Direct comparison of high‐temporal‐resolution CINE MRI with Doppler ultrasound for assessment of diastolic dysfunction in mice

Figures

• 

  Figure 1. Comparison of standard cinematic (CINE)‐derived cardiac systolic parameters in the control (n = 8) and infarcted (n = 13) groups. Error bars indicate standard error of the mean. Mann–Whitney U‐tests performed between groups for each parameter: ***p < 0.001; *p < 0.05. EDV, end diastolic volume; EF, ejection fraction; ESV, end systolic volume; SV, stroke volume

• 

  Figure 2. Representative filling rate curves from the same mouse measured using ultrasound and high‐temporal‐resolution cinematic magnetic resonance imaging (HTR‐CINE MRI) showing early‐filling (E) and atrial/late‐filling (A) peaks. (A) ultrasound image of mitral blood flow showing Doppler measurements overlaid. (B) HTR‐CINE image showing left ventricle blood pool segmentation. (C) ultrasound filling curve of blood velocity through the mitral valve. (D) HTR‐CINE‐derived left ventricle volume curve (black) and filling rate curve (blue)

• 

  Figure 3. (A) standard prospective cinematic (CINE) stack showing short‐axis images through a control mouse heart at diastole. Red boxes denote slices chosen for the determination of the most optimum slice in early‐filling to atrial/late‐filling ratio (E/A) measurements. (B) raw data from repeated measures of E/A in control mice using ultrasound and magnetic resonance imaging (MRI) at three different slice locations (apical, mid‐ventricle and basal) and when all three slices were combined (for all repeat measurements, p > 0.05, Wilcoxon signed‐rank test)

• 

  Figure 4. Bland–Altman plots showing repeatability of the early‐filling to atrial/late‐filling ratio (E/A) [and more routinely assessed cardiac parameters – Ejection fraction (EF), stroke volume (SV), end diastolic volume (EDV)] in the naïve cohort. (A) E/A measured using ultrasound. (B–D) EF, SV and EDV measured using magnetic resonance imaging (MRI). (E–H) E/A measured using high‐temporal‐resolution cinematic MRI (HTR‐CINE MRI) at different locations in the heart (basal, mid‐ventricle and apical) and with all three slices combined. Full lines represent the mean percentage difference between trials and broken lines represent 95% confidence intervals (±1.96 × standard deviation of the difference between the trials)

• 

  Figure 5. (A) short‐axis view of late‐gadolinium‐enhanced magnetic resonance imaging (LGE MRI) in an infarcted mouse heart (1, most apical slice; 8, most basal slice). Hyperintensity within the myocardium corresponds to infarcted tissue. (B) correlation of ejection fraction (EF) and area of gadolinium enhancement in the infarcted cohort for the determination of infarct severity. The red box denotes animals with minor cardiac impairment [EF > 50%, total infarct volume as a percentage of the left ventricular volume (LGE/LV) < 10%] where surgery was ineffective. These animals were excluded from subsequent analysis. LV, left ventricle

• 

  Figure 6. (A) comparison of early‐filling to atrial/late‐filling ratio (E/A) measurements determined by ultrasound (US) and high‐temporal‐resolution cinematic magnetic resonance imaging (HTR‐CINE MRI) in the infarcted cohort of mice with ejection fraction (EF) < 50% and total infarct volume as a percentage of the left ventricular volume (LGE/LV) > 10% (n = 9). There was a significant difference in E/A between control and infarcted mice when measured with MRI (p = 0.02, Mann–Whitney U‐test), which was not observed with ultrasound (p = 0.15). Ns, not significant. (B) Bland–Altman plot of E/A between ultrasound and MRI across all 13 animals in the infarcted cohort. E/A values tended to be lower when assessed with MRI relative to ultrasound (bias, 22%). Spearman's coefficient showed that the two modalities were strongly correlated (ρ s = 0.89, p = 0.00004)