1350735 - Feasibility of fast volumetric thermometry in the heart in a single cardiac phase

Background: CMR-thermometry shows promise for real-time temperature mapping in the heart and for the prediction of ablation lesion during MRI-guided catheter-based ablation of cardiac arrhythmias [1]. Current CMR-thermometry techniques enable the acquisition of multiple contiguous slices (typically N_slices=3-4) per heartbeat, necessary to cover the ablation region. The acquisition of this stack of slices takes about 300-400ms [2] as each slice requires pre-pulses (including fat suppression and saturation bands for in-flow blood suppression) in addition to an EPI readout. This results in different slices being acquired at different cardiac phases, which severely limits lesion assessment in the through-plane dimension [3]. In this study, we sought to develop a fast single-phase volumetric CMR-thermometry sequence using simultaneous multi-slice (SMS) imaging combined with a more time-efficient pre-pulse strategy.

Methods:

A diagram of the conventional and proposed CMR-thermometry sequences is shown in Figure 1. SMS single-shot EPI (with a multiband factor of 2) was used to acquire 4 slices. In-flow saturation slabs commonly played before each slice were replaced by a single double inversion recovery (DIR) pre-pulse for dark blood signal suppression (thickness of the slice-selective re-inversion pulse=40mm, delay between DIR pulses and first SMS shot=225ms). Fat suppression pulses were still played prior to each SMS-EPI shot.
This technique was evaluated in five healthy volunteers (4m, 1f) scanned on a 1.5T scanner (MAGNETOM Aera, Siemens Healthcare, Erlangen, Germany) under free breathing. All images were acquired with the following parameters: TE=24ms, TR=63ms, Flip Angle=60^o, Bandwidth=1420 Hz/px, FOV=250x250mm², in-plane voxel size=3.1x3.1mm², slice thickness=5mm, N_slices=4, multiband factor=2, number of dynamics=150, no slice gap. Motion corrected temperature maps were reconstructed offline using a multi-baseline approach (look-up table of 40 dynamics) for correction of respiratory induced phase variation, non-rigid image registration, and temporal temperature filtering, as previously described [4]. The stability of thermometry was calculated as the standard deviation over time of the temperature in the myocardium in all 4 slices.

 

Results: Successful blood signal suppression was achieved in all slices using the DIR pre-pulses, as shown in one subject (Figure 2).
Figure 3a shows the temperature map of one subject over the 4 slices. The distribution of thermometry stability over the 4 slices of each volunteer is shown in Figure 3b. Over all the subjects, the thermometry stability was 1.0±0.4°C.

Conclusion: The proposed method allows multiple (N_slices=4) contiguous slices per heartbeat to be acquired in the same cardiac phase, with good blood signal suppression and thermometry precision. Further studies evaluating the benefit of the technique for 3D lesion assessment are now warranted.