MRI-based interventional guidance of High-Intensity Focused Ultrasound Therapy (HIFU) relies on the stereotactic correspondence between the MR-image and the location of the HIFU ablation area. However, discrepancies between the planned focal point location and the actual focus of the HIFU beam are frequently observed, predominantly due to acoustic wave-front aberrations induced by variations of the acoustic impedance in the beam propagation path. As a consequence, current clinical practice employs a “low power thermal test shot” before the start of the ablation, to provide a 3D correction of the exact focus location. The power level of this test-shot is generally chosen in such a way that it allows observing a distinct temperature elevation without causing irreversible tissue damage. Although this approach is safe and accurate in low perfused target tissues such as uterine fibroids, the high heat-extraction rate renders this approach of limited value in high-perfused organs such as liver or kidney. Furthermore, for therapeutic HIFU ablations in the non-linear pressure range [1], shock-wave formation and non-linear absorption adds further discrepancy between linear test-shots and therapeutic sonications.

Recently, MR-acoustic radiation force imaging (MR-ARFI) has been proposed as an alternative method to locate the focal point non-invasively without giving rise to non-desired tissue damage [2-5]. MR-ARFI evidences the tissue displacement due to the acoustic radiation force of the HIFU beam instead of the temperature. However, since MR-ARFI employs a highly motion sensitized sequence and displays limited sensitivity, it is challenging in abdominal organs and in-vivo experience with this type of this approach is currently limited to the liver.

The first aim of this study was to investigate the feasibility of respiratory-gated MR-ARFI on clinical HIFU equipment for the exact beam localization in the kidney. Of particular interest was hereby potential interference with other motion sources, such as cardiac pulsations in the vicinity of larger vessels such as the renal artery. The second goal was to validate if MR-ARFI can perform this task for both linear and non-linear acoustic energy delivery.

A porcine model was chosen for its similarities in size and perfusion to the human kidney. The local animal board approved the study. The pig was anesthetized and mechanically ventilated at a frequency of 13/min. A clinical Sonalleve MR-HIFU therapy system (Philips Healthcare, Finland) integrated with a 1.5T Achieva MRI (Philips Healthcare, The Netherlands) with minor modifications was used for MR-HIFU. An MR-HIFU treatment cell was positioned at 4.5cm from the skin in the cortex of the kidney. A gradient-recalled echo planar imaging 2D dynamic scan (TR 100ms, TE 30ms, flip angle 20°, FOV 168x168mm2, matrix 112x108, voxel size 1.51x1.51mm2, slice thickness 7mm) was performed to visualize displacement due to the acoustic radiation force similar to [2]. Thermometry data was obtained using a gradient echo with EPI (TR 100ms, TE 15ms, flip angle 20°, matrix 160x160, FOV 400x400mm2, voxel size 2.5x2.5mm2, slice thickness 5mm). All MR sequences, including anatomical images, and the acoustic energy delivery were respiratory gated based on pencil beam navigator images of the contralateral kidney. For comparison of linear and non-linear effects, four experiments at two different power levels were performed: Two thermal ablations with 450W CW and 1000W pulsed (30% duty cycle, tone-burst of 10500 cycles), and two MR-ARFI experiments at equivalent acoustic intensity (tone-burst of 2400 cycles, bipolar MR-ARFI gradient 20mT/m, 3ms).Displacements due to the radiation force could be reproducibly measured for both acoustic intensities and coincided with the location of the temperature elevation of the equivalent 450W and 1000W ablations, respectively (Figure 1). No interference/artifacts with the cardiac cycle were observed during MR-ARFI. The focus of the beam cone was evidenced more posteriorly than planned for the 1000W, most likely due to shock-wave formation and non-linear absorption. No significant heating was observed during the MR-ARFI experiments (<2°C).The employed respiratory-gated MR-ARFI sequence in combination with a 450W excitation tone-burst of 2400 cycles is sensitive enough to exceed noise and to clearly display the focus of the HIFU beam at a fraction of the energy compared to the 450W thermal ablation. Both at 450W and at 1000W the displacement due to the radiation force coincided with the location of the temperature rise due to thermal ablation at equivalent power. Hence, radiation force in combination with a pencil beam navigator to compensate for respiratory motion is a reliable indicator of the location of the thermal lesion and might be an alternative to the low power thermal test shot in highly perfused organs such as the kidney.