Left ventricular wall thickness measurements by magnetic resonance: a validation study

Summary Left ventricular (LV) wall thickness was determined by magnetic resonance (MR) in 15 patients (7 controls and 8 patients with coronary artery disease). End-diastolic (ed) and end-systolic (es) wall thickness were measured in a short axis view perpendicular to the LV long axis. Wall thickness measurements were compared to data obtained by digital subtraction angiography (DSA) and M-mode echocardiography (Echo).End-diastolic and end-systolic wall thickness were significantly overestimated by MR (34% and 37%, respectively) when compared to DSA. In contrast, LV end-diastolic and end-systolic chamber diameter were significantly underestimated by MR (25% and 30%, respectively) when compared to DSA. However, fractioned shortening was similar (all NS) for MR (48±22%), DSA (54±15%) and Echo (44±10%), respectively.The mean difference (= accuracy) and the standard deviation of difference (= precision) for LV wall thickness was 0.4±0.2 cm between MR and DSA, 0.4±0.3 cm between MR and ECHO and 0.03±0.1 cm between DSA and ECHO. The correlation of wall thickness between MR and DSA (correlation coefficient r=0.74, p<0.001) and between MR and Echo (r=0.70, p<0.001) was good although the standard error of estimate (SEE) was 17% for MR vs. DSA and 21% for MR vs. Echo. The corresponding SEE for chamber diameter was 16% between MR and DSA and 19% between MR and Echo, respectively. Intraobserver variability for wall thickness determination by MR was excellent (correlation coefficient r=0.99, p<0.001) SEE of 4%. Interobserver variability was also good (correlation coefficient r=0.90, p<0.001) with a SEE of 12%.It is concluded that LV wall thickness and chamber diameter (short axis plane) can be determined by MR with good precision but only satisfactory accuracy. LV wall thickness is significantly overestimated probably due to signals from static blood which might be indistinguishable from the subendocardium.