1st supervisor: Sebastien Roujol, King’s College London
2nd supervisor: Rene Botnar, King’s College London
3rd supervisor: Reza Razavi, King’s College London
4th supervisor: Radhouene Neji, Siemens Healthineers
In this project, novel CMR-thermometry methods will be developed for advanced real-time monitoring of RFA lesions during RFA of cardiac arrhythmias. The proposed methodological developments will be evaluated in phantom, healthy volunteers and in-vivo in pre-clinical studies. In-vivo validation will be based on a Wellcome Trust funded study to investigate VT ablation inside MRI system (2016-2020, led by Prof. Razavi) and an EPSRC funded study on advanced MRI guidance of VT ablation (2018-2021, led by Dr. Roujol). The project is organized in 4 aims:
Aim 1: ECG-less MR-thermometry (12 months)
Current CMR-thermometry sequences are ECG-triggered. However the ECG signal and automatic detection of the R-wave is sometime unreliable (especially in the context of an EPI sequence). Our MR-compatible ablation catheter has two micro coils near its tip which enable real-time tracking of the catheter tip and orientation.
In Aim 1, a novel MR-thermometry sequence will be developed by acquiring active tracking data continuously outside the MR-thermometry readout. The temporal evolution of the catheter tip will represent the influence of both cardiac and respiratory motion and potential catheter drift. After removal of the drift/respiratory motion component (using band-pass filtering as we previously proposed in Roujol et al., Plos, 2013), the remaining signal will exploited for automatic determination of the cardiac phase and cardiac triggering.
Aim 2: CMR-thermometry with improved spatial coverage (12 months)
Current CMR thermometry sequences are unable to predict the amount of tissue destruction in through-plane dimension due to multiple slices being acquired at different cardiac phases.
In Aim 2, we will develop and investigate the potential of simultaneous multi-slice (SMS) for improved slice consistency during real-time MR-thermometry.
Aim 3: Advanced temperature map filtering (9 months)
Noise in temperature maps adversely affect thermal dose maps and the prediction of the extent of tissue destruction. The minimisation of noise in temperature maps is thus crucial.
In Aim 3, We will develop and investigate the potential of novel spatiotemporal filter using the Kalman filtering framework and a new empirical model of bio heat transfer based on an auto-calibrated 3D half Gaussian function, real-time RF power, and real-time catheter tip location measured from active tracking.
Aim 4: Novel algorithm for prediction of post-RFA tissue temperature maps (9 months)
Post-RFA tissue temperature is mainly influenced by heat diffusion and blood perfusion and should be well characterised for accurate thermal dose estimation. This would require to continue the acquisition after each RFA (for few minutes), prolonging the procedure.
In Aim 4, we will develop and evaluate the potential of a novel prediction algorithm of post-RFA temperature maps will be developed using a physical model of post-RFA heat transfer which is governed by heat diffusion and blood perfusion.