The influence of left bundle branch block on myocardial T1 mapping

Abstract

Tissue characterisation using T1 mapping has become an established magnetic resonance imaging (MRI) technique to detect myocardial diseases. This retrospective study aimed to determine the influence of left bundle branch block (LBBB) on T1 mapping at 1.5 T. Datasets of 36 patients with LBBB and 27 healthy controls with T1 mapping (Modified Look-Locker inversion-recovery (MOLLI), 5(3)3 sampling) were included. T1 relaxation times were determined on mid-cavity short-axis images. R² maps were generated as a pixel-wise indicator for the goodness of the fit of T1 maps. R² values were significantly lower in patients with LBBB than in healthy controls (whole myocardium/septum, 0.997, IQR, 0.00 vs. 0.998, IQR, 0.00; p = 0.008/0.998, IQR, 0.00 vs. 0.999, IQR, 0.00; p = 0.027). Manual correction of semi-automated evaluation tended to improve R² values but not significantly. Strain analysis was performed and the systolic dyssynchrony index (SDI_global) was calculated as a measure for left ventricular dyssynchrony. While MRI is generally prone to artefacts, lower goodness of the fit in LBBB may be mainly attributable to asynchronous contraction. Therefore, careful checking of the source data and, if necessary, manual post-processing is important. New techniques might improve the goodness of the fit of T1 mapping by reducing sampling in the motion prone diastole of LBBB patients.

Figure 1

Example of T1 mapping with mid-ventricular epi- and endomyocardial contours on the left and the corresponding R² map on the right. Brighter pixels on the R² map indicate a better goodness-of-the-fit. Areas with T1 values with a poor fit to the T1 recovery curve are displayed as darker pixels.

Figure 2

Comparison of R² values in patients with LBBB and controls in the whole mid-ventricular myocardium and the mid-ventricular septum after semi-automatic post processing. An asterisk (*) indicates a statistically significant difference (p < 0.05).

Figure 3

Single-shot acquisitions in T1 mapping (modified Look-Locker sequence) of a patient with LBBB in the upper two rows (a–h, T1 _{whole myocardium} 1012 ms, R² _{whole myocardium} 0.996) and of a healthy control in the lower two rows (i–p, T1 _{whole myocardium} 972 ms, R² _{whole myocardium} 0.999). The respective fitting curve of the whole myocardium is shown to the right, x-axis shows the TI in milliseconds (ms), y-axis shows the signal intensity. The white arrows indicate the deviation of the position of the septum during the eight sequential acquisitions and the resulting poorer fitting to the curve.