Skip to content
BY-NC-ND 3.0 license Open Access Published by De Gruyter September 12, 2015

US-tracked steered FUS in a respiratory ex vivo ovine liver phantom

  • Jan Strehlow , Xu Xiao EMAIL logo , Markus Domschke EMAIL logo , Michael Schwenke , Ioannis Karakitsios EMAIL logo , Senay Mihcin EMAIL logo , Julia Schwaab , Yoav Levy , Tobias Preusser and Andreas Melzer EMAIL logo


Organ motion is a major problem for Focused Ultrasound Surgery (FUS) of liver tumors. We present a liver phantom mimicking human respiratory motion (20 mm range, 3 − 7 s/cycle) and the evaluation of an ultrasound-tracked steered FUS system on that phantom. Temperature curves are recorded while sonicating in moving and static phantom. The temperature curves correlate well and show the ability of the system to compensate breathing like motion.

1 Introduction

While MR-guided Focused Ultrasound Surgery (FUS) is clinically established in the treatment of uterine myoma, brain and bone metastasis, new disease entities are continuously targeted by research [3]. A great challenge for the application of FUS in abdominal organs is organ motion due to respiration. If motion is ignored, the thermal dose spreads over a larger area, leading to insufficient heat for ablative necrosis. Gated approaches try to circumvent this by repeatedly halting respiration under general anesthesia. Besides increasing the complexity of the procedure these approaches also increase intervention time. To mitigate this problem, organ tracking for FUS beam steering approaches are investigated: Organ motion is sensed and the focus position is updated to compensate motion [2]. In [4], breathing belt signal and previously recorded motion data were used as a tracking surrogate. Relying on recorded data, this method is prone to changes in breathing motion. Ultrasound (US) based live-sensing of organ motion is used in [1] and [5]. These systems, however, were evaluated only in a slowly moving ex-vivo liver phantom (> 10 s/cycle) with a limited motion range (< 10 mm), and in anesthetized pigs with small liver motion range (6mm) and speed (5mm/s), respectively. In this study we present a breathing phantom exhibiting human breathing like motion (3−7 s/cycle, < 20 mm motion range) and an US imaging based steered FUS system that can compensate this motion.

2 Methods

2.1 Breathing phantom

A dynamic liver phantom was developed, simulating the human like breathing motion in terms of motion range and frequency. A fresh ex-vivo ovine liver was embedded in an 3 % agar gel (LOT-No.3 %, Fisher Scientific) block. To induce respiratory motion, the agar block was placed between an inflatable air balloon and a water balloon inside a water basin (Figure 1). The inflatable air balloon was connected to a lung ventilator (Pneupac®ventiPACTM Medical Ventilator, Smiths Medical, St Paul, USA). The ventilator blows air into the balloon to push the phantom from its initial (or expiration) position to a displaced (or inspiration) position. Once the lung ventilator stops filling the air balloon, the water balloon on the opposite side of the phantom acts as a repulsing device to move the liver phantom back to its initial position. The phantom moves forward and backward periodically, whereas overall displacement and cycle length per breath may be controlled precisely. The ovine liver was embedded into the agar block in a supine position such that the following directions of movement are given: The largest displacement direction caused by the lung ventilator represents the superior-inferior direction. A small gap between the agar block and the box allows a small motion in a perpendicular direction representing a left-right motion. Due to the construction of the box, an anterior-posterior displacement is neglected. Table 1 summarizes the accessible displacement within the described simulator in comparison to liver movement in the human body.

Figure 1 Ex-vivo ovine liver embedded into an agar phantom, showing (a) a schematic drawing and (b) a photograph of the arrangement, green balloons are sequentially filled and emptied to induce breathing like motion of the liver.
Figure 1

Ex-vivo ovine liver embedded into an agar phantom, showing (a) a schematic drawing and (b) a photograph of the arrangement, green balloons are sequentially filled and emptied to induce breathing like motion of the liver.

Table 1

Comparison between the motion of a liver in the respiratory motion simulator and a human body as reported in [6]

Human Liver17.9±5.13.0±2.05.1±3.13.9±0.7
Our phantom15.0 to 20.00.0 to 4.0N.A.3.0 to 7.0

2.2 Steering system

An MR-compatible digital phased array transducer (DiPhAS, Fraunhofer IBMT, St. Ingbert, Germany) was used to acquire the motion of the target. With a focal depth of 200 mm, the device can acquire images with 27 Hz. Tracking features, such as vessels or tissue interfaces, were defined manually and their position was tracked using an US tracking method described in [8]. To infer the position of the target from the tracking information liver motion had to be modeled. In this study a linear translational motion model was used, i.e. the translations of the tracking feature are assumed to be the transformations of the target. For adjustable focusing an electronically steerable multi-element transducer was used (Conformal Bone System, InSightec, Tirat Carmel, Israel) updating the focal spot position with up to 20 Hz. The high energy output of an active FUS renders the simultaneous US imaging unusable for tracking. To enable US tracking, the FUS was operated in a pulsed fashion, i.e. it was switched off for short periods to allow the diagnostic US to acquire a tracking image. Since the DiPhas does not allow for triggered image acquisition the FUS is switched off for the double US image acquisition time, ensuring the acquisition of one artifact free image. With an 25Hz imaging frequency the minimum shut-off period was thus 80 ms. To respect the relatively fast motion of the presented phantom a tracking frequency of 3 Hz was chosen resulting in a duty fraction of 0.76, i.e. the FUS is switched on 76% of the time. A 3 Hz tracking frequency was considerably smaller than the possible 20 Hz position update frequencies of the FUS transducer. The tracking positions were extrapolated linearly to allow for arbitrary FUS update frequencies.

2.3 Setup and calibration

The system components were connected to a therapy control component, orchestrating the steering. Tracking and steering information was sent via UDP network packages between the US tracking, therapy control, and FUS control. In order to estimate the target position for a given tracking datum the therapy control was calibrated to the setup at hand. MR images were acquired on a 1.5T scanner (Signa HDx, GE Medical Systems, Waukesha, USA) to determine the position and orientation of the FUS transducer and to define the target position. MR imaging in conjunction with the tracking coordinates was used to calibrate the translational motion model. The phantoms breathing motion was halted at two arbitrary but different positions in the breathing cycle, preferably at the positions that represent full inspiration and full expiration. The target position in those respiratory states was manually picked in the two MR images and stored with the tracking information at that points. Following the assumption of a translational motion the target positions in MR coordinates were estimated from the tracking positions.

2.4 Evaluation

The ability of the system to compensate motion was assessed by recording the temperature increase in the target spot. Specifically, the recordings were used to investigate if the temperature increase in a moving phantom resembles the temperature evolution in a static scenario. Since the whole setup was MR compatible MR thermometry for evaluating temperature can be used. However, this method is an indirect measure of temperature with limitations in spatial and temporal resolution. While MR thermometry will be the monitoring method of choice for clinical applications, the phantom case allows for a more direct measurement of temperature as a reference for thermometry. A PTFE fibre optical thermocouple (FOTEMP-4, OPTOCON, Dresden, Germany) was placed into the phantom to record the temperature every second. The tip of the thermocouple was used to define the target for sonication. To evaluate the system’s ability to compensate motion, pairs of 20 second long sonications were applied, each with precisely the same parameters, except one time with a moving phantom and one time with a halted (static) phantom. The static position of arrest was chosen to be full expiration, i.e. the ventilator of the phantom was switched off. To quantify the quality of motion compensation two measures were employed: The (Pearson) correlation coefficient (CC) was evaluated to give an estimate of the qualitative similarity of the temperature curves and the average temperature difference (ATD) was calculated as a measure of quantitative correlation. The system was evaluated with a number of experiment pairs with varying output power. The tracking frequency is fixed to 3 Hz in all experiments and the focal spot position was updated with 9Hz. For each output power in{ 15W, 30 W, 45W} a sonication within a moving and a static phantom is conducted.

3 Results

The employed US tracking worked reliably throughout experiment and the defined markers could be used over the course of the whole session. Table 2 shows the calculated CC and ATD values comparing the static and moving scenarios of the experiments.

4 Discussion and conclusion

The presented phantom exhibits motion comparable to human breathing in range and speed and is thus suitable for evaluating systems that aim at compensating for such motion. The presented steering system was able to compensate for the phantom motion as demonstrated by the good correlation of temperature curves acquired while sonicating in moving vs. static phantom. Only small (< 1.2°C) ATD values are observed, whereas the CC values are very close to 1. This was achieved by tracking the phantom motion via US images in gaps between sonications and employing a linear motion model to update the target position. However, real liver motion is more complex due to deformations and 3-dimensional motion of the organ and is influenced by perfusion. To compensate motion in a deformable liver, the linear motion model would need to be exchanged to one reflecting deformable organ motion, e.g. [7]. Since an active FUS renders the US imaging unusable, the system has an immanent trade-off between US imaging time and duty cycle. There are two mitigations to this: First, with an US device allowing for triggered image acquisition, the FUS device would have to to be switched off only half the time. Second, new advances in US imaging might lower the image acquisition time to a few milliseconds, improving the duty fraction of the presented system to 0.9with a 10 Hz tracking update. Since the whole setup is MR-compatible, MR thermometry can be employed for non-invasive temperature measurements in scenarios where thermocouples are inapplicable, e.g invivo experiments. In a next step, we are implementing per-fusion of the ex vivo liver to complete the phantom set up.

Table 2

Correlation coeflcient (CC) and average temperature difference (ATD) of temperature curves acquired in static and moving experiments for different output powers

Output powerCCATD in°C


The research leading to these results has received funding from the European Community’s Seventh Framework Programme FP7/2007-2013 under grant agreement n° 611889. The infrastructure at IMSaT was supported by Northern Research Partnership of Scotland.

Author's Statement

  1. Conflict of interest: Authors state no conflict of interest. Material and Methods: Informed consent: Informed consent has been obtained from all individuals included in this study. Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies and in accordance the tenets of the Helsinki Declaration, and has been approved by the authors’ institutional review board or equivalent committee.


[1] V. Auboiroux, L. Petrusca, M. Viallon, et al. Ultrasonography-based 2D motion-compensated HIFU sonication integrated with reference-free MR temperature monitoring: a feasibility study ex vivo.Physics in Medicine and Biology, 57(10):N159–71, 2012.10.1088/0031-9155/57/10/N159Search in Google Scholar PubMed

[2] J.-F. Aubry, K. Pauly, C. Moonen, et al. The road to clinical use of high-intensity focused ultrasound for liver cancer: technical and clinical consensus.Journal of Therapeutic Ultrasound, 1(1):13, 2013.10.1186/2050-5736-1-13Search in Google Scholar PubMed PubMed Central

[3] J. L. Foley, M. Eames, J. Snell, et al. Image-guided focused ultrasound: state of the technology and the challenges that lie ahead.Imaging in Medicine, 5(4):357–370, 2013.10.2217/iim.13.38Search in Google Scholar

[4] A. B. Holbrook, P. Ghanouni, et al. Respiration based steering for high intensity focused ultrasound liver ablation.Magnetic Resonance in Medicine, 806:797–806, 2013.10.1002/mrm.24695Search in Google Scholar PubMed PubMed Central

[5] F. Marquet, J. F. Aubry, M. Pernot, et al. Optimal transcostal high-intensity focused ultrasound with combined real-time 3D movement tracking and correction.Physics in Medicine and Biology, 56(22):7061–80, 2011.10.1088/0031-9155/56/22/005Search in Google Scholar PubMed

[6] J. C. Park, S. H. Park, J. H. Kim, et al. Liver motion during cone beam computed tomography guided stereotactic body radiation therapy.Medical Physics, 39(10):6431–6442, 2012.10.1118/1.4754658Search in Google Scholar PubMed

[7] G. Samei, C. Tanner, and G. Székely. Predicting liver motion using exemplar models. In H. Yoshida, D. Hawkes, and M. Vannier, editors,Abdominal Imaging. Computational and Clinical Applications, volume 7601 ofLecture Notes in Computer Science, pages 147–157. Springer Berlin Heidelberg, 2012.10.1007/978-3-642-33612-6_16Search in Google Scholar

[8] J. Schwaab, M. Prall, C. Sarti, et al. Ultrasound tracking for intra-fractional motion compensation in radiation therapy. Physica Medica, 30(5):578 – 582, 2014. Particle Radiosurgery Conference.10.1016/j.ejmp.2014.03.003Search in Google Scholar PubMed

Published Online: 2015-9-12
Published in Print: 2015-9-1

© 2015 by Walter de Gruyter GmbH, Berlin/Boston

This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Downloaded on 2.3.2024 from
Scroll to top button