Head phantoms for transcranial focused ultrasound Matthew D. C. Eames, Mercy Farnum, Mohamad Khaled, W. Jeff Elias, Arik Hananel, John W. Snell, Neal F. Kassell, and Jean-Francois Aubry Citation: Medical Physics 42, 1518 (2015); doi: 10.1118/1.4907959 View online: http://dx.doi.org/10.1118/1.4907959 View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/42/4?ver=pdfcov Published by the American Association of Physicists in Medicine Articles you may be interested in Dual-frequency ultrasound focal therapy for MRI-guided transurethral treatment of the prostate: Study in gel phantom AIP Conf. Proc. 1503, 212 (2012); 10.1063/1.4769946 Effects of human hair on trans-cranial focused ultrasound efficacy in an ex-vivo cadaver model AIP Conf. Proc. 1503, 195 (2012); 10.1063/1.4769943 Ultrasound focusing using magnetic resonance acoustic radiation force imaging: Application to ultrasound transcranial therapy Med. Phys. 37, 2934 (2010); 10.1118/1.3395553 Magnetic resonance imaging of boiling induced by high intensity focused ultrasound J. Acoust. Soc. Am. 125, 2420 (2009); 10.1121/1.3081393 A multimodality vascular imaging phantom with fiducial markers visible in DSA, CTA, MRA, and ultrasound Med. Phys. 31, 1424 (2004); 10.1118/1.1739300
Head phantoms for transcranial focused ultrasound Matthew D. C. Eamesa) and Mercy Farnum Focused Ultrasound Foundation, Charlottesville, Virginia 22903
Mohamad Khaled and W. Jeff Elias Department of Neurosurgery, University of Virginia, Charlottesville, Virginia 22908
Arik Hananel Focused Ultrasound Foundation, Charlottesville, Virginia 22903 and Department of Radiation Oncology, University of Virginia, Charlottesville, Virginia 22908
John W. Snell and Neal F. Kassell Focused Ultrasound Foundation, Charlottesville, Virginia 22903 and Department of Neurosurgery, University of Virginia, Charlottesville, Virginia 22908
Jean-Francois Aubry Department of Radiation Oncology, University of Virginia, Charlottesville, Virginia 22908 and Institut Langevin, ESPCI ParisTech, CNRS UMR 7587, INSERM U979, Paris 75005, France
(Received 16 May 2014; revised 18 November 2014; accepted for publication 13 January 2015; published 13 March 2015) Purpose: In the ongoing endeavor of fine-tuning, the clinical application of transcranial MR-guided focused ultrasound (tcMRgFUS), ex-vivo studies wlkiith whole human skulls are of great use in improving the underlying technology guiding the accurate and precise thermal ablation of clinically relevant targets in the human skull. Described here are the designs, methods for fabrication, and notes on utility of three different ultrasound phantoms to be used for brain focused ultrasound research. Methods: Three different models of phantoms are developed and tested to be accurate, repeatable experimental options to provide means to further this research. The three models are a cadaver, a gel-filled skull, and a head mold containing a skull and filled with gel that mimics the brain and the skin. Each was positioned in a clinical tcMRgFUS system and sonicated at 1100 W (acoustic) for 12 s at different locations. Maximum temperature rise as measured by MR thermometry was recorded and compared against clinical data for a similar neurosurgical target. Results are presented as heating efficiency in units (◦C/kW/s) for direct comparison to available clinical data. The procedure for casting thermal phantom material is presented. The utility of each phantom model is discussed in the context of various tcMRgFUS research areas. Results: The cadaveric phantom model, gel-filled skull model, and full head phantom model had heating efficiencies of 5.3, 4.0, and 3.9 ◦C/(kW/s), respectively, compared to a sample clinical heating efficiency of 2.6 ◦C/(kW/s). In the seven research categories considered, the cadaveric phantom model was the most versatile, though less practical compared to the ex-vivo skull-based phantoms. Conclusions: Casting thermal phantom material was shown to be an effective way to prepare tissue-mimicking material for the phantoms presented. The phantom models presented are all useful in tcMRgFUS research, though some are better suited to a limited subset of applications depending on the researchers needs. C 2015 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4907959] Key words: head phantom, focused ultrasound, tcMRgFUS, brain, treatment envelope
1. INTRODUCTION Focused ultrasound (FUS), or high-intensity focused ultrasound (HIFU), is a technology in which ultrasonic energy emitted from a transducer is focused at a target volume.1 The treatment is monitored either through magnetic resonance imaging (MRgFUS: MR-guided FUS) or ultrasound imaging (USgFUS: US-guided FUS).2 The effects at the focal area can vary from thermal to mechanical depending on the procedure.3 Currently, the most common clinical usage of FUS is noninvasive thermal ablation. This particular biological effect has 1518
Med. Phys. 42 (4), April 2015
been widely used to treat the indications of prostate cancer,4,5 liver cancer,6,7 breast,8 and symptomatic uterine fibroids.9,10 The neurological applications of FUS have been more problematic owing to the reflection, absorption, and refraction of the ultrasonic waves by the skull. However, the recent development of aberration correction techniques to compensate for this difficulty11,12 has led to the development of transcranial MRgFUS (tcMRgFUS) devices.13–15 Currently, it is used clinically as a functional neurosurgery tool for the treatment of essential tremor16,17 and neuropathic pain.18 However, the noninvasive, radiation-free, and precise properties of
0094-2405/2015/42(4)/1518/10/$30.00
© 2015 Am. Assoc. Phys. Med.
1518
1519
Eames et al.: Head phantoms for transcranial focused ultrasound
tcMRgFUS would make it an ideal candidate for many other brain indications such as Parkinson’s disease, trigeminal neuralgia, epilepsy, and brain tumors. Novel techniques are under development in laboratories in order to improve the aberration correction techniques,19–21 and to image the location of the ultrasonic beam with low energy deposition using acoustic radiation force imaging.22–25 In order to further expand the treatment envelope—which is to say to increase the volume of the brain throughout which a single focal spot of sufficient intensity to achieve therapeutic effect can be achieved—and to translate to clinic novel guidance modalities, studies need to be conducted using ex-vivo skulls to accurately replicate the ultrasonic distortions induced by the skull in neurological treatment procedures. To advance this process, we tested three head phantom models in order to provide accurate and repeatable options for conducting the necessary experiments. The first phantom model is a cadaver. The second model is a gel-filled skull. The third model is a plastic mannequin head model into which a skull has been placed filled with the same tissue-mimicking material used in the second model, referred to here as the full head phantom. A total of nine phantoms were built and tested in a tcMRgFUS system.
2. METHODS 2.A. Cadaver
Full human cadavers were obtained from Virginia’s State Anatomical Program through the University of Virginia department of Neurosurgery. All specimens provided by this mechanism were full organ donors at the time of death. For each cadaver, a CT scan of the head was taken for the purposes of the skull correction algorithm. Trans-skull sonications could then be performed under MR image-guidance in the same manner as clinical tcMRgFUS treatments. 2.B. Description and preparation of tissue-mimicking phantom hydrogel
The phantom material used to mimic human tissue was a tissue-mimicking hydrogel (ATS Laboratories, Bridgeport CT). Acoustical and thermal properties of the hydrogel are summarized in Table I and compared to a previously published
1519
milk powder based phantom26 and to measurements in brain27 and skin.27 In order to pour the phantom, the hydrogel was first liquefied by heating to 55 ◦C in a conventional microwave oven dedicated to research use. 2.C. Gel-filled skull
The skull of previous cadaver was first defleshed and cleaned by the department of neurosurgery. A portion of the vault of the skull (later referred to as the skull vault) was cut away from the skull on a plane above the eye orbits and above the external occipital protuberance. Holes were drilled into the skull vault using a drill (3000 Variable Speed Rotary Tool, Dremel, Mount Prospect, IL) on the sides of the skull and the back and holes were also drilled into the skull base adjacent to those drilled into the skull vault. The insides of the skull cap and skull base were then lined with plastic film to create a watertight mold into which the gel was poured and cooled in a refrigerator to approximately 5 ◦C for 8 h to solidify. Once the gel had solidified, the plastic was peeled off and both gels were stored in the fridge. The two gels formed two halves of the brain tissue-mimicking phantom that fit closely to the inside surface of the skull (Fig. 1). The human ex-vivo skull was then degassed in water for at least 2 h in a vacuum chamber (Acrylic Round Vacuum Chamber, Abbess, Holliston, MA) at 230 mmHg. Once degassed, the cured gels for the skull vault and skull base were gently placed within the respective portions of the skull. The two halves of the skull were connected together while submerged in water by threading cable ties through the holes in the skull base and vault and connecting the two together. The skull was then attached to a base frame using cable ties, as shown in Fig. 2. The skull was then removed from the container and set up in the midfrequency ExAblate-Neuro system (InSightec, Tirat Hakarmel, Israel), which operates at a central frequency of 710 kHz. The setup was then promptly filled with degassed water to minimize oxygen exposure. The phantom fit is sufficiently close to trap a thin film of water between the gel and the inside surface of the skull, allowing for experimental setup maneuverability in air before submerging the setup in the transducer water bath in preparation for sonication. The quality of fit of this phantom in
T I. Acoustical and thermal properties of the brain (Ref. 27) and skin (Ref. 27), of the tissue-mimicking hydrogel (Ref. 28), and of a milk powder based phantom (Ref. 26). N/A: not available. Brain and skin tissue correspond to human, otherwise stated.
Speed of sound (m s−1)
Skin (Ref. 27)
Brain (Ref. 27)
1537 @ 17.5 ◦C
1532 @ 22 ◦C 1562 @ 37 ◦C 0.28 (gray matter) @ 37 ◦C 0.48 (white matter) @ 37 ◦C 3680 (gray matter) 3600 (white matter) 0.478 ± 0.015
Attenuation (dB cm−1 MHz−1)
3.5 ± 1.2 @ 23 ◦C
Specific heat capacity (J kg−1 K−1)
3215–3280 (pig)
Thermal conductivity (W m−1 K−1)
0.293 ± 0.016
Medical Physics, Vol. 42, No. 4, April 2015
Tissue-mimicking hydrogel (Ref. 28)
Milk powder based phantom (Ref. 26)
1537 @ 23 ◦C
1625 @ N/A◦C
0.5 ± 0.2 @ 23 ◦C
0.46 @ N/A◦C
3500 ± 500
N/A
0.5 ± 0.1
N/A
1520
Eames et al.: Head phantoms for transcranial focused ultrasound
1520
F. 1. From left to right: cured gel, skull, and gel fitted in the skull. Upper row displays the base and the lower row displays the vault.
an ex-vivo human skull is illustrated in Fig. 3 (center and right). 2.D. Full head phantom
First, a portion of the top of a plastic mannequin head (Fig. 4, left; Unisex head, Hard plastic clear, Inflatable Mannequins, Lindon, UT) was cut away. A complete human skull (Skulls Unlimited, Oklahoma City, OK) was then inserted into the mannequin head and settled into place (Fig. 4, center). The
removed piece from the top of the mannequin head was then reattached and sealed using tape (Fig. 4, right). The model was then inverted, immersed in the same degassing water tank displayed in Fig. 2 left, and digassed for 2 h. Immediately after removing the model from the degassing tank, water was evacuated and tissue-mimicking hydrogel (prepared as previously specified) was poured into mannequin head until the skull had been completely immersed. The mannequin mold was then wrapped in plastic film and secured in a container filled with cold water in order to facilitate the
F. 2. Degassing of the cadaver skull (left), tissue phantom mold beside skull (upper left), and gel-filled skull set up in frame (lower right). Medical Physics, Vol. 42, No. 4, April 2015
1521
Eames et al.: Head phantoms for transcranial focused ultrasound
1521
F. 3. Bare skull strapped to holder (left), and as seen on MR T2w sagittal (center), and coronal (right) image, which illustrates the quality of the fit of the phantom.
solidification of the gel (Fig. 5, left). The plastic film prevented the leakage of phantom from the model by allowing the water pressure to counteract the weight of the phantom in the model. It also prevented the water from leaking into the model and potentially diluting the tissue-mimicking gel or loosening the tape. The setup was left in place overnight in a refrigerator to allow the gel to cool and solidify. The mold was then removed and the plastic dried off. A CT scan was then taken of the completed model in order to provide the baseline data needed for the skull correction sequence (Fig. 6). The model was then set up in the ExAblate-Neuro transducer in the same manner that a patient’s head would be (Fig. 7). 2.E. Planning
Phantoms were positioned in the midfrequency ExAblateNeuro brain system. CT images of the skulls were aligned manually with T2 MR images with the InSightec planning software. The target was chosen at the location of the ventral intermediate nucleus (VIM), in the thalamus. The targeting
was performed on T2 images of the brain for the cadavers (Fig. 8, left). As the gel-filled skull model was poured in the same skull as the cadaver (Fig. 8, right), T2 images of the cadaver brain were fused to allow MR-based targeting of the VIM. Concerning the full head phantom, the VIM was located based on the skull anatomy landmarks on MR and CT. Compared to humans and cadavers, frontal sinuses are filled with gel in all our phantoms. Nevertheless, the clinical procedure needs to be applied to all phantoms, including the required drawing of “nopass regions” corresponding to sinuses so that the InSightec planning software disables ultrasound elements whose output would pass primarily through the sinuses. This planning step is mandatory for clinical sonication in order to avoid ultrasonic reflection on the air and consequent overheating, and it is recommended when using the phantoms. In both cases, no ultrasound beam will go through the sinuses, mitigating the fact that the phantom sinuses do not comprise air. 2.F. MR thermometry
Before, during, and after energy deposition, gel thermal rise in the focal point was evaluated using the proton resonance
F. 4. From left to right: the intact head mold, the mold with the upper portion removed and skull inserted, and the mold containing the skull with the upper portion reattached with tape. Medical Physics, Vol. 42, No. 4, April 2015
1522
Eames et al.: Head phantoms for transcranial focused ultrasound
1522
F. 5. The mold immersed in water (left) and the completed mold (right).
frequency (PRF) shift method of MR temperature mapping. The MR system was a 3T Discovery 750 (GE, Milwaukee, WI). The MR thermometry scan parameters used were parameters: TR/TE 27.6/12.8 ms, flip angle 30◦, bandwidth 5.68 kHz, FOV 28 cm, slice thickness 3 mm, matrix 256×128, scan time 3 s. A temperature sensitivity of −0.009 ppm/◦C was used29 and the initial temperature was arbitrarily set to 37 ◦C for all measurements.
full head phantom made out of the intact skull bone from Skulls Unlimited®. Thermal rises measured during the clinical essential tremor showed an average 2.59 ± 1.13 ◦C/kJ thermal rise in the patient’s VIM. It is lower than the ones observed in the cadaveric phantoms (4.85 ± 1.18 ◦C/kJ) and the gel phantoms (3.98 ± 0.27 ◦C/kJ).
4. DISCUSSION 2.G. Relevance of the head phantoms in regard to clinical data
Recently, a MRgFUS pilot study has been conducted at the University of Virginia on 15 drug resistant essential tremor patients.16 For each patient, the location of the focus was first adjusted and the power was ramped up until a lesion could be seen on T2 images, with an average of 18 successive sonications. We processed the corresponding anonymized data to extract the maximum temperature elevation induced in the VIM during the procedure. For all patients, we selected only the highest power sonication associated with an axial orientation plane for MR temperature monitoring.
3. RESULTS Table II summarizes all the temperature elevations induced by the ultrasonic beam at the VIM location in four different cadavers, in the four corresponding gel-filled skulls and in the
The cadaver model geometry is the most similar to the clinical environment: not only it includes air-filled sinuses, skin, and hair, but the degrees of freedom of the head phantom are constrained by its attachment to the patient body; in particular, the limited three-axis rotation of the head constrains the location of the transducer in regard to the head to clinically realistic positions. The anatomy is similar to that of a patient, the main difference being that perfusion is absent. Nevertheless, access to cadavers is not straightforward, and the model can only be used for a limited time period, typically one or two days. Gel-filled head phantoms advantageously provide a skull-specific, reusable phantom. Even though they do not reproduce the heterogeneities of the brain structures and of the skin fat and muscles surrounding the head, they are made of real human skull. The presence of the skull has long hampered the clinical implementation of focused ultrasound. The skullmimicking structure is thus a key component to be able to test new techniques for MRgFUS. This is the reason why no compromise was made on the skull itself. Taking advantage
F. 6. The full head phantom. Left: positioned in the CT scanner. Right: sagittal MR slice. Medical Physics, Vol. 42, No. 4, April 2015
1523
Eames et al.: Head phantoms for transcranial focused ultrasound
1523
F. 7. The setup for the full head phantom in the ExAblate-Neuro transducer.
of recent developments in composite material design, Wydra and Maev mimicked most of the acoustical properties of human bones.30 The human skulls used in this study could thus be replaced by composite material in the future. An alternative to the commercially available hydrogel (ATS Laboratories, Bridgeport CT) used in this study has also been proposed by Maruvada et al.31 The corresponding recipe can be found in King et al.32 The phantoms could be used to investigate various novel developments for tcMRgFUS ranging from checking skin heating, to developing new MR thermal mapping sequences in order to perform faster imaging or larger field of view. Some of the applications are listed in Table III and the appropriateness of the phantoms presented in this paper is rated. 4.A. Treatment envelope
Extending the treatment envelope consists of optimizing the target volume that can be treated with tcMRgFUS.33 Current treatments are limited to the thalamus. But promising treatments could be envisioned for epilepsy or trigeminal neu-
ralgia34 if the system was shown to be able to efficiently and safely deliver energy at other locations in the brain. The phantoms could be used to compare the maximum temperature elevation and the temperature distribution, as compared to the one obtained in the thalamus. For such a targeting study, the ability to visualize the anatomy of the brain is crucial. The cadaveric phantom model is the only one presented in this study that could provide MR imagery of brain anatomy, including trigeminal nerve, hippocampus, globus pallidus, or anterior/posterior commissure for determination of the VIM location. The tissue-mimicking phantom can be used for investigating the treatment envelope only if the skull has been extracted from a cadaver from which a T2 image has been acquired and that can be superimposed on the CT image of the skull in the InSightec planning software, as presented in Sec. 2.E. 4.B. MR-ARFI spot localization
While in current clinical trials, low-temperature thermal sonications are performed to assess focal spot quality and
F. 8. Target planning in a cadaver model (left) and in a gel-filled skull (right). Medical Physics, Vol. 42, No. 4, April 2015
1524
Eames et al.: Head phantoms for transcranial focused ultrasound
1524
T II. Summary of tcMRgFUS temperature measurements. Type
◦
Subject number
VIM side
Power (W)
Duration (s)
Tmax ( C)
∆Tmax/kW/s
Cadaver Cadaver Cadaver Cadaver Cadaver Cadaver Mean Std
1 2 3 3 4 4
R R R L R L
1100 1100 1100 1100 1100 1100
12 12 12 12 12 12
61 60 72 70 81 77 70.2 8.4
1.8 1.7 2.7 2.5 3.3 3.0 2.5 0.6
Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Mean Std
1 1 2 2 3 3 4 4
R L R L R L R L
1100 1100 1100 1100 1100 1100 1100 1100
12 12 12 12 12 12 12 12
49 50 50 55 50 51 57 58 52.5 3.6
0.9 1.0 1.0 1.4 1.0 1.1 1.5 1.6 1.2 0.3
1100 125 250 525 550 650 650 650 700 750 900 1000 1025 1100 1300 1300
12 10 10 10 10 15 10 10 10 10 10 10 12 12 15 15
51 43 45 56 53 59 63 58 57 57 62 48 60 57 58 55 55.4 5.8
1.1 4.7 3.4 3.5 2.9 2.2 4.1 3.2 2.8 2.6 2.8 1.1 1.9 1.5 1.1 0.9 2.6 1.1
Full head phantom ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient Mean Std
A B C D E F G H I J K L M N O
targeting accuracy, research has demonstrated the feasibility of detecting ARFI-induced tissue displacement with MR as a nonthermal means of evaluating these focal characteristics.21,35,36 Any of the three presented phantoms could be used to test and refine this nonthermal technique, however, there is a common compromise in tissue or material stiffness that affects all phantoms models. Because the cadaveric tissue is both deceased and colder than body temperature, it is generally stiffer than living brain tissue. This serves to decrease the sensitivity of this technique because the same ARFI sonication results in a smaller tissue displacement in the phantom than would occur in a living patient. While the gel-based phantoms presented also suffer from an increased stiffness compared to living brain tissue, it is possible to dilute the tissue-mimicking gel as much as 1:1 with water to decrease stiffness and thereby increase MR-ARFI signal for spot localization. This method of diluting the phantom mixture would need to be calibrated to accurately mimic the stiffness of brain tissue. Medical Physics, Vol. 42, No. 4, April 2015
4.C. MR-ARFI aberration correction
While similar to the MR-ARFI technique above, the aberration correction implementation aims to improve or replace the CT-based aberration correction in order to achieve a tighter focal spot in the brain or to altogether eliminate the need for a CT scan prior to a tMRgFUS procedure. As above, the same limitations with phantom material stiffness apply, again suggesting the need to calibrate the stiffness of the tissuemimicking phantom material. 4.D. Investigation of skin heating
Assessing the degree of skin heating that occurs during the course of clinical tMRgFUS procedures is important in order to prevent skin burns. Of the phantoms presented, the cadaveric model is the best suited to these studies as it is the most accurate model of the skin, fat, and muscle layers on the skull. Additionally, the fact that the cadaver is colder than
1525
Eames et al.: Head phantoms for transcranial focused ultrasound
body temperature does not compromise the use of this model because the chilled water bath surrounding the patient’s head in such FUS systems ensures that the skin approaches a temperature equal to the chilled water. The full head phantom model could be a low-cost alternative for investigating skin heating as it provides a layer of tissue-mimicking phantom material exterior to the skull surface, while the gel-filled skull is not suitable for this application. Acoustical and thermal properties of the tissue-mimicking hydrogel are close to that of the brain (Table I), but are also similar to that of the skin, except for attenuation (Table I). The full head phantom would thus mimic more closely heat release from the skull toward surrounding tissues than ultrasonic absorption by the skin itself. 4.E. New MR thermometry sequences
The phantoms described could also be valuable tools for evaluating techniques to improve on existing thermometry sequences or to introduce altogether new sequences. While purely imaging studies could be carried out on healthy human volunteers, it is hard to imagine a scenario in which FUSinduced temperature rise measurements using experimental MR thermometry sequences would not, at least at some stage, require a phantom model. All three phantoms presented have in common the following disadvantages as compared to clinical reality: their baseline temperatures will be cooler than body temperature and they lack perfusion and pulsatile flow. Nevertheless, Pulkkinnen et al.37 performed numerical simulations of clinical focused ultrasound thalamotomy and showed that perfusion impacts temperature elevation by 4% only when adding blood perfusion in their model. Because they degrade over a period of weeks to months, the gel-filled skull and full head phantoms are more durable than the cadaver model. Between the gel-based phantoms, the full head phantom has the advantage of permitting thermometry in the skin portion of the phantom, prompting us to recommend its use for most MR thermometry sequence development. 4.F. Novel bone imaging techniques 38
Ultrashort TE (time-to-echo) MR sequences (UTE) are commonly used in research environments to image bony anatomy. In the context of focused ultrasound, MR bone imaging could replace CT as the default source of information on
1525
skull geometry needed for aberration correction in the clinic. Because UTE is currently a research-only MR pulse sequence, there is an interest in exploring novel bone imaging techniques for FUS that could immediately be made available for clinical use. The gel-filled skull does not provide MR-signal at the outer bone surface, and is therefore the least preferred option of the gel-based phantoms. The cadaveric phantom offers both human soft tissue (blood, fat, muscle, brain) and bone (including bone marrow within the dipole), providing the most realistic imaging phantom for early pulse sequence development in the presence of the nonfiring but possibly MRartifact-producing transducer (i.e., before initiating a study on healthy human volunteers). The disadvantage of the cadaver is its limited longevity as compared to the gel-based phantom, making it the preferred choice only when accurate soft-tissue anatomy is required. 4.G. Development of FUS-compatible MR coils
Imaging in MR is improved by anatomy-specific imaging coils, but the water bath and need for an acoustically transparent path between anatomical target and transducer and of a focused ultrasound system prevents the use of commercially available MR imaging coils in clinical applications. These head phantoms serve as tools to design and evaluate prototype MR coils. While the gel-based phantoms, with good intraphantom homogeneity, can be used for preliminary calculation of a coil’s SNR improvement over alternative MR imaging options, the gel was not specifically designed to mimic the electromagnetic properties of living tissue, so that coil performance would need to be assessed on volunteer to better assess the SNR improvement. Cadavers serve both to load MR coils in a more realistic manner and to guide prototype decisions by means of the anatomically correct limitations a patient’s neck or shoulders may place on design options. 4.H. Relevance of the head phantoms in regard to clinical data
An average 2.6 ± 1 ◦C/kJ thermal rise was induced in the patient’s VIM. It is lower than the ones observed in the cadaveric phantoms (4.8 ± 1.2 ◦C/kJ) and the gel phantoms (4 ± 0.3 ◦C/kJ). All thermal rises per unit of time and acoustical power (∆Tmax/kW/s in Table II) obtained either with one
T III. Comparison of the best indication of each phantom for various applications. Application
Cadaver
Gel-filled skull
Full head gel phantom
Treatment envelope
+++
MR-ARFI spot localization MR-ARFI aberration correction Investigation of skin heating New MR thermometry sequences Novel bone imaging techniques Development of FUS-compatible MR coils
++ ++ +++ ++ + +++
++ (if skull extracted from a cadaver) ++ ++ − ++ ++ ++
++ (if skull extracted from a cadaver) ++ ++ + +++ ++ ++
Medical Physics, Vol. 42, No. 4, April 2015
1526
Eames et al.: Head phantoms for transcranial focused ultrasound
of the phantoms or measured during clinical essential tremor treatments overlap within the standard deviation, indicating that the models are relevant for preclinical developments. In the case of specific preclinical studies for which absolute thermal rise is of interest, such as treatment envelop extension, a 50% correction factor would need to be applied. 5. CONCLUSIONS This method for casting a commercially available thermal phantom material has proven to be an efficient, straightforward, and effective way to prepare head phantoms for tcMRgFUS studies. Peak thermal rises have been reported here in the phantoms when targeting the VIM and have been compared to clinical data on essential tremor treatments with the same anatomical target. Variations of the methods might be used for other clinically relevant preclinical or technical studies, making this technique relatively portable across anatomical applications—provided the gel is casted with a relevant geometry30—including breast, bone, or prostate. a)Author
to whom correspondence should be addressed. Electronic mail: [email protected] 1J. E. Kennedy, G. R. ter Haar, and D. Cranston, “High intensity focused ultrasound: Surgery of the future?,” Br. J. Radiol. 76, 590–599 (2003). 2J. L. Foley, M. Eames, J. Snell, A. Hananel, N. Kassell, and J.-F. Aubry, “Image-guided focused ultrasound: State of the technology and the challenges that lie ahead,” Imaging Med. 5, 357–370 (2013). 3G. ter Haar, “Ultrasound bioeffects and safety,” Proc. Inst. Mech. Eng., Part H 224, 363–373 (2010). 4S. Crouzet, X. Rebillard, D. Chevallier, P. Rischmann, G. Pasticier, G. Garcia, O. Rouviere, J. Y. Chapelon, and A. Gelet, “Multicentric oncologic outcomes of high-intensity focused ultrasound for localized prostate cancer in 803 patients,” Eur. Urol. 58, 559–566 (2010). 5Y. Inoue, K. Goto, T. Hayashi, and M. Hayashi, “Transrectal high-intensity focused ultrasound for treatment of localized prostate cancer,” Int. J. Urol. 18, 358–363 (2011). 6J.-F. Aubry, K. Pauly, C. Moonen, G. t. Haar, M. Ries, R. Salomir, S. Sokka, K. Sekins, Y. Shapira, F. Ye, H. Huff-Simonin, M. Eames, A. Hananel, N. Kassell, A. Napoli, J. Hwang, F. Wu, L. Zhang, A. Melzer, Y.-s. Kim, and W. Gedroyc, “The road to clinical use of high-intensity focused ultrasound for liver cancer: Technical and clinical consensus,” J. Ther. Ultrasound 1:13 (2013). 7F. Wu, Z. B. Wang, W. Z. Chen, W. Wang, Y. Gui, M. Zhang, G. Zheng, Y. Zhou, G. Xu, M. Li, C. Zhang, H. Ye, and R. Feng, “Extracorporeal high intensity focused ultrasound ablation in the treatment of 1038 patients with solid carcinomas in China: An overview,” Ultrason. Sonochem. 11, 149–154 (2004). 8H. Furusawa, K. Namba, S. Thomsen, F. Akiyama, A. Bendet, C. Tanaka, Y. Yasuda, and H. Nakahara, “Magnetic resonance–guided focused ultrasound surgery of breast cancer: Reliability and effectiveness,” J. Am. Coll. Surg. 203, 54–63 (2006). 9J. Rabinovici, M. David, H. Fukunishi, Y. Morita, B. S. Gostout, and E. A. Stewart, “Pregnancy outcome after magnetic resonance-guided focused ultrasound surgery (MRgFUS) for conservative treatment of uterine fibroids,” Fertil. Steril. 93, 199–209 (2008). 10E. A. Stewart, J. Rabinovici, C. M. Tempany, Y. Inbar, L. Regan, B. Gostout, G. Hesley, H. S. Kim, S. Hengst, and W. M. Gedroyc, “Clinical outcomes of focused ultrasound surgery for the treatment of uterine fibroids,” Fertil. Steril. 85, 22–29 (2006). 11J. F. Aubry, M. Tanter, M. Pernot, J. L. Thomas, and M. Fink, “Experimental demonstration of noninvasive transskull adaptive focusing based on prior computed tomography scans,” J. Acoust. Soc. Am. 113, 84–93 (2003). 12K. Hynynen and J. Sun, “Trans-skull ultrasound therapy: The feasibility of using image-derived skull thickness information to correct the phase Medical Physics, Vol. 42, No. 4, April 2015
1526
distortion,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control 46, 752–755 (1999). 13K. Hynynen, N. Mcdannold, G. Clement, F. A. Jolesz, E. Zadicario, R. Killiany, T. Moore, and D. Rosen, “Pre-clinical testing of a phased array ultrasound system for MRI-guided noninvasive surgery of the brain–a primate study,” Eur. J. Radiol. 59, 149–156 (2006). 14W. J. Elias, M. Khaled, J. D. Hilliard, J.-F. Aubry, R. C. Frysinger, J. P. Sheehan, M. Wintermark, and M. B. Lopes, “A magnetic resonance imaging, histological, and dose modeling comparison of focused ultrasound, radiofrequency, and Gamma Knife radiosurgery lesions in swine thalamus,” J. Neurosurg. 119, 307–317 (2013). 15D. Chauvet, L. Marsac, M. Pernot, A.-L. Boch, R. Guillevin, N. Salameh, L. Souris, L. Darrasse, M. Fink, and M. Tanter, “Targeting accuracy of transcranial magnetic resonance–guided high-intensity focused ultrasound brain therapy: A fresh cadaver model: Laboratory investigation,” J. Neurosurg. 118(5), 1046–1052 (2013). 16W. J. Elias, D. Huss, T. Voss, J. Loomba, M. Khaled, E. Zadicario, R. C. Frysinger, S. A. Sperling, S. Wylie, S. J. Monteith, J. Druzgal, B. B. Shah, M. Harrison, and M. Wintermark, “A pilot study of focused ultrasound thalamotomy for essential tremor,” N. Engl. J. Med. 369, 640–648 (2013). 17N. Lipsman, M. L. Schwartz, Y. Huang, L. Lee, T. Sankar, M. Chapman, K. Hynynen, and A. M. Lozano, “MR-guided focused ultrasound thalamotomy for essential tremor: A proof-of-concept study,” Lancet Neurol. 12, 462–468 (2013). 18D. Jeanmonod, B. Werner, A. Morel, L. Michels, E. Zadicario, G. Schiff, and E. Martin, “Transcranial magnetic resonance imaging-guided focused ultrasound: Noninvasive central lateral thalamotomy for chronic neuropathic pain,” Neurosurg. Focus 32, 1–11 (2012). 19J. Gateau, L. Marsac, M. Pernot, J. F. Aubry, M. Tanter, and M. Fink, “Transcranial ultrasonic therapy based on time reversal of acoustically induced cavitation bubble signature,” IEEE Trans. Biomed. Eng. 57, 134–144 (2010). 20E. A. Kaye, Y. Hertzberg, M. Marx, B. Werner, G. Navon, M. Levoy, and K. B. Pauly, “Application of Zernike polynomials towards accelerated adaptive focusing of transcranial high intensity focused ultrasound,” Med. Phys. 39, 6254–6263 (2012). 21L. Marsac, D. Chauvet, B. Larrat, M. Pernot, B. Robert, M. Fink, A.-L. Boch, J.-F. Aubry, and M. Tanter, “MR-guided adaptive focusing of therapeutic ultrasound beams in the human head,” Med. Phys. 39, 1141–1149 (2012). 22E. A. Kaye, J. Chen, and K. B. Pauly, “Rapid MR—ARFI method for focal spot localization during focused ultrasound therapy,” Magn. Reson. Med. 65, 738–743 (2011). 23E. Dervishi, B. Larrat, M. Pernot, C. Adam, Y. Marie, M. Fink, J.-Y. Delattre, A.-L. Boch, M. Tanter, and J.-F. Aubry, “Transcranial high intensity focused ultrasound therapy guided by 7 TESLA MRI in a rat brain tumour model: A feasibility study,” Int. J. Hyperthermia 29, 598–608 (2013). 24N. Mcdannold and S. E. Maier, “Magnetic resonance acoustic radiation force imaging,” Med. Phys. 35, 3748–3758 (2008). 25R. Paquin, A. Vignaud, L. Marsac, Y. Younan, S. Lehéricy, M. Tanter, and J.-F. Aubry, “Keyhole acceleration for magnetic resonance acoustic radiation force imaging (MR ARFI),” Magn. Reson. Imaging 31, 1695–1703 (2013). 26N. Mcdannold, G. T. Clement, P. Black, F. Jolesz, and K. Hynynen, “Transcranial magnetic resonance imaging–guided focused ultrasound surgery of brain tumors: Initial findings in 3 patients,” Neurosurgery 66, 323–332; discussion 332 (2010). 27F. A. Duck, Physical Properties of Tissues: A Comprehensive Reference Book (Academic, San Diego, CA, 1990). 28ATS Laboratories (2014), see http://www.atslaboratories-phantoms.com/ page8/index.html. 29T. Wu, K. R. Kendell, J. P. Felmlee, B. D. Lewis, and R. L. Ehman, “Reliability of water proton chemical shift temperature calibration for focused ultrasound ablation therapy,” Med. Phys. 27, 221–224 (2000). 30A. Wydra and R. G. Maev, “A novel composite material specifically developed for ultrasound bone phantoms: Cortical, trabecular and skull,” Phys. Med. Biol. 58, N303–N319 (2013). 31S. Maruvada, Y. Liu, W. Pritchard, B. Herman, and G. Harris, “Comparative study of temperature measurements in ex vivo swine muscle and a tissuemimicking material during high intensity focused ultrasound exposures,” Phys. Med. Biol. 57, 1–19 (2012). 32R. L. King, Y. Liu, S. Maruvada, B. A. Herman, K. A. Wear, and G. R. Harris, “Development and characterization of a tissue-mimicking material
1527
Eames et al.: Head phantoms for transcranial focused ultrasound
for high-intensity focused ultrasound,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control 58(7), 1397–1405 (2011). 33M. D. C. Eames, A. Hananel, N. F. Kassell, and J. W. Snell, “Intracranial treatment envelope mapping of transcranial focused ultrasound,” AIP Conf. Proc. 1503, 181–184 (2012). 34S. Monteith, J. Sheehan, R. Medel, M. Wintermark, M. Eames, J. Snell, N. F. Kassell, and W. J. Elias, “Potential intracranial applications of magnetic resonance–guided focused ultrasound surgery,” J. Neurosurg. 118, 215–221 (2013). 35Y. Hertzberg, A. Volovick, Y. Zur, Y. Medan, S. Vitek, and G. Navon, “Ultrasound focusing using magnetic resonance acoustic radiation force
Medical Physics, Vol. 42, No. 4, April 2015
1527
imaging: Application to ultrasound transcranial therapy,” Med. Phys. 37, 2934–2942 (2010). 36E. A. Kaye, J. Chen, and K. B. Pauly, “Rapid MR-ARFI method for focal spot localization during focused ultrasound therapy,” Magn. Reson. Med. 65, 738–743 (2011). 37A. Pulkkinen, B. Werner, E. Martin, and K. Hynynen, “Numerical simulations of clinical focused ultrasound functional neurosurgery,” Phys. Med. Biol. 59, 1679–1700 (2014). 38M. D. Robson, P. D. Gatehouse, M. Bydder, and G. M. Bydder, “Magnetic resonance: An introduction to ultrashort TE (UTE) imaging,” J. Comput. Assisted Tomogr. 27, 825–846 (2003).
Head phantoms for transcranial focused ultrasound Matthew D. C. Eamesa) and Mercy Farnum Focused Ultrasound Foundation, Charlottesville, Virginia 22903
Mohamad Khaled and W. Jeff Elias Department of Neurosurgery, University of Virginia, Charlottesville, Virginia 22908
Arik Hananel Focused Ultrasound Foundation, Charlottesville, Virginia 22903 and Department of Radiation Oncology, University of Virginia, Charlottesville, Virginia 22908
John W. Snell and Neal F. Kassell Focused Ultrasound Foundation, Charlottesville, Virginia 22903 and Department of Neurosurgery, University of Virginia, Charlottesville, Virginia 22908
Jean-Francois Aubry Department of Radiation Oncology, University of Virginia, Charlottesville, Virginia 22908 and Institut Langevin, ESPCI ParisTech, CNRS UMR 7587, INSERM U979, Paris 75005, France
(Received 16 May 2014; revised 18 November 2014; accepted for publication 13 January 2015; published 13 March 2015) Purpose: In the ongoing endeavor of fine-tuning, the clinical application of transcranial MR-guided focused ultrasound (tcMRgFUS), ex-vivo studies wlkiith whole human skulls are of great use in improving the underlying technology guiding the accurate and precise thermal ablation of clinically relevant targets in the human skull. Described here are the designs, methods for fabrication, and notes on utility of three different ultrasound phantoms to be used for brain focused ultrasound research. Methods: Three different models of phantoms are developed and tested to be accurate, repeatable experimental options to provide means to further this research. The three models are a cadaver, a gel-filled skull, and a head mold containing a skull and filled with gel that mimics the brain and the skin. Each was positioned in a clinical tcMRgFUS system and sonicated at 1100 W (acoustic) for 12 s at different locations. Maximum temperature rise as measured by MR thermometry was recorded and compared against clinical data for a similar neurosurgical target. Results are presented as heating efficiency in units (◦C/kW/s) for direct comparison to available clinical data. The procedure for casting thermal phantom material is presented. The utility of each phantom model is discussed in the context of various tcMRgFUS research areas. Results: The cadaveric phantom model, gel-filled skull model, and full head phantom model had heating efficiencies of 5.3, 4.0, and 3.9 ◦C/(kW/s), respectively, compared to a sample clinical heating efficiency of 2.6 ◦C/(kW/s). In the seven research categories considered, the cadaveric phantom model was the most versatile, though less practical compared to the ex-vivo skull-based phantoms. Conclusions: Casting thermal phantom material was shown to be an effective way to prepare tissue-mimicking material for the phantoms presented. The phantom models presented are all useful in tcMRgFUS research, though some are better suited to a limited subset of applications depending on the researchers needs. C 2015 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4907959] Key words: head phantom, focused ultrasound, tcMRgFUS, brain, treatment envelope
1. INTRODUCTION Focused ultrasound (FUS), or high-intensity focused ultrasound (HIFU), is a technology in which ultrasonic energy emitted from a transducer is focused at a target volume.1 The treatment is monitored either through magnetic resonance imaging (MRgFUS: MR-guided FUS) or ultrasound imaging (USgFUS: US-guided FUS).2 The effects at the focal area can vary from thermal to mechanical depending on the procedure.3 Currently, the most common clinical usage of FUS is noninvasive thermal ablation. This particular biological effect has 1518
Med. Phys. 42 (4), April 2015
been widely used to treat the indications of prostate cancer,4,5 liver cancer,6,7 breast,8 and symptomatic uterine fibroids.9,10 The neurological applications of FUS have been more problematic owing to the reflection, absorption, and refraction of the ultrasonic waves by the skull. However, the recent development of aberration correction techniques to compensate for this difficulty11,12 has led to the development of transcranial MRgFUS (tcMRgFUS) devices.13–15 Currently, it is used clinically as a functional neurosurgery tool for the treatment of essential tremor16,17 and neuropathic pain.18 However, the noninvasive, radiation-free, and precise properties of
0094-2405/2015/42(4)/1518/10/$30.00
© 2015 Am. Assoc. Phys. Med.
1518
1519
Eames et al.: Head phantoms for transcranial focused ultrasound
tcMRgFUS would make it an ideal candidate for many other brain indications such as Parkinson’s disease, trigeminal neuralgia, epilepsy, and brain tumors. Novel techniques are under development in laboratories in order to improve the aberration correction techniques,19–21 and to image the location of the ultrasonic beam with low energy deposition using acoustic radiation force imaging.22–25 In order to further expand the treatment envelope—which is to say to increase the volume of the brain throughout which a single focal spot of sufficient intensity to achieve therapeutic effect can be achieved—and to translate to clinic novel guidance modalities, studies need to be conducted using ex-vivo skulls to accurately replicate the ultrasonic distortions induced by the skull in neurological treatment procedures. To advance this process, we tested three head phantom models in order to provide accurate and repeatable options for conducting the necessary experiments. The first phantom model is a cadaver. The second model is a gel-filled skull. The third model is a plastic mannequin head model into which a skull has been placed filled with the same tissue-mimicking material used in the second model, referred to here as the full head phantom. A total of nine phantoms were built and tested in a tcMRgFUS system.
2. METHODS 2.A. Cadaver
Full human cadavers were obtained from Virginia’s State Anatomical Program through the University of Virginia department of Neurosurgery. All specimens provided by this mechanism were full organ donors at the time of death. For each cadaver, a CT scan of the head was taken for the purposes of the skull correction algorithm. Trans-skull sonications could then be performed under MR image-guidance in the same manner as clinical tcMRgFUS treatments. 2.B. Description and preparation of tissue-mimicking phantom hydrogel
The phantom material used to mimic human tissue was a tissue-mimicking hydrogel (ATS Laboratories, Bridgeport CT). Acoustical and thermal properties of the hydrogel are summarized in Table I and compared to a previously published
1519
milk powder based phantom26 and to measurements in brain27 and skin.27 In order to pour the phantom, the hydrogel was first liquefied by heating to 55 ◦C in a conventional microwave oven dedicated to research use. 2.C. Gel-filled skull
The skull of previous cadaver was first defleshed and cleaned by the department of neurosurgery. A portion of the vault of the skull (later referred to as the skull vault) was cut away from the skull on a plane above the eye orbits and above the external occipital protuberance. Holes were drilled into the skull vault using a drill (3000 Variable Speed Rotary Tool, Dremel, Mount Prospect, IL) on the sides of the skull and the back and holes were also drilled into the skull base adjacent to those drilled into the skull vault. The insides of the skull cap and skull base were then lined with plastic film to create a watertight mold into which the gel was poured and cooled in a refrigerator to approximately 5 ◦C for 8 h to solidify. Once the gel had solidified, the plastic was peeled off and both gels were stored in the fridge. The two gels formed two halves of the brain tissue-mimicking phantom that fit closely to the inside surface of the skull (Fig. 1). The human ex-vivo skull was then degassed in water for at least 2 h in a vacuum chamber (Acrylic Round Vacuum Chamber, Abbess, Holliston, MA) at 230 mmHg. Once degassed, the cured gels for the skull vault and skull base were gently placed within the respective portions of the skull. The two halves of the skull were connected together while submerged in water by threading cable ties through the holes in the skull base and vault and connecting the two together. The skull was then attached to a base frame using cable ties, as shown in Fig. 2. The skull was then removed from the container and set up in the midfrequency ExAblate-Neuro system (InSightec, Tirat Hakarmel, Israel), which operates at a central frequency of 710 kHz. The setup was then promptly filled with degassed water to minimize oxygen exposure. The phantom fit is sufficiently close to trap a thin film of water between the gel and the inside surface of the skull, allowing for experimental setup maneuverability in air before submerging the setup in the transducer water bath in preparation for sonication. The quality of fit of this phantom in
T I. Acoustical and thermal properties of the brain (Ref. 27) and skin (Ref. 27), of the tissue-mimicking hydrogel (Ref. 28), and of a milk powder based phantom (Ref. 26). N/A: not available. Brain and skin tissue correspond to human, otherwise stated.
Speed of sound (m s−1)
Skin (Ref. 27)
Brain (Ref. 27)
1537 @ 17.5 ◦C
1532 @ 22 ◦C 1562 @ 37 ◦C 0.28 (gray matter) @ 37 ◦C 0.48 (white matter) @ 37 ◦C 3680 (gray matter) 3600 (white matter) 0.478 ± 0.015
Attenuation (dB cm−1 MHz−1)
3.5 ± 1.2 @ 23 ◦C
Specific heat capacity (J kg−1 K−1)
3215–3280 (pig)
Thermal conductivity (W m−1 K−1)
0.293 ± 0.016
Medical Physics, Vol. 42, No. 4, April 2015
Tissue-mimicking hydrogel (Ref. 28)
Milk powder based phantom (Ref. 26)
1537 @ 23 ◦C
1625 @ N/A◦C
0.5 ± 0.2 @ 23 ◦C
0.46 @ N/A◦C
3500 ± 500
N/A
0.5 ± 0.1
N/A
1520
Eames et al.: Head phantoms for transcranial focused ultrasound
1520
F. 1. From left to right: cured gel, skull, and gel fitted in the skull. Upper row displays the base and the lower row displays the vault.
an ex-vivo human skull is illustrated in Fig. 3 (center and right). 2.D. Full head phantom
First, a portion of the top of a plastic mannequin head (Fig. 4, left; Unisex head, Hard plastic clear, Inflatable Mannequins, Lindon, UT) was cut away. A complete human skull (Skulls Unlimited, Oklahoma City, OK) was then inserted into the mannequin head and settled into place (Fig. 4, center). The
removed piece from the top of the mannequin head was then reattached and sealed using tape (Fig. 4, right). The model was then inverted, immersed in the same degassing water tank displayed in Fig. 2 left, and digassed for 2 h. Immediately after removing the model from the degassing tank, water was evacuated and tissue-mimicking hydrogel (prepared as previously specified) was poured into mannequin head until the skull had been completely immersed. The mannequin mold was then wrapped in plastic film and secured in a container filled with cold water in order to facilitate the
F. 2. Degassing of the cadaver skull (left), tissue phantom mold beside skull (upper left), and gel-filled skull set up in frame (lower right). Medical Physics, Vol. 42, No. 4, April 2015
1521
Eames et al.: Head phantoms for transcranial focused ultrasound
1521
F. 3. Bare skull strapped to holder (left), and as seen on MR T2w sagittal (center), and coronal (right) image, which illustrates the quality of the fit of the phantom.
solidification of the gel (Fig. 5, left). The plastic film prevented the leakage of phantom from the model by allowing the water pressure to counteract the weight of the phantom in the model. It also prevented the water from leaking into the model and potentially diluting the tissue-mimicking gel or loosening the tape. The setup was left in place overnight in a refrigerator to allow the gel to cool and solidify. The mold was then removed and the plastic dried off. A CT scan was then taken of the completed model in order to provide the baseline data needed for the skull correction sequence (Fig. 6). The model was then set up in the ExAblate-Neuro transducer in the same manner that a patient’s head would be (Fig. 7). 2.E. Planning
Phantoms were positioned in the midfrequency ExAblateNeuro brain system. CT images of the skulls were aligned manually with T2 MR images with the InSightec planning software. The target was chosen at the location of the ventral intermediate nucleus (VIM), in the thalamus. The targeting
was performed on T2 images of the brain for the cadavers (Fig. 8, left). As the gel-filled skull model was poured in the same skull as the cadaver (Fig. 8, right), T2 images of the cadaver brain were fused to allow MR-based targeting of the VIM. Concerning the full head phantom, the VIM was located based on the skull anatomy landmarks on MR and CT. Compared to humans and cadavers, frontal sinuses are filled with gel in all our phantoms. Nevertheless, the clinical procedure needs to be applied to all phantoms, including the required drawing of “nopass regions” corresponding to sinuses so that the InSightec planning software disables ultrasound elements whose output would pass primarily through the sinuses. This planning step is mandatory for clinical sonication in order to avoid ultrasonic reflection on the air and consequent overheating, and it is recommended when using the phantoms. In both cases, no ultrasound beam will go through the sinuses, mitigating the fact that the phantom sinuses do not comprise air. 2.F. MR thermometry
Before, during, and after energy deposition, gel thermal rise in the focal point was evaluated using the proton resonance
F. 4. From left to right: the intact head mold, the mold with the upper portion removed and skull inserted, and the mold containing the skull with the upper portion reattached with tape. Medical Physics, Vol. 42, No. 4, April 2015
1522
Eames et al.: Head phantoms for transcranial focused ultrasound
1522
F. 5. The mold immersed in water (left) and the completed mold (right).
frequency (PRF) shift method of MR temperature mapping. The MR system was a 3T Discovery 750 (GE, Milwaukee, WI). The MR thermometry scan parameters used were parameters: TR/TE 27.6/12.8 ms, flip angle 30◦, bandwidth 5.68 kHz, FOV 28 cm, slice thickness 3 mm, matrix 256×128, scan time 3 s. A temperature sensitivity of −0.009 ppm/◦C was used29 and the initial temperature was arbitrarily set to 37 ◦C for all measurements.
full head phantom made out of the intact skull bone from Skulls Unlimited®. Thermal rises measured during the clinical essential tremor showed an average 2.59 ± 1.13 ◦C/kJ thermal rise in the patient’s VIM. It is lower than the ones observed in the cadaveric phantoms (4.85 ± 1.18 ◦C/kJ) and the gel phantoms (3.98 ± 0.27 ◦C/kJ).
4. DISCUSSION 2.G. Relevance of the head phantoms in regard to clinical data
Recently, a MRgFUS pilot study has been conducted at the University of Virginia on 15 drug resistant essential tremor patients.16 For each patient, the location of the focus was first adjusted and the power was ramped up until a lesion could be seen on T2 images, with an average of 18 successive sonications. We processed the corresponding anonymized data to extract the maximum temperature elevation induced in the VIM during the procedure. For all patients, we selected only the highest power sonication associated with an axial orientation plane for MR temperature monitoring.
3. RESULTS Table II summarizes all the temperature elevations induced by the ultrasonic beam at the VIM location in four different cadavers, in the four corresponding gel-filled skulls and in the
The cadaver model geometry is the most similar to the clinical environment: not only it includes air-filled sinuses, skin, and hair, but the degrees of freedom of the head phantom are constrained by its attachment to the patient body; in particular, the limited three-axis rotation of the head constrains the location of the transducer in regard to the head to clinically realistic positions. The anatomy is similar to that of a patient, the main difference being that perfusion is absent. Nevertheless, access to cadavers is not straightforward, and the model can only be used for a limited time period, typically one or two days. Gel-filled head phantoms advantageously provide a skull-specific, reusable phantom. Even though they do not reproduce the heterogeneities of the brain structures and of the skin fat and muscles surrounding the head, they are made of real human skull. The presence of the skull has long hampered the clinical implementation of focused ultrasound. The skullmimicking structure is thus a key component to be able to test new techniques for MRgFUS. This is the reason why no compromise was made on the skull itself. Taking advantage
F. 6. The full head phantom. Left: positioned in the CT scanner. Right: sagittal MR slice. Medical Physics, Vol. 42, No. 4, April 2015
1523
Eames et al.: Head phantoms for transcranial focused ultrasound
1523
F. 7. The setup for the full head phantom in the ExAblate-Neuro transducer.
of recent developments in composite material design, Wydra and Maev mimicked most of the acoustical properties of human bones.30 The human skulls used in this study could thus be replaced by composite material in the future. An alternative to the commercially available hydrogel (ATS Laboratories, Bridgeport CT) used in this study has also been proposed by Maruvada et al.31 The corresponding recipe can be found in King et al.32 The phantoms could be used to investigate various novel developments for tcMRgFUS ranging from checking skin heating, to developing new MR thermal mapping sequences in order to perform faster imaging or larger field of view. Some of the applications are listed in Table III and the appropriateness of the phantoms presented in this paper is rated. 4.A. Treatment envelope
Extending the treatment envelope consists of optimizing the target volume that can be treated with tcMRgFUS.33 Current treatments are limited to the thalamus. But promising treatments could be envisioned for epilepsy or trigeminal neu-
ralgia34 if the system was shown to be able to efficiently and safely deliver energy at other locations in the brain. The phantoms could be used to compare the maximum temperature elevation and the temperature distribution, as compared to the one obtained in the thalamus. For such a targeting study, the ability to visualize the anatomy of the brain is crucial. The cadaveric phantom model is the only one presented in this study that could provide MR imagery of brain anatomy, including trigeminal nerve, hippocampus, globus pallidus, or anterior/posterior commissure for determination of the VIM location. The tissue-mimicking phantom can be used for investigating the treatment envelope only if the skull has been extracted from a cadaver from which a T2 image has been acquired and that can be superimposed on the CT image of the skull in the InSightec planning software, as presented in Sec. 2.E. 4.B. MR-ARFI spot localization
While in current clinical trials, low-temperature thermal sonications are performed to assess focal spot quality and
F. 8. Target planning in a cadaver model (left) and in a gel-filled skull (right). Medical Physics, Vol. 42, No. 4, April 2015
1524
Eames et al.: Head phantoms for transcranial focused ultrasound
1524
T II. Summary of tcMRgFUS temperature measurements. Type
◦
Subject number
VIM side
Power (W)
Duration (s)
Tmax ( C)
∆Tmax/kW/s
Cadaver Cadaver Cadaver Cadaver Cadaver Cadaver Mean Std
1 2 3 3 4 4
R R R L R L
1100 1100 1100 1100 1100 1100
12 12 12 12 12 12
61 60 72 70 81 77 70.2 8.4
1.8 1.7 2.7 2.5 3.3 3.0 2.5 0.6
Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Gel-filled skull Mean Std
1 1 2 2 3 3 4 4
R L R L R L R L
1100 1100 1100 1100 1100 1100 1100 1100
12 12 12 12 12 12 12 12
49 50 50 55 50 51 57 58 52.5 3.6
0.9 1.0 1.0 1.4 1.0 1.1 1.5 1.6 1.2 0.3
1100 125 250 525 550 650 650 650 700 750 900 1000 1025 1100 1300 1300
12 10 10 10 10 15 10 10 10 10 10 10 12 12 15 15
51 43 45 56 53 59 63 58 57 57 62 48 60 57 58 55 55.4 5.8
1.1 4.7 3.4 3.5 2.9 2.2 4.1 3.2 2.8 2.6 2.8 1.1 1.9 1.5 1.1 0.9 2.6 1.1
Full head phantom ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient ET patient Mean Std
A B C D E F G H I J K L M N O
targeting accuracy, research has demonstrated the feasibility of detecting ARFI-induced tissue displacement with MR as a nonthermal means of evaluating these focal characteristics.21,35,36 Any of the three presented phantoms could be used to test and refine this nonthermal technique, however, there is a common compromise in tissue or material stiffness that affects all phantoms models. Because the cadaveric tissue is both deceased and colder than body temperature, it is generally stiffer than living brain tissue. This serves to decrease the sensitivity of this technique because the same ARFI sonication results in a smaller tissue displacement in the phantom than would occur in a living patient. While the gel-based phantoms presented also suffer from an increased stiffness compared to living brain tissue, it is possible to dilute the tissue-mimicking gel as much as 1:1 with water to decrease stiffness and thereby increase MR-ARFI signal for spot localization. This method of diluting the phantom mixture would need to be calibrated to accurately mimic the stiffness of brain tissue. Medical Physics, Vol. 42, No. 4, April 2015
4.C. MR-ARFI aberration correction
While similar to the MR-ARFI technique above, the aberration correction implementation aims to improve or replace the CT-based aberration correction in order to achieve a tighter focal spot in the brain or to altogether eliminate the need for a CT scan prior to a tMRgFUS procedure. As above, the same limitations with phantom material stiffness apply, again suggesting the need to calibrate the stiffness of the tissuemimicking phantom material. 4.D. Investigation of skin heating
Assessing the degree of skin heating that occurs during the course of clinical tMRgFUS procedures is important in order to prevent skin burns. Of the phantoms presented, the cadaveric model is the best suited to these studies as it is the most accurate model of the skin, fat, and muscle layers on the skull. Additionally, the fact that the cadaver is colder than
1525
Eames et al.: Head phantoms for transcranial focused ultrasound
body temperature does not compromise the use of this model because the chilled water bath surrounding the patient’s head in such FUS systems ensures that the skin approaches a temperature equal to the chilled water. The full head phantom model could be a low-cost alternative for investigating skin heating as it provides a layer of tissue-mimicking phantom material exterior to the skull surface, while the gel-filled skull is not suitable for this application. Acoustical and thermal properties of the tissue-mimicking hydrogel are close to that of the brain (Table I), but are also similar to that of the skin, except for attenuation (Table I). The full head phantom would thus mimic more closely heat release from the skull toward surrounding tissues than ultrasonic absorption by the skin itself. 4.E. New MR thermometry sequences
The phantoms described could also be valuable tools for evaluating techniques to improve on existing thermometry sequences or to introduce altogether new sequences. While purely imaging studies could be carried out on healthy human volunteers, it is hard to imagine a scenario in which FUSinduced temperature rise measurements using experimental MR thermometry sequences would not, at least at some stage, require a phantom model. All three phantoms presented have in common the following disadvantages as compared to clinical reality: their baseline temperatures will be cooler than body temperature and they lack perfusion and pulsatile flow. Nevertheless, Pulkkinnen et al.37 performed numerical simulations of clinical focused ultrasound thalamotomy and showed that perfusion impacts temperature elevation by 4% only when adding blood perfusion in their model. Because they degrade over a period of weeks to months, the gel-filled skull and full head phantoms are more durable than the cadaver model. Between the gel-based phantoms, the full head phantom has the advantage of permitting thermometry in the skin portion of the phantom, prompting us to recommend its use for most MR thermometry sequence development. 4.F. Novel bone imaging techniques 38
Ultrashort TE (time-to-echo) MR sequences (UTE) are commonly used in research environments to image bony anatomy. In the context of focused ultrasound, MR bone imaging could replace CT as the default source of information on
1525
skull geometry needed for aberration correction in the clinic. Because UTE is currently a research-only MR pulse sequence, there is an interest in exploring novel bone imaging techniques for FUS that could immediately be made available for clinical use. The gel-filled skull does not provide MR-signal at the outer bone surface, and is therefore the least preferred option of the gel-based phantoms. The cadaveric phantom offers both human soft tissue (blood, fat, muscle, brain) and bone (including bone marrow within the dipole), providing the most realistic imaging phantom for early pulse sequence development in the presence of the nonfiring but possibly MRartifact-producing transducer (i.e., before initiating a study on healthy human volunteers). The disadvantage of the cadaver is its limited longevity as compared to the gel-based phantom, making it the preferred choice only when accurate soft-tissue anatomy is required. 4.G. Development of FUS-compatible MR coils
Imaging in MR is improved by anatomy-specific imaging coils, but the water bath and need for an acoustically transparent path between anatomical target and transducer and of a focused ultrasound system prevents the use of commercially available MR imaging coils in clinical applications. These head phantoms serve as tools to design and evaluate prototype MR coils. While the gel-based phantoms, with good intraphantom homogeneity, can be used for preliminary calculation of a coil’s SNR improvement over alternative MR imaging options, the gel was not specifically designed to mimic the electromagnetic properties of living tissue, so that coil performance would need to be assessed on volunteer to better assess the SNR improvement. Cadavers serve both to load MR coils in a more realistic manner and to guide prototype decisions by means of the anatomically correct limitations a patient’s neck or shoulders may place on design options. 4.H. Relevance of the head phantoms in regard to clinical data
An average 2.6 ± 1 ◦C/kJ thermal rise was induced in the patient’s VIM. It is lower than the ones observed in the cadaveric phantoms (4.8 ± 1.2 ◦C/kJ) and the gel phantoms (4 ± 0.3 ◦C/kJ). All thermal rises per unit of time and acoustical power (∆Tmax/kW/s in Table II) obtained either with one
T III. Comparison of the best indication of each phantom for various applications. Application
Cadaver
Gel-filled skull
Full head gel phantom
Treatment envelope
+++
MR-ARFI spot localization MR-ARFI aberration correction Investigation of skin heating New MR thermometry sequences Novel bone imaging techniques Development of FUS-compatible MR coils
++ ++ +++ ++ + +++
++ (if skull extracted from a cadaver) ++ ++ − ++ ++ ++
++ (if skull extracted from a cadaver) ++ ++ + +++ ++ ++
Medical Physics, Vol. 42, No. 4, April 2015
1526
Eames et al.: Head phantoms for transcranial focused ultrasound
of the phantoms or measured during clinical essential tremor treatments overlap within the standard deviation, indicating that the models are relevant for preclinical developments. In the case of specific preclinical studies for which absolute thermal rise is of interest, such as treatment envelop extension, a 50% correction factor would need to be applied. 5. CONCLUSIONS This method for casting a commercially available thermal phantom material has proven to be an efficient, straightforward, and effective way to prepare head phantoms for tcMRgFUS studies. Peak thermal rises have been reported here in the phantoms when targeting the VIM and have been compared to clinical data on essential tremor treatments with the same anatomical target. Variations of the methods might be used for other clinically relevant preclinical or technical studies, making this technique relatively portable across anatomical applications—provided the gel is casted with a relevant geometry30—including breast, bone, or prostate. a)Author
to whom correspondence should be addressed. Electronic mail: [email protected] 1J. E. Kennedy, G. R. ter Haar, and D. Cranston, “High intensity focused ultrasound: Surgery of the future?,” Br. J. Radiol. 76, 590–599 (2003). 2J. L. Foley, M. Eames, J. Snell, A. Hananel, N. Kassell, and J.-F. Aubry, “Image-guided focused ultrasound: State of the technology and the challenges that lie ahead,” Imaging Med. 5, 357–370 (2013). 3G. ter Haar, “Ultrasound bioeffects and safety,” Proc. Inst. Mech. Eng., Part H 224, 363–373 (2010). 4S. Crouzet, X. Rebillard, D. Chevallier, P. Rischmann, G. Pasticier, G. Garcia, O. Rouviere, J. Y. Chapelon, and A. Gelet, “Multicentric oncologic outcomes of high-intensity focused ultrasound for localized prostate cancer in 803 patients,” Eur. Urol. 58, 559–566 (2010). 5Y. Inoue, K. Goto, T. Hayashi, and M. Hayashi, “Transrectal high-intensity focused ultrasound for treatment of localized prostate cancer,” Int. J. Urol. 18, 358–363 (2011). 6J.-F. Aubry, K. Pauly, C. Moonen, G. t. Haar, M. Ries, R. Salomir, S. Sokka, K. Sekins, Y. Shapira, F. Ye, H. Huff-Simonin, M. Eames, A. Hananel, N. Kassell, A. Napoli, J. Hwang, F. Wu, L. Zhang, A. Melzer, Y.-s. Kim, and W. Gedroyc, “The road to clinical use of high-intensity focused ultrasound for liver cancer: Technical and clinical consensus,” J. Ther. Ultrasound 1:13 (2013). 7F. Wu, Z. B. Wang, W. Z. Chen, W. Wang, Y. Gui, M. Zhang, G. Zheng, Y. Zhou, G. Xu, M. Li, C. Zhang, H. Ye, and R. Feng, “Extracorporeal high intensity focused ultrasound ablation in the treatment of 1038 patients with solid carcinomas in China: An overview,” Ultrason. Sonochem. 11, 149–154 (2004). 8H. Furusawa, K. Namba, S. Thomsen, F. Akiyama, A. Bendet, C. Tanaka, Y. Yasuda, and H. Nakahara, “Magnetic resonance–guided focused ultrasound surgery of breast cancer: Reliability and effectiveness,” J. Am. Coll. Surg. 203, 54–63 (2006). 9J. Rabinovici, M. David, H. Fukunishi, Y. Morita, B. S. Gostout, and E. A. Stewart, “Pregnancy outcome after magnetic resonance-guided focused ultrasound surgery (MRgFUS) for conservative treatment of uterine fibroids,” Fertil. Steril. 93, 199–209 (2008). 10E. A. Stewart, J. Rabinovici, C. M. Tempany, Y. Inbar, L. Regan, B. Gostout, G. Hesley, H. S. Kim, S. Hengst, and W. M. Gedroyc, “Clinical outcomes of focused ultrasound surgery for the treatment of uterine fibroids,” Fertil. Steril. 85, 22–29 (2006). 11J. F. Aubry, M. Tanter, M. Pernot, J. L. Thomas, and M. Fink, “Experimental demonstration of noninvasive transskull adaptive focusing based on prior computed tomography scans,” J. Acoust. Soc. Am. 113, 84–93 (2003). 12K. Hynynen and J. Sun, “Trans-skull ultrasound therapy: The feasibility of using image-derived skull thickness information to correct the phase Medical Physics, Vol. 42, No. 4, April 2015
1526
distortion,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control 46, 752–755 (1999). 13K. Hynynen, N. Mcdannold, G. Clement, F. A. Jolesz, E. Zadicario, R. Killiany, T. Moore, and D. Rosen, “Pre-clinical testing of a phased array ultrasound system for MRI-guided noninvasive surgery of the brain–a primate study,” Eur. J. Radiol. 59, 149–156 (2006). 14W. J. Elias, M. Khaled, J. D. Hilliard, J.-F. Aubry, R. C. Frysinger, J. P. Sheehan, M. Wintermark, and M. B. Lopes, “A magnetic resonance imaging, histological, and dose modeling comparison of focused ultrasound, radiofrequency, and Gamma Knife radiosurgery lesions in swine thalamus,” J. Neurosurg. 119, 307–317 (2013). 15D. Chauvet, L. Marsac, M. Pernot, A.-L. Boch, R. Guillevin, N. Salameh, L. Souris, L. Darrasse, M. Fink, and M. Tanter, “Targeting accuracy of transcranial magnetic resonance–guided high-intensity focused ultrasound brain therapy: A fresh cadaver model: Laboratory investigation,” J. Neurosurg. 118(5), 1046–1052 (2013). 16W. J. Elias, D. Huss, T. Voss, J. Loomba, M. Khaled, E. Zadicario, R. C. Frysinger, S. A. Sperling, S. Wylie, S. J. Monteith, J. Druzgal, B. B. Shah, M. Harrison, and M. Wintermark, “A pilot study of focused ultrasound thalamotomy for essential tremor,” N. Engl. J. Med. 369, 640–648 (2013). 17N. Lipsman, M. L. Schwartz, Y. Huang, L. Lee, T. Sankar, M. Chapman, K. Hynynen, and A. M. Lozano, “MR-guided focused ultrasound thalamotomy for essential tremor: A proof-of-concept study,” Lancet Neurol. 12, 462–468 (2013). 18D. Jeanmonod, B. Werner, A. Morel, L. Michels, E. Zadicario, G. Schiff, and E. Martin, “Transcranial magnetic resonance imaging-guided focused ultrasound: Noninvasive central lateral thalamotomy for chronic neuropathic pain,” Neurosurg. Focus 32, 1–11 (2012). 19J. Gateau, L. Marsac, M. Pernot, J. F. Aubry, M. Tanter, and M. Fink, “Transcranial ultrasonic therapy based on time reversal of acoustically induced cavitation bubble signature,” IEEE Trans. Biomed. Eng. 57, 134–144 (2010). 20E. A. Kaye, Y. Hertzberg, M. Marx, B. Werner, G. Navon, M. Levoy, and K. B. Pauly, “Application of Zernike polynomials towards accelerated adaptive focusing of transcranial high intensity focused ultrasound,” Med. Phys. 39, 6254–6263 (2012). 21L. Marsac, D. Chauvet, B. Larrat, M. Pernot, B. Robert, M. Fink, A.-L. Boch, J.-F. Aubry, and M. Tanter, “MR-guided adaptive focusing of therapeutic ultrasound beams in the human head,” Med. Phys. 39, 1141–1149 (2012). 22E. A. Kaye, J. Chen, and K. B. Pauly, “Rapid MR—ARFI method for focal spot localization during focused ultrasound therapy,” Magn. Reson. Med. 65, 738–743 (2011). 23E. Dervishi, B. Larrat, M. Pernot, C. Adam, Y. Marie, M. Fink, J.-Y. Delattre, A.-L. Boch, M. Tanter, and J.-F. Aubry, “Transcranial high intensity focused ultrasound therapy guided by 7 TESLA MRI in a rat brain tumour model: A feasibility study,” Int. J. Hyperthermia 29, 598–608 (2013). 24N. Mcdannold and S. E. Maier, “Magnetic resonance acoustic radiation force imaging,” Med. Phys. 35, 3748–3758 (2008). 25R. Paquin, A. Vignaud, L. Marsac, Y. Younan, S. Lehéricy, M. Tanter, and J.-F. Aubry, “Keyhole acceleration for magnetic resonance acoustic radiation force imaging (MR ARFI),” Magn. Reson. Imaging 31, 1695–1703 (2013). 26N. Mcdannold, G. T. Clement, P. Black, F. Jolesz, and K. Hynynen, “Transcranial magnetic resonance imaging–guided focused ultrasound surgery of brain tumors: Initial findings in 3 patients,” Neurosurgery 66, 323–332; discussion 332 (2010). 27F. A. Duck, Physical Properties of Tissues: A Comprehensive Reference Book (Academic, San Diego, CA, 1990). 28ATS Laboratories (2014), see http://www.atslaboratories-phantoms.com/ page8/index.html. 29T. Wu, K. R. Kendell, J. P. Felmlee, B. D. Lewis, and R. L. Ehman, “Reliability of water proton chemical shift temperature calibration for focused ultrasound ablation therapy,” Med. Phys. 27, 221–224 (2000). 30A. Wydra and R. G. Maev, “A novel composite material specifically developed for ultrasound bone phantoms: Cortical, trabecular and skull,” Phys. Med. Biol. 58, N303–N319 (2013). 31S. Maruvada, Y. Liu, W. Pritchard, B. Herman, and G. Harris, “Comparative study of temperature measurements in ex vivo swine muscle and a tissuemimicking material during high intensity focused ultrasound exposures,” Phys. Med. Biol. 57, 1–19 (2012). 32R. L. King, Y. Liu, S. Maruvada, B. A. Herman, K. A. Wear, and G. R. Harris, “Development and characterization of a tissue-mimicking material
1527
Eames et al.: Head phantoms for transcranial focused ultrasound
for high-intensity focused ultrasound,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control 58(7), 1397–1405 (2011). 33M. D. C. Eames, A. Hananel, N. F. Kassell, and J. W. Snell, “Intracranial treatment envelope mapping of transcranial focused ultrasound,” AIP Conf. Proc. 1503, 181–184 (2012). 34S. Monteith, J. Sheehan, R. Medel, M. Wintermark, M. Eames, J. Snell, N. F. Kassell, and W. J. Elias, “Potential intracranial applications of magnetic resonance–guided focused ultrasound surgery,” J. Neurosurg. 118, 215–221 (2013). 35Y. Hertzberg, A. Volovick, Y. Zur, Y. Medan, S. Vitek, and G. Navon, “Ultrasound focusing using magnetic resonance acoustic radiation force
Medical Physics, Vol. 42, No. 4, April 2015
1527
imaging: Application to ultrasound transcranial therapy,” Med. Phys. 37, 2934–2942 (2010). 36E. A. Kaye, J. Chen, and K. B. Pauly, “Rapid MR-ARFI method for focal spot localization during focused ultrasound therapy,” Magn. Reson. Med. 65, 738–743 (2011). 37A. Pulkkinen, B. Werner, E. Martin, and K. Hynynen, “Numerical simulations of clinical focused ultrasound functional neurosurgery,” Phys. Med. Biol. 59, 1679–1700 (2014). 38M. D. Robson, P. D. Gatehouse, M. Bydder, and G. M. Bydder, “Magnetic resonance: An introduction to ultrashort TE (UTE) imaging,” J. Comput. Assisted Tomogr. 27, 825–846 (2003).
