MR-Guided, Focused Ultrasound: Applications to Essential Tremor and Other Neurologic Conditions

In this issue of the American Journal of Neuroradiology, a novel approach by means of MR-guided, focused sonography surgery (MRgFUS) is used to treat essential tremor.¹ The results indicate that clinical improvement is significantly related to total lesion size. No relationship was found between the imaging characteristics of the lesion and sonication number, power, or maximal temperature. Although the authors describe an important advance in the use of this procedure, the study also raises a number of questions regarding the broad application of this technique to various neurologic conditions.

The use of focused sonography to treat brain disorders has evolved over the past 70 years. In the 1950s, Francis and William Fry developed a system of converging sonography beams to produce focal ablations in the brains of pigs and cats when applied through a craniotomy acoustic window.² The major limitation of this technology was the inability to focus sufficient sonography energy through the bony skull because of attenuation of acoustic energy. By the 1970s, their lab described the acoustic properties of the human skull³ and successfully achieved trans-skull transmission of an intensely focused ultrasonic beam.⁴

During the past decade, sonography therapy has emerged as a minimally invasive therapy for movement disorders, neuropathic pain, and malignancies. In combination with MR imaging and MR thermometry, MRgFUS can produce focused ablations in the brain by thermal and nonthermal effects with millimeter accuracy.⁵ Thermal (ablative) effects of MRgFUS occur when tissue is heated above 57–60°C, resulting in coagulative necrosis and tissue destruction. The degree of tissue necrosis is related to the focused sonography beam and can be monitored in real time with MR thermometry. Nonthermal (nonablative) effects of focused sonography result from acoustically induced interactions of microscopic gas bubbles or “microbubbles” with the surrounding vascular endothelium, a process termed “cavitation.” These interactions cause disruption of endothelial cell tight junctions and result in disruption of the blood-brain barrier. Because the sonography intensity needed to produce microbubble-induced cavitation is several orders of magnitude lower than the intensity needed for thermal ablation, this disruption of the blood-brain barrier is only temporary and has been shown to be safe and effective in an animal model.⁶

Both thermal and nonthermal mechanisms of MRgFUS can provide novel therapeutic opportunities for the treatment of brain disorders. Focused sonography is ideal for ablation therapy because it can target deep brain structures including the thalamus, subthalamus, and pallidum regions. However, it is limited in treatment of lesions near the calvaria because of the attenuation effects of the skull, which are more pronounced at locations nearer to bone. Ablative therapies have been investigated as suitable minimally invasive alternatives for glioblastoma,⁷ neuropathic pain,⁸ and essential tremor.^9,10 Investigations for the treatment of Parkinson disease are currently underway.¹¹

The short-lived disruption of the blood-brain barrier by MRgFUS provides a means to target delivery of drugs, antibodies, and stem cells to brain tissue.^12⇓–14 Sonography has also been used to enhance revascularization in a process termed “sonothrombolysis.” A recent meta-analysis of the use of sonography in ischemic stroke showed the therapy to be safe and effective.¹⁵ MRgFUS enables targeted delivery of sonography to the clot location and has the potential to improve the treatment of acute ischemic stroke. MR imaging can identify clot location and serve as a treatment map for immediate focused sonography therapy. Focused sonography sonothrombolysis has also been proposed for the treatment of intracerebral hemorrhage.¹⁶ In this setting, sonothrombolysis is used to liquefy the clotted blood within the intracerebral hemorrhage with consequent minimally invasive MR imaging–guided drainage of the liquefied clot.

The effectiveness and utility of sonography therapy can be augmented with nanotechnology. Thermal ablation is being evaluated by use of multifunctional drug delivery systems capable of triggering local hyperthermia in the presence of low-frequency sonography.¹⁷ These systems provide a unique synergistic combination of chemotherapy, thermal therapy, and real-time imaging and are being investigated for the treatment of CNS malignancies. The present study by Wintermark et al¹ demonstrates the importance of lesion size in achieving symptom relief. Although total lesion size was significantly correlated to clinical improvement, the value of the imaging findings remains unclear. The time-dependent imaging characteristics of MRgFUS-induced brain lesions on T2-weighted imaging consists of 3 concentric zones: zones I and II appear as a result of coagulation and necrosis, and zone III appears as the most peripheral of the concentric zones and represents transient edema.¹⁸ A larger zone III area is correlated with clinical improvement, but some of this improvement is lost as the edema resolves. Of interest, 2 patients with limited clinical improvement had imaging characteristics that were not very different from those with clinical improvement. This raises the concern of difficulties associated with accurately locating therapeutic targets. The ventrointermediate nuclei (Vim) are the thalamic relays of the cerebellothalamocortical tract and are the principal targets of MRgFUS in the treatment of essential tremor. Two methods may be used to locate the Vim: image-based coordinate targeting (direct method) and atlas-based targeting (indirect method). The latter approach is subject to potential inaccurate localization of the anterior and posterior commissures, an error that can be >5 mm. Direct identification is considered to be more accurate in identifying Vim and may be achieved with fractional anisotropy and color-coded vector maps.¹⁹ Lesion identification in the current study was determined by atlas coordinates and clinical parameters evaluated in real time with sublesional sonication. In the 2 patients with limited therapeutic benefit, the MRgFUS was not repositioned, and the patients did not show sensory symptoms during treatment. Future studies may incorporate direct methods of Vim location during sonication to confirm target identification. Diffusion tractography may also be useful in evaluating the integrity of these tracts over time and in correlating their integrity with clinical symptoms. This approach could potentially help to identify valuable imaging information and provide useful targets for repeat therapy. Last, in the current study, total lesion size appeared to be unrelated to sonication number, power, or maximal temperature, presumably because of the small effect size and underpowered study. Determining the optimal use of these variables may improve the clinical utility of MRgFUS.

• © 2014 by American Journal of Neuroradiology