Contrast-enhanced magnetic resonance imaging assessment of scar size in patients with chronic myocardial infarction.

Background: In the assessment of myocardial infarction (MI) mass, contrast-enhanced magnetic resonance imaging (CE-MRI) is comparable to single-photon emission computed tomography (SPECT). The aim of the present study was to determine whether the MI area, as assessed at CE-MRI and SPECT, is comparable to mass evaluation. We also compared CE-MRI and SPECT estimates of the MI area with functional evaluations made at echocardiography and kinetic MRI (cine-MRI).

Methods: We used a 1.0 Tesla MRI scanner and an inversion-recovery turboFLASH sequence, a tomographic gamma-camera and second-harmonic ultrasound systems. Two blinded operators assessed the extent of scarring, expressed as a percentage of the whole left ventricle (LV), using a 16-segment model. We studied 55 consecutive patients with a clinically stable healed MI (50 Q wave, 5 non-Q wave).

Results: The scar mass was 19+/-23% of the LV at CE-MRI and 21+/-25% at SPECT; the scar area was 29+/-23% of the LV at CE-MRI, 41+/-28% at SPECT, 29+/-31% at cine-MRI, and 32+/-29% at echocardiography. The Bland-Altman bias between CE-MRI and SPECT mass estimations was -2% of the LV with a+/-23% limit of agreement (LOA), while the bias between the area assessments was -12% with a+/-42% LOA. Bias between CE-MRI and functional evaluation by cine-MRI and echocardiography was 0% with a+/-39% LOA and -3% with a+/-36% LOA respectively. Comparing SPECT with cine-MRI and echocardiography the bias was 12% with a+/-52% LOA and 9% with a+/-56% LOA respectively.

Conclusions: CE-MRI has proved to be comparable to SPECT in the assessment of the healed MI mass. Conversely, a high systematic error (high bias and LOA) renders CE-MRI and SPECT assessments of the MI area incomparable. Similarly (high bias and/or LOA) CE-MRI and SPECT estimations of the MI area cannot be compared with functional evaluation by echocardiography or cine-MRI.