Proof of concept of extracorporeal cardiac stimulation with HIFU has recently been reported on a large animal model [1]. Short duration (millisecond range) ultrasound bursts with sufficient pressure at the focus have been shown to induce premature ventricular contractions (PVC) when applied to the ventricle after the refractory period (so-called “ST” period within the cardiac cycle). Such a new application of non-invasive MR-HIFU may enable the development of innovative diagnostic and therapeutic approaches in cardiology. However, depending on the anatomical structure embedded in the HIFU beam path (ribs, lungs, fat layer,…), the resulting pressure applied to the cardiac muscle at the targeted location may vary importantly. We propose here to use MR-Acoustic Radiation Force Imaging (MR-ARFI) technique to visualize and to measure the net local tissue displacement induced by the HIFU pulse [2]. A fast MR-ARFI method was developed including synchronization of the sequence at selectable timings within the cardiac cycle. The method was evaluated on an ex-vivo beating heart model from pig, allowing evaluation of the imaging method under well controlled experimental conditions. Displacement maps computed from MR-ARFI measurements were exploited for localization purpose of the HIFU pulse within the cardiac muscle and for quantification of the resulting displacement during refractory (contraction) and non-refractory (resting time) periods. Pig heart (45kg, N=1) was extracted and installed on a MR-compatible isolated beating heart setup and positioned on top of a HIFU transducer (256 elements, 1MHz operating frequency, 13/13 cm focal/aperture, Imasonic, France), inside the MR scanner (1.5T Avanto Siemens, Germany) (Fig1). The heart was electrically paced on the epicardium at 120bpm and local electrophysiology (EP, using MR-compatible catheter) and left intra-ventricular pressure were monitored continuously. ARFI displacement was monitored using a modified spin echo Single-Shot EPI sequence integrating two bipolar motion encoding gradients (MEG) located symmetrically from the refocusing RF pulse. The sequence was trigged on the EP signal of the heart with user-defined adjustable delays. ARFI maps were computed in real-time on Thermoguide™ (IGT,Pessac,France) as followed:$$ Dn>10= (Φn - Φ {ref})/ (γ.|A|.2. Δ) $$ with Φn and Φref respectively the current and reference (meaned over first 10 dynamics) phase maps, n the dynamic number, |A| and Δ respectively the MEG amplitude and duration of one lobe and γ the gyromagnetic ratio. Sequence parameters were: FOV=257x257mm², one slice, TE/TR/FA=42ms/98ms/75°, refocusing angle = 180° , 7/8 partial Fourier, a spatial resolution of 2.3x2.3x4mm3, |A| =25mT/m and Δ =3ms. Two 6 element surface coils and a 19cm loop coil were positioned around the heart for image acquisition. HIFU sonications were triggered by the scanner and lasted 13ms to cover the refocusing pulse duration and one lobe of MEG on both of its sides. The sequence was run continuously (100 repetitions) and 5 HIFU pulses (with several acoustic power values ranging from 180 W to 270W) were applied once every 15 repetitions to allow recovering of the normal sinus rhythm in case of successful HIFU stimulation. Experiments were performed in the left ventricle and triggered to the non-refractory period (N=40) and to the refractory period (N=10).

HIFU applied during the refractory period fig2A) induced no premature extra systole even at the maximal power used. ARFI maps confirmed that sonication was performed in the myocardium and a maximum displacement of 22±2.6µm was measured. Premature extra systole were induced when the ultrasonic pulse was sent after the absolute refractory period, confirmed by EP and pressure measurements of LV: on LV EP, ultrasound induced a depolarization prior do the pace signal and the pressure pattern was modified fig2B). The success rate of stimulation over all sets of experiments was 95%. Under these conditions, displacement maps confirmed the localization of the stimulation site with a maximum displacement of 38±2.7µm. Fig3 shows results of displacement values during refractory and non-refractory periods at maximal HIFU power. A significant increase of displacement was observed in the non-refractory period when the heart is in resting time and not anymore under pressure, attesting of a change in tissue stiffness.

This study demonstrates the ex-vivo beating heart proof of feasibility of controlled noninvasive ultrasound-based cardiac stimulation with accurate localization of the depolarization site using a fast MR-ARFI sequence . The quantitative measurement of local tissue displacement during refractory and non-refractory periods of the cardiac cycle from displacement maps may provide additional information on the cardiac elasticity changes between diastole and systole.