1 Introduction

With surgery, radiation therapy, and drugs having significant limitations in the brain, new treatment approaches for diseases and disorders of the central nervous system (CNS) are desperately needed. An important novel option is magnetic resonance-guided focused ultrasound surgery (MRgFUS) of the brain, a disruptive technology, through which clinicians can precisely focus a wide range of acoustic energy levels to targeted locations in the brain to produce non-invasively a variety of bioeffects that can transform practically every area of clinical neuroscience. Indeed, MRgFUS can revolutionize CNS disease treatment and radically shift treatment paradigms (Colen and Jolesz 2010; McDannold et al. 2010).

MRgFUS of the brain, known as transcranial (Tc) MRgFUS, is at a tipping point: it is on the cusp of becoming a clinical reality and has the potential to significantly impact multiple fields in clinical and basic neurosciences including neurology, neurosurgery, neuro-oncology, and radiation therapy. TcMRgFUS is moving from animal research to clinical testing, after which time it can be a clinical treatment method practiced not just in exclusive research domains but across the clinical community in academic centers.

Since its first demonstration by Lynn et al. (1942), therapeutic ultrasound technology has substantially advanced. Beginning in the 1940s, FUS has been investigated as a potential alternative to surgical resection and radiosurgery (Lynn et al. 1942; Fry and Fry 1960; Lele 1962; Meyers et al. 1959; Ballantine et al. 1960). However, it was abandoned as a lesioning device since removal of a relatively large section of the skull (craniotomy) was required to create an appropriate acoustic window. Clinical use of FUS for brain tumor treatment required a craniotomy (Ram et al. 2006; Guthkelch et al. 1991). The intact skull creates two major challenges. Firstly, it has high ultrasound absorption and thus loss of signal and heating of the bone occurs. Secondly, its irregular thickness and inhomogeneous density cause beam aberration that prevents focusing. The presence of the bony skull made MRgFUS an invasive method not unlike surgery and therefore was not seen as practical. This is despite the fact that after creating an acoustic window by craniotomy, the dura can remain intact and deep seating tumors can be treated, or lesions for functional neurosurgery can be generated. Another limiting factor for the use of FUS on the brain was the lack of imaging modalities to accurately visualize the target and the focus by monitoring temperature elevations during thermal ablation.

Two fundamental advances had to occur for transcranial MRgFUS (TcMRgFUS) to become practical and eventually clinically relevant: the discovery of an advanced imaging method for target definition and treatment monitoring, and the development of an FUS device that can focus the acoustic energy delivered through the irregular skull bone and provide a focal temperature elevation inside the intact skull. The development of the MRI in the 80s was the most important advancement needed to allow for accurate visualization of target lesions, such as tumors, and treatment monitoring of thermal deposition using MRI-based thermal maps (Jolesz et al. 1988; Panych et al. 1992; Kuroda et al. 1997; McDannold and Jolesz 2000; Chung et al. 1999). By the 1990s, through the development of large ultrasound-phased array transducers, completely non-invasive TcMRgFUS was able to focus through the intact skull after correcting phase distortions based on CT-based measurements of skull thickness and density (Clement et al. 2000; Hynynen and Jolesz 1998).

The phases of each element in the phased array ultrasound can be adjusted via transcranial focusing (Hynynen et al. 2004) and acoustic modeling. Overheating of the skull is addressed by the following two methods : First, a helmet-shaped, hemispherically-configured phased array transducer allows for spatial spreading of the transducer elements to distribute heat across the largest possible surface area. Second, a cold water cap containing degassed water is chilled to about 15 °C and circulates between the patient’s head and transducer to provide an active cooling system for the scalp and the skull (Hynynen et al. 2006a). Third, a relatively low frequency between under 1 MHz (250–650 kHz) is used for less absorption than would occur at higher frequencies.

MRgFUS therefore resulted from the fusion of FUS and MRI and became a single image-guided therapy delivery system through a landmark cooperation between the Brigham and Women’s Hospital (BWH), Insightec (Haifa, Israel) and General Electric (GE Healthcare, Milwaukee, WI). The development of MRgFUS technology in the early 1990s and the construction of the first prototype was the result of this successful collaboration (Cline et al. 1992, 1993, 1994).

In the past two decades MRgFUS has been introduced for the treatment of several benign (breast fibroadenoma, uterine fibroid) and malignant tumors (breast, liver, prostate and bone) (Tempany et al. 2011; Jolesz 2009). The commencement of clinical trials in the brain using MRgFUS began in 1994 at the Brigham and Women’s Hospital (McDannold et al. 2010; Cline et al. 1993, 1994) in patients with high-grade glioma, glioblastoma multiforme. Using the TcMRgFUS device, treatment of brain cancer through an intact skull was possible. This system has been successfully tested in patients with chronic pain treated with thalamotomy (Jeanmonod et al. 2012). It is currently being tested in clinical trials on essential tremor, pain, glioma, and metastatic brain tumor patients. Pre-clinical studies are currently underway for blood–brain barrier (BBB) disruption for the targeted delivery of chemotherapy (Treat et al. 2007; Kinoshita et al. 2006; Ting et al. 2012; Chen et al. 2010; Liu et al. 2010), nanoparticles (Liu et al. 2010b), interfering RNA (RNAi) (Frenkel 2008; Kinoshita and Hynynen 2005), and antibodies (Kinoshita et al. 2006). Its use for neuromodulation (Yoo et al. 2011), vascular malformations (Vaezy et al. 1998) and ischemic or hemorrhagic stroke (Medel et al. 2009) is also undergoing pre-clinical investigation.

2 Advantages and Limitations

MRgFUS provides advantages over other current treatment modalities including surgery, radiation therapy, and targeted drug delivery. When compared to surgery for tumor treatment, MRgFUS thermal ablations are non-invasive and thus do not carry the complication risks of hemorrhage, infection and collateral damage to non-targeted normal functioning tissue. The latter is especially important in treating CNS diseases. When compared to radiation therapy, MRgFUS does not carry the risks associated with ionizing radiation to the patient including the increased risk of secondary tumors. Of even greater importance is the ability of FUS to deliver unlimited retreatment sessions, an option not present in radiation therapy, where cumulative doses prevent repeated treatments. In radiation treatment, the dose is statistically determined based on accumulated prior experience and has toxic cumulative effects that prevent exposure past a single treatment session (irrelevant of the treatment success). FUS allows unlimited treatments in a single session and unlimited sessions over a period of time; this is important, particularly when a tumor recurs in the same anatomic region. FUS’s narrow thermal gradients allow for more precise targeting, such as when treating tumors adjacent to the optic or other cranial nerves. Furthermore, MR guidance provides real-time intra-procedural feedback and monitoring to prevent under- and over-treatment and to determine a therapeutic endpoint during the procedure (Hynynen et al. 1997; Vykhodtseva et al. 2000). When compared to laser ablation treatments that use invasive percutaneous heat-conducting probes, such as radiofrequency and laser therapy (Wood et al. 2002), FUS can non-invasively ablate geometrically asymmetric or non-circular target lesions while the laser probe ablation zone extends in each direction equally (spherically) and, therefore, cannot be tailored to fit the complex shape of a target volume. With lasers, shallow temperature gradients are seen, and heat dissipates from the single-source probe; as hyperthermia distributes at relatively lower temperatures (43 °C) over a longer period of time (30–60 minutes), selective cell death of malignant cells occurs while non-tumoral cells can survive the treatment (Diederich and Hynynen 1999). This is unlike the “all-or-none” ablative effect of FUS, irrespective of cell type.

However, certain limitations exist in MRgFUS. Current research is trying to address these issues to push the technology forward. Presently, treatment sessions are lengthy and require considerable manpower. MRgFUS thermal ablation is also currently limited to deep and centrally-located targets in the brain due to an inability of ultrasound to penetrate the skull when at extreme angles, a phenomenon noted in the first TcMRgFUS thermal ablation treatment clinical trial (McDannold et al. 2010). Research is under way to address these issues.

3 Treatment Paradigm of MRgFUS of the Brain

MRgFUS is the integration of two modalities into a closed therapy delivery system in which the acoustic component becomes the therapeutic modality and the imaging component takes on the role of targeting and monitoring as a guidance modality. The FUS device itself causes direct non-thermal (non-ablative) and thermal (ablative) effects in the target tissue.

3.1 TcMRgFUS for Thermal Tumor Ablation

3.1.1 Planning Phase

Pre-treatment planning is central to the MRgFUS patient’s workup. Conventional diagnostic neuroimaging is important for accurate anatomical localization of the target site. With the advent of more advanced techniques, such as diffusion and perfusion, these MRI techniques are helpful in evaluating increases in cell density and angiogenesis, respectively. In the future, these might serve as imaging surrogates for tumor detection outside and beyond the region of enhancement to allow for more complete thermal ablation and/or disruption of the BBB for targeted delivery of chemotherapy or gene therapy. Pre-procedural MRI biomarkers will likely be helpful in monitoring therapy and treatment response. Although not yet studied, perfusion and diffusion MRI might, better than conventional MRI, evaluate the extent of tumor and peritumoral regions. MRI perfusion in particular might help guide the delivery of thermal ablation since increase in blood flow and perfusion are associated with cooling effects. Other advanced techniques, similar to current pre-surgical planning, are diffusion tensor imaging (DTI) to determine the location of white matter tracts and its relation to the tumor and functional MRI to determine regions of viable eloquent cortex in relation to the tumor. These techniques are important particularly when the tumor or target is adjacent to or invades white matter tracts or eloquent cortex. Overall, MRI, in the pre-procedural phase, provides accurate definition and coordinates of the target site, including its relation to adjacent structures and helps determine the feasible target volume to deliver the FUS sonications.

Unique to the MRgFUS pre-treatment workup is the acquisition of a CT scan to calculate and compute the inhomogeneous skull density and thickness so that the phase of the phase array ultrasound elements can be adjusted (Clement et al. 2000, 2005; Hynynen and Jolesz 1998; Hynynen et al. 2006).

3.1.2 Targeting and Monitoring Phase

An immediate pre-treatment MRI is obtained on the day of the procedure with the patient positioned in the MRI suite and secured in the TcMRgFUS apparatus, which consists of the hemispherical phased array transducer with an underlying cranial cooling system. Conventional MRI sequences are obtained for same-day re-evaluation of the tumor and for accurate re-targeting of the tumor or target lesion.

MRI thermometry is the imaging modality of choice for intra-procedural monitoring and determining the therapeutic endpoint. By detecting proton resonance frequency shifts (McDannold and Jolesz 2000), MR thermometry determines small changes in temperature (2–3 °C) caused by focused sonications (McDannold et al. 2000) and can be used as a surrogate for tissue viability (Chen et al. 2001).

Initial non-treatment, sub-threshold, non-coagulative low-power sonications are delivered to the target region to identify and confirm correct targeting of the region of interest. These sub-threshold sonications, which do not cause tissue changes or damage, cause short, small temperature elevations (below 60 °C) and are detected by MR thermal maps, thus confirming correct targeting (Hynynen et al. 1997). During all stages of the procedure, patients are awake and responsive, ensuring neurological integrity.

Subsequently, treatment with high-power focused sonications (500–20,000 W/cm²) of short duration (1–60 s) is administered. During and between each high-power therapeutic sonication (McDannold et al. 2000), MR temperature-sensitive measurements are obtained to monitor treatment. FUS-induced thermal energy depositions are exploited in MRgFUS ablations. High temperatures (above 57–60°) over a few seconds cause irreversible cell death due to protein denaturing in a non-selective fashion (including both malignant and non-malignant cells). MR thermometry’s ability to accurately detect temperature changes in real-time allows for controlled energy deposition in thermal ablations whereby changes in sonication therapy can be performed immediately and in real-time to avoid over- or under-treatment. Cooling between sonication intervals is necessary to avoid heat effect summation and ablation of a larger tissue volume than the intended focal volume. Unlike probe-delivered laser ablations, FUS has a steep thermal gradient. The advantage of short sonication times and steep thermal gradients is that substantial cooling effects from blood flow and perfusion are circumvented, while heat build-up decreases in the tissue at the same time.

Conventional and, depending on the situation and tumor type, advanced MR imaging sequences, may be obtained intra-operatively to provide real-time updates on the progress and extent of tumor ablation. However, these MRI sequences should be tailored to confer the least amount of intra-procedural acquisition time. Extent of ablation, similar to the surgically termed extent of resection (Sanai and Berger 2008), is important as the percentage of tumor removed remains one of the important independent predictors of tumor recurrence and prognosis.

3.1.3 Post-Procedure Validation

A final post-procedural MRI scan is obtained when the desired therapeutic endpoint is achieved. Post-contrast MR imaging is acquired for treatment validation and to establish a post-treatment baseline MRI using simple post-contrast T1WI and T2WI to determine the success of therapy. In thermal ablation, a focal region of non-enhancement in the region of sonication histologically corresponds to tissue necrosis and blocked capillaries (McDannold et al. 2000) and reflects successful US-induced thermal ablation. In non-thermal therapy such as with the disruption of the BBB (McDannold et al. 2008), enhancement of the sonicated area reflects increased permeability and successful BBB opening (Fig. 1). T2WI helps to document post-procedure edema that typically resolves after 48–74 h (Morocz et al. 1998). However, with more refined MRI techniques, such as diffusion and perfusion, enhanced visualization of isolate foci of tumor left behind after tumor resection or ablation is possible.

Fig. 1
figure 1

T1-weighted axial image taken after BBBD demonstrates enhancement in the region of sonication which was performed in the area of the tumor and the previously non-enhancing peritumoral region

Full size image

4 Mechanism of Thermal Effects

In thermal ablation, FUS’s thermal heating effects on tissue are used to produce a thermolesion and destroy tissue at the point of convergence of the ultrasound beams where the summation of the acoustic intensity and concentration of energy is maximal. Overlapping high-power sonications are delivered in an iterative process to reach 56–60 °C, the threshold at which coagulation necrosis and irreversible non-selective cell destruction occurs. The US-induced thermal lesion achieved per sonication is small, typically 1–3 mm in diameter (perpendicular to the beam) and about twice as long in length (parallel to the beam) (Hynynen et al. 2006). The use of “cavitation-enhanced ablation” and the simultaneous treatment of multiple sites are two strategies employed to increase the focal point size and decrease overall treatment time. In the brain, “cavitation-enhanced ablation” has the potential to achieve tissue killing effect with lower energy levels than simple thermal coagulation to then result in less skull heating, which is the limitation of thermal TcMRgFUS. At the same time, cavitation increases the risk of secondary cavitation—induced hemorrhage.

5 Mechanism of Non-Thermal Effects

MRgFUS has non-thermal, non-ablative effects that are exploited in the brain primarily for targeted drug delivery through a transiently disrupted BBB (Hynynen et al. 2003; McDannold et al. 2006). Transient disruption can occur due to multiple complex mechanisms, the most important being cavitation defined as acoustically-induced interactions from microscopic gas bubbles in the medium (Nyborg 1968, 2001), which result in the following: (1) bubble oscillation; (2) acoustic streaming; (3) mechanical (acoustic radiation) forces; and (4) inertial (transient) cavitation (Deng et al. 2004; Hynynen et al. 2001; Mitragotri 2005; Sheikov et al. 2004), the last of which is believed to be responsible for most of the biological effects that cause blood brain barrier disruption (BBBD) (Fig. 2).

Fig. 2
figure 2

Interactions of microbubbles on the blood vessel. The shear stresses, resulting from bubble oscillation, microstreaming, and radiation forces, cause localized stretching of the cell membrane and stimulation of the endothelial cells via activation of stretch-sensitive ion channels, thereby, inducing its biological effects on the BBB. The BBB disruption results in agents passing through the tight junctions (arrow) and in active transport across the endothelial cells. At high acoustic pressures, rapid growth and oscillation of microbubbles can result in bubble implosion causing a release of stored energy and unpredictable deposition of thermal energy

Full size image

Microbubbles, cavities filled with gas or vapor (Minnaert 1933), can cause any of the four above-mentioned mechanisms. Microbubbles can be (1) seeded and induced by the ultrasound itself and generated within the native sonicated tissue (internally generated), or (2) intravenously administered using preformed microbubbles, the latter termed “microbubble-enhanced therapeutic ultrasound” (Hynynen et al. 2003; McDannold et al. 2006). Preformed microbubbles are typically made from human serum albumin shells filled with perfluorocarbon gas perflutren (mean bubble diameter, 2.0–4.5 μm). Two common commercially available microbubbles that are both FDA-approved as echocardiograph ultrasound contrast agents are Definity (Lantheus Medical Imaging Inc) and Optison (GE Healthcare).

Ultrasound bioeffects can be linear or non-linear. Linear effects are stable, predictable, and controlled. At low acoustic power, microbubble growth in size and oscillation via rectified diffusion (Nyborg 2000, 2001), also known as “stable” cavitation, cause BBBD. The shear stresses, resulting from bubble oscillation, microstreaming, and radiation forces, cause localized stretching of the cell membrane and stimulation of the endothelial cells via activation of stretch-sensitive ion channels, thereby inducing its biological effects on the BBB (Nyborg 1968, 2001). Eddying circulation motions of fluid around the bubble cause microstreaming, also called acoustic streaming. Radiation forces occur when the bubble moves in the direction of the wave propagation; it exerts mechanical force on and deformation of the endothelium perpendicular to the direction of the blood flow and length of the vessel (Leighton 1994).

Non-linear effects are unstable, non-predictable, and uncontrolled. At high acoustic pressures, FUS causes rapid growth and oscillation of microbubbles that subsequently undergo violent collapse (implosion), also known as inertial cavitation, which produces high velocity jets (Brujan et al. 2005), free radicals( Riesz and Kondo 1992), and an unpredictable release of stored energy in the form of shock waves (Nyborg 1968; Vykhodtseva et al. 2008) that disrupts the cell membrane and endothelial tight junctions. The unpredictable, non-linear uncontrolled deposition of thermal energy can lead to hemorrhage and unwanted tissue destruction outside the focal area (Vykhodtseva et al. 1995). Preformed contrast microbubble agents (Sheikov et al. 2004, 2008; Hynynen et al. 2003), by acting as cavitation nucleation sites, allow the same amount of BBBD or tissue ablation to occur at lower energy levels without the side effects of uncontrolled heating, hemorrhage and unwanted tissue destruction (Hynynen et al. 2001).

In addition to non-thermal treatments such as BBBD, the effects of cavitation can be exploited in thermal ablations, termed "cavitation-enhanced ablation," which is currently used in uterine fibroid ablations. With cavitation, the tissue heats faster, allowing for a larger thermal ablation volume to be ablated in a relatively shorter time than would occur without cavitation.

5.1 The BBB Challenge

The BBB serves as protective measure of the brain against systemic toxins, both innate and iatrogenic. Formed by a single continuous layer of endothelial cells bound by tight junctions, the basal lamina, and glial cell processes (Rubin and Staddon 1999), the intact BBB excludes more than 98 % of large-molecular neurotherapeutic drugs (Pardridge 2005). In addition to size, the charge and lipid solubility determine a molecule’s ability to cross the BBB. As an example, while lipophilic agents with a molecular mass of less than 400–500 Da can cross the BBB in significant pharmacological amounts (Pardridge 2003), ionized hydrophilic molecules larger than 180 Da cannot (Kroll and Neuwelt 1998).

In brain cancer, the global failure of effective neuro-oncology therapy continues due to an almost completely impenetrable BBB and the very aggressive and therapy-resistant character of most malignant primary brain tumors. In brain tumors, there is a gross misconception that the BBB is open in the region of active tumor enhancement seen on MRI that corresponds to the active tumor. However, contrary to general belief, BBB integrity is maintained in the well-vascularized, actively proliferating tumor (Neuwelt 2004). In fact, the region of largest BBB permeability is in the region of necrosis where there is tumor cell death, and chemotherapy is not needed (Neuwelt 2004). Even more importantly, compared to the region of tumor enhancement, the perilesional non-enhancing T2/FLAIR hyperintense regions and normal-appearing white matter where tumor infiltration is present, has a BBB with greater impermeability that decreases flow of chemotherapeutic drugs, a significant finding since it is these areas of tumor infiltration that are the major causes of tumor recurrence and patient death in glioma patients (Albert et al. 1994).

Intensive efforts have been undertaken to increase the delivery of drugs into the brain, including (1) drug modification, (2) mechanical disruption of the BBB, and (3) direct parenchymal drug injection and implantation. Modification of existing drugs and development of more lipophilic agents and water-soluble drugs with high affinities for carriers that penetrate the BBB are being investigated (Pardridge 2003). However, drug modifications carry a significant economic cost, are modified to fit a single specific drug, and take a long time to become usable in clinical practice. Furthermore, treatment of brain tumors, autism, and neurodegenerative diseases such as Alzheimer’s and Huntington’s, remains limited since these do not respond to small-molecular lipophilic drugs (Pardridge 2009). On the other hand, multiple diseases such as affective disorders and epilepsy consistently respond to such therapeutics (Pardridge 2009). Mechanical disruption of the BBB using intra-arterial infusion of hyperosmotic solutions, such as mannitol, is now in clinical trials (Guillaume et al. 2010; Doolittle et al. 2000). Mannitol causes shrinkage of the endothelial cells, resulting in the opening of the tight junctions and increased, though reversible, permeability of the BBB over a 5-h window (Doolittle et al. 2000). The BBB can also be bypassed via direct parenchymal injections of drugs and implanted delivery systems into the resection cavity (Kroll and Neuwelt 1998; Guerin et al. 2004). However, mechanical disruption of the BBB using intra-arterial infusion results in diffuse BBBD within the territory perfused by the injected vessel and requires invasive intra-arterial catheterization. Direct injection and implantation, although localized and targeted to the lesion, are invasive and destroy the overlying non-targeted trajectorial parenchyma, increasing the risk of complications. By contrast, MRgFUS is non-invasive, targeted, and localized to cause only transient effects. Moreover, it is an ideal “drug delivery system” that is generic, meaning that it is not tailored to a specific drug.

6 Clinical Applications

6.1 MRgFUS in Brain Tumors: Thermal Ablation and Targeted Delivery of Chemotherapy

In brain cancer, the "ideal therapeutic modality" for resection would allow for what can be described as the “ideal tumor surgery” – that is, complete removal of the tissue volume of the target lesion with complete functional and structural preservation of the surrounding tissue, including sparing of the tissue in the surgical path. Current neurosurgical resection can thus only approximate the "ideal surgery," and radiotherapy, although non-invasive, damages the tissue within the radiation beam path. MRgFUS, however, is non-invasive and destroys only targeted tissue at the point of acoustic convergence, sparing the overlying tissue. It has the capacity to deposit sufficient thermal energy to ablate even deep-seated lesions (Colen and Jolesz 2010).

Currently, in clinical trials, MRgFUS’ role in brain tumors is solely for thermal ablation of primary malignant CNS neoplasms (McDannold et al. 2010). Patients entered into both these trials have had high grade glioma, the most common primary malignant brain tumor. Gliomas grow through an infiltrative process beyond the area of abnormal enhancement into the area of abnormal non-enhancing peritumoral regions and even into the normal appearing brain parenchyma. The entire tumor, and all tumor cells, cannot be completely removed or ablated without associated injury to normal tissue and related functional damage. Many interspersed tumor cells are not destroyed and are invariably left behind to become the major cause of tumor recurrence and ultimately patient mortality (Sanai and Berger 2008). In glioma, debulking to decrease tumor volume is thus considered by many to be a satisfactory outcome.

However, those regions in which tumor resection or ablation is precluded might have a foreseeable viable treatment option in the future. Although still in its pre-clinical phases (Treat et al. 2007; Kinoshita et al. 2006) the MRgFUS BBBD technique may be able to target and treat the infiltrative tumor cells that are interspersed within normal tissue or tissue within regions of eloquent cortex. Targeted increases in permeability of the BBB may allow the passage of chemotherapy, nanoparticles, and other therapeutics to enter and target those infiltrative cells that are interspersed in those non-permeable regions in which chemotherapy would not otherwise reach (Kinoshita et al. 2006; Liu et al. 2010; Sheikov et al. 2008). Chemotherapeutic drugs can also be encapsulated in microbubbles or liposomes and attached to nanoparticles before administration (Unger et al. 2004; Hynynen et al. 1996). At the moment of bubble rupture, the encapsulated drug is released locally in the sonicated region. Furthermore, the ability to confirm treatment using contrast MRI makes for an unparalleled combination. In addition, the ability of MRgFUS to reversibly disrupt the BBB and increase its permeability can potentially transform and replace current non-selective drug delivery methods that cause systemic toxicity, an important limiting factor in chemotherapy today. Studies are investigating the ability of large molecular drugs, such as Herceptin in CNS breast cancer metastasis and Doxorubicin, to pass through the BBB after sonication with the hope that these protocols can be successfully used in the future to treat primary and secondary CNS malignancies (Treat et al. 2007; Kinoshita et al. 2006). Due to the non-invasiveness of the procedure and its possibility of enhancing the therapeutic delivery of drugs, this technique has the potential to translate into clinical practice in patients with CNS malignancies.

Benign tumors and most metastases, on the other hand, demonstrate well-defined borders, and thus complete destruction of tumor volume would be possible if the tumor is not within the eloquent cortex. In benign tumors in which tumor removal is the absolute treatment, MRgFUS provides a non-invasive method. Furthermore, in benign vascular malformations, MRgFUS provides a means to not only non-invasively, thermally ablate these lesions, but also to occlude the intralesional vasculature (Vaezy et al. 1998; Zderic et al. 2006; Barnard et al. 1955). In addition, with FUS’s sharp thermal gradient, along with its accurate treatment monitoring capabilities (i.e., MRI thermometry), MRgFUS treatment of tumors adjacent to nerves (i.e. optic and other cranial nerves) can be achieved.

MRgFUS thermal ablation alone might be sufficient for treating most benign tumors; however, in malignant infiltrative gliomas, MRgFUS thermal ablation, in conjunction with BBBD, might be a more effective approach. Early research focused on thermal ablation of brain tumors; however, non-thermal, non-ablative BBBD in brain tumor treatment for delivery of therapeutic agents and chemotherapy is being developed. In 1955, the first study of ultrasound-induced BBBD took place (Hynynen et al. 2006). Recent studies performed at our institution demonstrate and continue to confirm the ability of MRgFUS to produce selective, targeted, reversible BBBD, and, therefore, increase BBB permeability (Vykhodtseva et al. 2008; Martin et al. 2009). Currently, pre-clinical trials are under way at the BWH for the delivery of Herceptin and Doxorubicin in metastasis and glioblastoma multiforme, respectively, (Treat et al. 2007; Kinoshita et al. 2006) and delivery of Temozolomide and Bortezomib for the treatment of glioblastoma multiforme.

6.2 Functional Neurosurgery

Given that FUS has been shown to play a part in the treatment of certain functional neurological disorders, such as movement disorders, epilepsy, or pain (Meyers et al. 1959; Foley et al. 2004; Moser et al. 2012), it might have a major role in functional neurosurgery.

6.2.1 Chronic Neuropathic Pain

The results of the landmark clinical trial from the FUS research team in Zurich, Switzerland, have been published, demonstrating successful treatment of chronic neuropathic pain using MRgFUS for thermal ablation of the central lateral thalamus, confirming FUS’s role in functional neurosurgery (Foley et al. 2004). This first paper of this kind demonstrated positive short-term results (Foley et al. 2004), and more recently the encouraging long-term follow-up outcomes of these patients were published, showing significant somatosensory improvement that persisted at the 3- month and 1- year follow- ups (Jeanmonod et al. 2012). Given that the targets in functional neurosurgery are in normal tissue (for the most part), precise targeting measurements within the range of a millimeter is needed. A similar method employed in currently used radiofrequency or other therapeutic ablation techniques used in functional neurosurgery have been shown to be translatable to FUS, allowing for ongoing precision control (Fry 1958).

6.2.2 Spastic Disease and Pain

Ever since Fry et al. (Colucci et al. 2009) demonstrated the nerve effects possible with ultrasound, FUS has been viewed as having clinical potential as a non-invasive treatment of spastic diseases and pain control. An in-vivo study performed by Foley et al. demonstrated the ability of FUS to effectively block nerve conduction in the peripheral nerves in rabbits (Moser et al. 2012). Although the exact mechanism remains to be elucidated, it is hypothesized that the bioeffects of FUS-induced neurolysis are due to a combination of demyelination, axon disruption and other structural damage to the nerve fibers caused by thermal and mechanical effects (Moser et al. 2012). An in vitro study demonstrated the ability of FUS to temporarily and reversibly block nerve conduction at a frequency suitable for transkull sonication (Kluger and Triggs 2007), making FUS a potential future clinical tool for non-invasive intracranial functional testing and mapping that is similar to the currently investigated modality of transcranial magnetic stimulation (TMS) (Ferrara et al. 2007). FUS offers certain advantages over TMS; it allows for better localization and targeting (smaller effective volume) and can be localized and monitored using MRI (Hynynen et al. 1997).

6.2.3 Essential Tremors and Movement Disorders

In the 1950s, Meyers et al. (1959) investigated ultrasound’s use at the basal ganglia to alleviate the hyperkinetic and hypertonic disorders including those symptoms seen in Parkinson’s disease. Today a clinical study is underway at the Center of Ultrasound Functional Neurosurgery in Solothurn, Switzerland, to investigate the precision and safety of MRgFUS for tissue ablation in the thalamus, subthalamus and pallidum and its initial effectiveness for the treatment of long-term chronic therapy for resistant movement disorders of Parkinson’s disease and essential tremor. Another important current, first-of-its-kind study, is one with Insightec (Insightec Ltd), in collaboration with the University of Virginia, to evaluate the safety and initial effectiveness of MRgFUS thermal ablation for the treatment of essential tremors. In this study, 15 patients have already undergone non-invasive MRgFUS treatment using the Insightec Exblate TcMRgFUS (Insightec Ltd) system and experienced encouraging results; publication of the study’s findings is pending.

6.3 Stroke and Hemorrhage

FUS potentiates the thrombolytic effects of tPA and other thrombolytic drugs to increase clot lysis and improve the effectiveness of therapy in acute stroke patients (Medel et al. 2009). This lytic benefit comes with an increase in the risk of cavitation-induced hemorrhage in ischemic stroke that can be reduced by using FUS-induced arterial occlusion effects on the hemorrhaging vessels to generate hemostasis (Zderic et al. 2006; Barnard et al. 1955). This latter FUS effect can be exploited for the treatment of vascular malformations (Vaezy et al. 1998) for which FUS can also be used for ablations.

6.4 Targeted Delivery of Non-Oncological Agents

Besides the use of the MRgFUS BBBD technique for targeted delivery of chemotherapy, there are numerous non-neoplastic applications for MRgFUS, such as targeted delivery of iRNA (Frenkel 2008; Moonen 2007; Huang et al. 2012), DNA (Raymond et al. 2008), antibodies (Kinoshita et al. 2006), as well as diagnostic and therapeutic agents for Parkinson’s and Alzheimer’s diseases (Kinoshita et al. 2006; Jordao et al. 2010).

With antibodies being too large to cross the BBB, their delivery as therapeutics using MRgFUS is potentially significant. In the treatment of Alzheimer’s disease, animal model studies (Kinoshita et al. 2006; Jordao et al. 2010) have demonstrated promising results for the use of MRgFUS for detecting and decreasing protein aggregates, respectively. Most recently, for Alzheimer’s disease, using the MRgFUS BBBD technique, anti-amyloid beta antibody delivery resulted in significant plaque reduction in transgenic mice (Burgess et al. 2011). In antibody therapy for cancer targeting, antitumor monoclonal antibodies delivery against breast cancer have shown encouraging results in pre-clinical small animal studies (Kinoshita et al. 2006). Stem cells, which have promising potential in the future to treat neurodegenerative disease, traumatic brain injury, stroke, and brain tumors, were shown to enter the brain after BBBD (Paciotti et al. 2004).

While the BBB has remained a bottleneck for nanotechnology-based targeting for malignant tumors (Fulci and Chiocca 2007), the use of FUS BBBD was recently found to enhance delivery of gold nanoparticles. In 2010, Lie and colleagues demonstrated the feasibility and efficacy of FUS to increase the delivery of an iron oxide MNP conjugated to an antineoplastic drug into the tumor (Liu et al. 2010).

Though there are recent advances in genomics, gene therapies that can be-viral based have limited utility in brain tumors due to their inability to cross the BBB (Sheikov et al. 2006). FUS can, however, increase cell membrane permeability for use in gene therapy (Frenkel 2008; Deng et al. 2004). The first demonstration of FUS-mediated viral vectors was performed using the radiolabeled HSV vector (Shimamura et al. 2004). More recently, enhanced delivery and expression of the luciferase gene was seen in the rat brain using this method (Etame et al. 2012). Also, delivery of short interfering RNA (siRNA) and genes into the brain and cell membrane using FUS (Kinoshita and Hynynen 2005; Raymond et al. 2008) was successfully demonstrated. Given these successes, it follows logically that this technology could help deliver microRNA or short hairpin RNAs to silence and downregulate the expression of an aberrantly expressed protein.

7 The Future

In the future, TcMRgFUS can be used for the treatment of a potentially extensive range of CNS diseases and disorders. Not only is it non-invasive, but it is targeted and repeatable. Its use within an MRI unit enables precise anatomical and functional guidance unlike any other technology. Capabilities of FUS include the ability to non-invasively ablate tissue volume (replacing neurosurgery and radiosurgery), deliver drugs to targeted brain regions through a temporary disruption of the BBB (revolutionizing neuro-oncology and neuropharmacology), and reversibly modulate neuronal function (providing a tool with unprecedented abilities that can transform neuroscience). In recent years, with the development of devices capable of focusing ultrasound through the human skull, the demonstration of feasibility in humans, and the large number of pre-clinical studies that have been published, it has become clear that this technique is mature and ready to be moved to patients. However, this translation will be difficult since, for most people, the therapeutic use of ultrasound on the brain is a radical concept—a game changer. Significant work is therefore needed to prove that it can be applied safely before wide-scale testing and, ultimately, adoption, can take place.

We have seen exponential progress and advancements in MRgFUS. Multiple clinical trials have now begun for essential tremor, movement disorders, and chronic neuropathic pain, as well as for brain metastasis. Thermal ablation using MRgFUS can be expected to change the neurosurgical and radiation oncology fields. In regards to MRgFUS BBBD, preliminary pre-clinical results demonstrate both feasibility and efficacy to increase the delivery of molecules and therapeutics across the BBB (Min et al. 2011), a situation that should lead to many therapeutic opportunities to enter the brain, using treatments in sufficient concentrations to cause an effective therapeutic response. TcMRgFUS-induced neuromodulation has the potential to replace transcranial magnetic stimulation (TMS) as the testing method for brain function and functional connectivity (Yoo et al. 2011). Potentially, epilepsy can also be localized and even treated with TcMRgFUS(98).

While the investigation of therapeutic applications of FUS in the brain is not new, clinical translation of this technology has been hampered by its novelty, the huge expense needed for clinical translation of new treatments for CNS disease, and the perceived high risks involved in developing and applying new technologies for brain treatments, particularly for non-oncological applications. Despite these obstacles, today, MRgFUS has evolved into a clinical device that is being tested with patients, and the pre-clinical data on the system is largely mature. Compared to the enormous clinical potential, the investment needed to demonstrate that these uses of FUS are possible and effective is small. FUS is a classic “high-risk, high-reward” technology with the reward being uniquely high and truly “transformative.”

In conclusion, as it progressively demonstrates an increase in its clinical applicability and visibility, MRgFUS can be anticipated to radically shift the therapeutic paradigms for brain disease treatment The treatments possible with TcMRgFUS, applied alone or in combination, can completely change clinical practice and open up entirely new directions in the treatment of CNS diseases.