Methods

Prospective, observational study of clinical patients referred for CMR imaging for left ventricular evaluation, performed on a Siemens Magnetom Aera 1.5T scanner in a large, tertiary clinical service. All patients had a standard left ventricular short axis stack using standard ECG-triggered TrueFISP imaging (TE 1.3ms, TR 36ms, inplane spatial resolution 1.2 x 1.2 mm2), requiring 10-14 breath-holds of 10-15s, and then repeated with an ECG-triggered prototype TrueFISP sequence (TE 1.2ms, TR 31ms, inplane spatial resolution 1.5x1.5 mm2) with a constrained undersampling resulting in a net acceleration factor of 4 (CV_sparse), acquired in a single breath-hold or two breath-holds (depending on cardiac size). For the CV_sparse approach, coil sensitivity maps were calculated from the temporal average of the input data in a central region of k-space and image data were then reconstructed using a non-linear iterative reconstruction with k-t regularization of the SENSE type dataset. Datasets were analysed by two experienced readers (>10 years CMR analysis), with papillary muscles and trabeculae included in the blood pool. Comparison of TrueFISP and CV_sparse was assessed using paired t-test and Pearson's correlation, and inter-observery reproducibility by Bland-Altman

Results

26 clinical patients had both TrueFISP and CV_sparse imaging successfully performed (figure 1A,B), mean LVEF 49% (range 16-65%). Strong positive correlation was present for end-diastolic volumes (r=0.98, P< 0.0001), end-systolic volume (r=0.99, P< 0.0001), LVmass (r=0.97, P< 0.0001) and LVEF (r=0.97, P< 0.0001), with no difference between the groups (P=0.95 NS), figure 1C. Reproducibility was excellent with bias 0.3%, limits-of-agreement -3.7 to 4.3% (figure 1D). Acquisition time for the LVSAX series was significantly shorter for CV_sparse compared to TrueFISP (p<0.001), however CV_sparse images require longer reconstruction times (~5 minutes), during which clinical image acquisition can continue.

Figure 1

A. TrueFisp B. CV_SPARSE C. Pearson's Correlation; D. Bland-Altman LVEF%

Conclusions

Highly accelerated imaging using TrueFISP with sparse sampling and iterative reconstruction provides excellent image quality with accurate, reproducible results equalling those of standard TrueFISP imaging, but with significantly more rapid acquisition times. This accelerated technique appears valid and ready for application to clinical practice.

Funding

Smart Futures Fellowship Early Career Grant (Queensland State Government #ISF783).

Copyright information

© Hamilton-Craig et al; licensee BioMed Central Ltd. 2015

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/4.0), 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.

Authors and Affiliations

  • Christian R Hamilton-Craig
    • 2
    • 1
  • Robyn Riley
    • 2
  • Andrew Trotter
    • 2
  • Johanne Neill
    • 2
  • Michael O Zenge
    • 4
  • Michaela Schmidt
    • 4
  • Benjamin Schmitt
    • 3
  • Wendy E Strugnell
    • 2
  1. 1.Centre for Advanced ImagingUniversity of QueenslandBrisbaneAustralia
  2. 2.Richard Slaughter Centre of Excellence in CVMRIThe Prince Charles HospitalBrisbaneAustralia
  3. 3.Imaging & Therapy Systems DivisionSiemens Healthcare AustraliaBrisbaneAustralia
  4. 4.Magnetic ResonanceSiemens Healthcare AGErlangenGermany

Personalised recommendations