Marylène Delcey1,2,3,4, Isabelle Saniour5, Pierre Bour1,2,4, Fanny Vaillant1,2,4, Emma Abell1,2,4, Wadie Benhassen3, Marie Poirier-Quinot5, and Bruno Quesson1,2,4
1IHU Liryc, Electrophysiology and Heart Modeling Institute, Fondation Bordeaux Université, Pessac, France, 2Univ. Bordeaux, Centre de recherche Cardio-Thoracique de Bordeaux, U1045, Bordeaux, France, 3Siemens Healthcare SAS, Saint-Denis, France, 4INSERM, Centre de recherche Cardio-Thoracique de Bordeaux, U1045, Bordeaux, France, 5IR4M, UMR8081, Université Paris-Sud/CNRS, Université Paris-Saclay, Orsay, France
Synopsis
In the context
of cardiovascular diseases, precise determination of the extent and locations of the
arrhythmogenic substrate could significantly improve diagnosis and treatment for
both atrial and ventricular electrical diseases. However, the current spatial
resolution and signal-to-noise ratio (SNR) in clinical scanners remain
insufficient to provide relevant information of small structures like atrial
wall or sub-millimeter fatty infiltration in ventricle. To address this
limitation in SNR, two receiver coils were implemented at 1.5T for high
resolution cardiac imaging, with different active decoupling techniques (safety
aspects). Images at 200 µm in-plane spatial resolution were successfully
obtained on a beating heart.
Introduction
Atrial and ventricular cardiac arrhythmias
constitute the main cause of death worldwide. Using ex vivo MRI and
histological validation, the extent of fibrosis was shown to be highly
correlated with AF occurrence, severity and duration [1]. A precise description of the arrhythmogenic substrate
could significantly improve the diagnosis of cardiovascular diseases in both
atria and ventricles. However, the current spatial resolution and
signal-to-noise ratio (SNR) in clinical scanners remain insufficient to provide
relevant information, particularly for imaging small structures like the atrial
wall (thickness ranging 2 to 5mm) or in presence of sub-millimeter fatty
infiltrations in ventricles. To address the intrinsic limitation in SNR,
several research groups proposed to use a local RF coil positioned close to the
region of interest [2–4]. In this
study, we developed small receiver coils interfaced with a clinical 1.5T MRI, using
two active decoupling techniques based on electrical or optical approaches. Gains
in SNR and spatial selectivity compared to conventional multi-element cardiac
coils were measured on a phantom. 3D high-resolution images were obtained with
the developed instrumentation on an ex vivo cardiac sample from sheep. Finally,
preliminary results of high resolution MRI on ex vivo beating heart from pig are presented. Since in the context of arrhythmia, triggering on electrical
activity is difficult, we also implemented a pressure sensor to trigger the
acquisition on the mechanical status of the selected cardiac chamber to be
imaged.Methods
Two receive-only loop coils (2-cm in diameter)
were built (35 µm-thick copper on a FR4 substrate). These coils were
matched to 50 Ω and tuned to 63.6 MHz using capacitors and connected to an Aera
1.5T MR (Siemens Healthcare, Erlangen, Germany) with a dedicated interface box
(figure 1.C). One coil was decoupled using conventional PIN diode driven by Direct
Current (DC) while the second one was detuned using a photodiode (figures 1.A
and 1.B) activated through a laser fiber driven by an homemade electrical-to-optical
(25mW modulated laser diode, Laser Components, Germany, λ = 650 nm) conversion
circuit (figure 1.D) [5].
The reflection coefficient modulus |S11| and Q
factor were measured using a Vector Network Analyzer (VNA) when coils were
loaded with a quality assurance phantom. SNR maps were measured on this phantom
for both coils and compared to values obtained with the clinical 18 elements
cardiac coil. Imaging parameters were: 2D Gradient Echo (GRE), TR/TE = 1041/4.41ms,
TA = 50s, FA = 15°, pixel size = 1*1mm², thickness = 2.5mm, FOV =
200*200mm², BW = 293Hz/Px.
3D 200µm
isotropic resolution images were obtained using the galvanic detuned coil on an
ex vivo sample of sheep septum fixed into formalin mixed with gadolinium
(figure 3.A). Imaging parameters were: 3D GRE, TR/TE = 40/22ms, TA = 15min, FA = 33°,
pixel size = 200µm3 isotropic, FOV = 80*50*22.4mm3, BW = 172Hz/Px.
Beating heart (N=3, protocol approved by ethic committee) perfused in
Langendorf mode [6] were imaged with the implemented coil for testing high resolution
imaging in presence of motion. The local coil was placed in contact with the
left ventricle wall using a home-made holder. Left intraventricular pressure
was recorded using a small tubing inserted into the left ventricle connected to
a recording system (Labchart, ADInstruments, Sydney, Australia). From these
readings, a threshold was defined to generate a TTL signal that trigged MRI
acquisitions on the diastolic phase. High resolution GRE images with reduced
FOV were acquired with the following parameters: TR/TE = 162/25ms, A = 2min56’,
FA = 60°, pixel size = 200*200µm², thickness = 2.5mm, FOV = 70*70mm²,
BW = 131Hz/Px.Results
Matching of the two implemented coils was optimized
to be lower than -20dB (|S11| = -23dB and -27dB for the galvanic and optical
coils, respectively). Qloaded were 87 and 82 for the galvanic and
optical coil, respectively. Decoupling was efficient and similar for
both solutions. The measured SNR was higher than 1000 for both coils (figure 2.A).
SNR gain with respect to the 18-element coil was higher than 15 for both loop
coils, with similar performances whatever the detuning strategy. The signal
profile was uniform over a 2*3*1cm3 (length*width*depth) region
(figure 2.B) giving an adequate coverage for cardiac wall imaging (typically 2-5mm
thickness for atrium and up to 15 mm for the ventricle). Successful 3D high
resolution (200µm isotropic) images acquired on an excised septum revealed
sub-millimeter structural details (figure 3B). On ex vivo beating heart from
pig, the MRI sequences were successfully triggered on the intraventricular
pressure, allowing acquisition of high-resolution images (200µm² in-plane
resolution) and visualization of small vessels (figure 4).Conclusion and Discussion
The local loop coil allowed a significant increase in SNR with an
achievable gain of 15 compared to a clinical chest coil. We demonstrate that
optical detuning offers similar performances than conventional electrical
decoupling, increasing safety in perspective of future catheter-based coil
development (avoiding use of DC). We also demonstrate that intra-chamber
pressure can be used to trigger the acquisition and results in high quality
images. Combination of high sensitivity and limited spatial coverage using this
implementation allowed to obtain images on conventional clinical MRI scanner with
a higher spatial resolution than state-of-the-art cardiac imaging techniquesAcknowledgements
This work received financial support from the French National Investments for the Future Programs: ANR-10-IAHU-04 (IHU Liryc),Laboratory of Excellence ANR-10-LABX-57(Labex TRAIL) and ANR-11-INBS-0006 (France Life Imaging, Carcoi), Siemens Healthcare.
Helmut Stark is gratefiully thanks for technical support in coil interface.
References
[1]. Oakes
RS, et al. Detection and Quantification of Left
Atrial Structural Remodeling With Delayed-Enhancement Magnetic Resonance
Imaging in Patients With Atrial Fibrillation. Circulation 2009;119:1758–1767
[2]. Volland NA, et al. Initial feasibility testing of limited field of view magnetic resonance
thermometry using a local cardiac radiofrequency coil: Limited FOV MR
Thermometry with Local Coil. Magnetic Resonance in Medicine 2013;70:994–1004.
[3]. Homagk A-K, et al. An
expandable catheter loop coil for intravascular MRI in larger blood vessels.
Magnetic Resonance in Medicine 63:517–523.
[4]. Schmidt EJ, et al.
Intra-Cardiac MRI Catheter for EP Ablation Monitoring: Preliminary Studies. ISMRM 2011
[5]. Saniour I, et al. Active optical-based detuning circuit for
receiver endoluminal coil. Biomedical Physics & Engineering Express 2017;3:025002
[6]. Vaillant F, Magat J, Bour P, et al. Magnetic
resonance-compatible model of isolated working heart from large animal for
multimodal assessment of cardiac function, electrophysiology, and metabolism.
American Journal of Physiology-Heart and Circulatory Physiology 2016.