Comparison of velocity vector imaging echocardiography with magnetic resonance imaging in mouse models of cardiomyopathy

Abstract

Background: Myocardial strain imaging using echocardiography can be a cost-effective method to quantify ventricular wall motion objectively, but few studies have compared strain measured with echocardiography against magnetic resonance imaging (MRI) in small animals.

Methods and results: We compared circumferential strain (CS) and radial strain (RS) measured with echocardiography (velocity vector imaging [VVI]) to displacement encoding with stimulated-echo MRI in 2 mouse models of cardiomyopathy. In 3-month-old mice with gene targeted deficiency of cardiac myosin-binding protein-C (cMyBP-C(-/-), n=6) or muscle LIM protein (MLP(-/-), n=6), and wild-type mice (n=8), myocardial strains were measured at 3 cross-sectional levels and averaged to obtain global strains. There was modest correlation between VVI and MRI measured strains, with global CS yielding stronger correlation compared with global RS (CS R(2)=0.4452 versus RS R(2)=0.2794, both P<0.05). Overall, strain measured by VVI was more variable than MRI (P<0.05) and the limits of agreement were slightly, but not significantly (P=0.14), closer for global CS than RS. Both VVI and MRI strain measurements showed significantly lower global CS strain in the knockout groups compared with the wild type. The VVI (but not MRI) CS strain measurements were different between the 2 knockout groups (-14.5±3.8% versus -6.6±4.0%, cMyBP-C(-/-) versus MLP(-/-) respectively, P<0.05).

Conclusions: Measurements of left ventricular CS and RS are feasible in small animals using 2-dimensional echocardiography. VVI and MRI strain measurements correlated modestly and the agreement between the modalities tended to be greater for CS than RS. Although VVI and MRI strains were able to differentiate between wild-type and knockout mice, only global CS VVI differentiated between the 2 models of cardiomyopathy.

Figure 1

DENSE MRI (1A) and VVI echocardiographic (1B) vector maps for regional strain in wild type (left) vs. KO (right). The KO group has decreased strain compared to the wild type, demonstrated by smaller size of the vectors.

Figure 1

DENSE MRI (1A) and VVI echocardiographic (1B) vector maps for regional strain in wild type (left) vs. KO (right). The KO group has decreased strain compared to the wild type, demonstrated by smaller size of the vectors.

Figure 2

Bland-Altman Plots comparing the MRI- and echo-derived Strain. (A) Circumferential strain measurements. (B) Radial strain measurements. There is closer agreement between the MRI and echo imaging modalities for circumferential strain compared to radial strain measurements. Upper and lower dotted lines are the 95% limits of agreement.

Figure 2

Bland-Altman Plots comparing the MRI- and echo-derived Strain. (A) Circumferential strain measurements. (B) Radial strain measurements. There is closer agreement between the MRI and echo imaging modalities for circumferential strain compared to radial strain measurements. Upper and lower dotted lines are the 95% limits of agreement.

Figure 3

Regional radial strain: (*)P < 0.05 Echo vs. MRI, (†)P < 0.05 Wild type vs. KO, (**)P < 0.05 MyBPC^−/− vs. MLP^−/− (Echo only). Regional radial strain measurements using MRI and echo in the wild type group are similar. Regional radial strain measurements by MRI in the KO groups are significantly lower compared to the wild type group but more variable via echo.

Figure 4

Regional Circumferential Strain: (*)P < 0.05 Echo vs. MRI, (†)P < 0.05 Wild type vs. KO, (**)P < 0.05 MyBPC^−/− vs. MLP^−/− (Echo only). Regional circumferential strain measurements using MRI and echo in the wild type group are similar. Regional circumferential strain measurements by MRI in the KO groups are significantly lower compared to the wild type group but more variable via echo.