S. J. Gandy, S. A. Waugh, R. S. Nicholas, N. Rajendra, P. Martin, J. G. Houston



Publication year



International Journal of Cardiovascular Imaging

Periodical Number






Author Address

Gandy, SJ Ninewells Hosp, Dept Med Phys, NHS Tayside, Dundee DD1 9SY, Scotland Ninewells Hosp, Dept Med Phys, NHS Tayside, Dundee DD1 9SY, Scotland Ninewells Hosp, Dept Clin Radiol, NHS Tayside, Dundee DD1 9SY, Scotland Univ Dundee, Sch Med, Ninewells Hosp, Dundee DD1 9SY, Scotland

Full version

Aim Quantitative MRI assessments of cardiac structure and function are possible and potentially useful for longitudinal clinical monitoring. The aim of this study was to compare the magnitude and repeatability of left ventricular (LV) ejection fraction (EF) and mass (LVM) measurements in patients with clinically distinct cardiac conditions. Materials and Methods Patients were recruited into four groups: (i) congestive heart failure (CHF), (ii) left ventricular hypertrophy (LVH), (iii) recent post myocardial infarct (PMI), and (iv) healthy normal volunteers (HNV). LV short-axis images were acquired on a 1.5T MRI scanner and analysed on a satellite workstation. EF and LVM (at ED) values were derived from myocardial segmentations, and intra- observer test-retest coefficients of repeatability (CoR) were determined for each cohort. Results The mean EF for the CHF patients (30.3%) was lower than for the other cohorts (LVH 72.7%, PMI 53.0%, HNV 67.0%; P < 0.0002). As expected, the mean LVM for the CHF patients (143 g) was greater than for the other cohorts (LVH 122 g, PMI 124 g, HNV 107 g), but only significant when compared to the HNV cohort (P = 0.004). The intra-observer CoR values for EF were 1.5% (LVH), 1.6% (HNV), 2.6% (PMI) and 5.5% (CHF), and 4.6 g (HNV), 6.7 g (PMI), 8.3 g (CHF) and 9.8 g (LVH) for LVM. Conclusion The EF, LVM and associated repeatability parameters are variable and dependent upon the clinical condition under investigation. It is important that reproducibility data for EF and LVM are acquired individually and specifically on a per-cohort basis if the parameters are to form reliable endpoints for longitudinal clinical follow-up assessments.