Three-dimensional analysis of left ventricular geometry using magnetic resonance imaging: feasibility and comparison with echocardiographic analysis

OBJECTIVES: Reliability of left ventricular geometry assessed by echocardiography (Echo) using an assumed left ventricular mass (LVM) and one-dimensional eccentricity (relative wall thickness: RWT), remains questionable. This study evaluated the feasibility of three-dimensional left ventricular geometric analysis using magnetic resonance imaging (MRI). METHODS: Echocardiography and MRI were performed on 55 patients with hypertension. LVM was calculated using 0.8 (American Society of Echocardiography-cube LVM) + 0.6 g for Echo and the slice summation method for MRI. Eccentricity was determined by RWT (septal wall thickness + posterior wall thickness/left ventricular inner diameter) for Echo and LVM/1.05/left ventricular end-diastolic volume (LVEDV) ratio [MRI-mass volume/cavity (M/C) ratio] for MRI. Left ventricular geometry was classified into four patterns according to the presence/absence of left ventricular hypertrophy and abnormal/normal eccentricity (partition value: RWT = 0.44, MRI; M/C ratio = 2.0), and the patient distribution was compared between the two methods. RESULTS: Although the mean values for LVM were similar, the mean value for LVEDV by echocardiography was significantly higher (p < 0.0001) and the mean M/C ratio was significantly lower (r = 0.004) than those by MRI. There were widely dispersed LVM values at higher underlying values of LVM and significant correlations between MRI-LVEDV and MRI-LVM (r = 0.87) and between Echo-LVEDV and Echo-LVM (r = 0.75). There was a significant difference in patient distribution according to left ventricular geometric pattern between the two methods (p < 0.01). Concentric (n = 18) and eccentric hypertrophy (n = 12) were dominant patterns in Echo analysis, and concentric hypertrophy (n = 23) and concentric remodeling (n = 21) were dominant in MRI analysis. The left ventricular geometric patterns were different in 32 patients (58.0%). Inadequate LVEDV values in Echo were the primary cause of this phenomenon. CONCLUSIONS: Left ventricular geometric analysis by Echo results in inaccurate values. Three-dimensional left ventricular geometric analysis using MRI provides more accurate information about left ventricular geometry.