If a high enough temperature can be reached to create a thermoablative lesion of the correct size and at the correct location, considerable clinical efficacy can be obtained for conditions such as PD (9), ET (10), OCD/major depressive disorder (MDD) (11), and epilepsy (12). Although modern MRgFUS techniques generally allow high spatial and temporal resolution of temperature characterization and effective control of temperature distribution in the brain, several studies have reported difficulties in achieving a temperature that reaches the ablative level in patients with ET and PD because of the physical properties of the skull (13, 14).

The skull's acoustic properties are different from those of soft tissue. The intrinsic attenuation of ultrasonic waves in the skull is ~20 dB/cm^*MHz, which is higher than the attenuation of waves in brain tissue (~0.8 dB/cm^*MHz) (15). Attenuation arises from acoustic absorption and scattering in all directions in the medium. Acoustic scattering refers to that part of an incident acoustic wave that is reflected from interfaces between different tissues due to inhomogeneities in their density and compressibility (16, 17). This scattering can be substantial at major interfaces, such as between bone and soft tissue. Reflection is also high at interfaces between the outer/inner tables of cortical bones and the central cancellous bone because of their different bone structures (18). Increased attenuation implies decreased heating power available at the target, in addition to increased deposition at nontarget soft tissues such as the scalp and skull. Such unwanted heating is exacerbated if the total incident power is increased to compensate for increased attenuation and maintain the desired temperature at the target. There can be great variability in attenuation between portions of a single skull and also between the skulls of different patients. Some patients are effectively untreatable with this method because of potential scalp burns or damage to the underlying bone, in addition to the painful heating of nontarget tissues. Thus, a practical simple measure that can predict which patients may benefit from HIFU is needed.

Along any ultrasound ray traversing the skull, one can calculate the ratio of skull density between the mean cortical and mean cancellous bone using the Hounsfield units (HUs) that result from a CT scan with high resolution and using a bone kernel for image reconstruction. These ratios can be averaged over all rays traversing the skull in a HIFU configuration to provide a single measure called the skull density ratio (SDR), which is a useful global index for identifying patients who are eligible to undergo MRgFUS-mediated lesion creation (Figure 3). Although currently used MRgFUS techniques can compensate for skull factors with CT-based phase-correction software on multiarray systems, some patients will still demonstrate sufficiently low SDR to restrict them from treatment. To this end, practice guidelines from the American Society for Stereotactic and Functional Neurosurgery state that patients with an SDR <0.4 should not undergo MRgFUS for lesion creation (19), as insufficient heating at the intracranial target will lengthen the time needed to achieve ablative temperatures and lead to excessive heating at nontarget tissues.

Researchers have also assessed other factors such as skull morphology (20) and volume (21) as potential predictors of MRgFUS treatment success. For instance, in a retrospective analysis of 189 patients who underwent MRgFUS, D'Souza et al. (22) found that patients across different SDR categories (SDR <0.4, 0.4 ≤ SDR <0.45, and SDR ≥ 0.45) demonstrated similar improvements in clinical outcomes indexed by 1-year follow-up Clinical Rating Scale for Tremor (CRST) scores even though the percentage of patients achieving the peak temperature of 54°C was substantially higher in patients with SDR ≥ 0.45 (91%) than in those with 0.4 ≤ SDR <0.45 (64%) and those with SDR <0.4 (55%). Several other studies have reported cases of patients with low SDR in whom permanent ablation was still achieved (23, 24). Thus, there is an urgent clinical need to investigate factors such as skull thickness, incidence angle, and skull heterogeneity as potential predictors of treatment success, thereby identifying new metrics that may be more accurate in predicting success among patients with low SDR.

For patients with low SDR, the thalamic target energy deposition will be inefficient, leading to an increased risk of overheating the scalp and skull at the expense of therapeutic heating in the thalamic target. This unwanted heating will cause the patient pain and could lead to irreversible tissue damage (Figure 4). In addition, the efficiency of energy deposition decreases over the course of multiple treatment cycles, potentially due to effects from skull heating (25). A recent study of patients treated with MRgFUS reported that seven out of 30 patients demonstrated multiple new asymptomatic calvarial marrow injuries 3 months after attempted treatment (26). This study found no correlation between SDR and the presence of skull lesions, but the maximum power used was substantially higher for patients with lesions than for those without. For instance, one patient with a skull lesion had undergone prolonged sonication for 31 s with 1,100 W maximum power, but the maximum temperature achieved was only 48°C. There is a well-established time-temperature relationship that can change various time-temperature profiles into a standardized single measurement to estimate the degree of the thermal dose while allowing for tissue necrosis (27). For future MRgFUS procedures, clinicians must be able to tailor parameters such as maximum energy, sonication duration, and number of sonication sessions so that they can prevent thermal hotspots in the skull and thus prevent long-term skull injuries, especially for patients with unfavorable skull bone characteristics.

InSightec's hemispheric phased array system is composed of 1,024 element of transducers, and besides the skull bone property, i.e., SDR and thickness, the successful delivery of ultrasound beam from each transducer depend on the angle of incidence to the skull surface. Since the shape of the skull is not spherical, if the targeting brain structure is located far from the center of the brain, some of the ultrasound rays may likely have an incidence angle to the skull of >25°. The increased angle of incidence will deactivate the corresponding transducer and lead to a less cumulative number of active elements as the incidence angle of >25° will increase reflection and thereby likely decrease energy deposition inside of the planned target and increase deposition in the scalp. As a specific example, a recent study by Jung et al. (28) showed that an increased number of elements were deactivated when targeting globus pallidus interna (Gpi) and anterior limb of the internal capsule (ALIC), which are more lateral brain structures from the center of the brain compared to the thalamus. Several simulation studies have suggested that significant attenuation may be attributed from longitudinal-shear (transverse) mode conversion (29, 30); however, another study also reported that the conversion of longitudinal waves to shear waves inside the skull is insignificant when the incidence angle is <20°, with the assumption that amplitude loss during shear wave conversion from incident rays at the skull is not critical (31). An investigation using an ex vivo human skull demonstrated a significant reduction (nearly 31% loss of normal incidence) in transmitted amplitude when the incident angle was 31° at 0.548 MHz (32). In the study by Jung et al. (28), at an incidence angle >25°, energy transmission sharply decreased when the SDR was <0.6, but the energy transmission started to recover when the SDR was >0.6, indicating that a high SDR compensates for the influence of a higher incidence angle. This study highlights that even though SDR provides a useful standard value for screening eligible patients, the role of incidence angle also must be considered, especially when the focal region is distant from the transducer's geometric focus as in cases of capsulotomy (33) and pallidotomy (34).

Jung at al. (28) also demonstrated higher energy transmission (by a factor of ~3) at a lower frequency (230 kHz) than at a mid-frequency (680 kHz) for all SDR and incidence angle ranges. These findings helped to initiate subsequent studies developing a low-frequency system to circumvent skull limitations at higher frequencies, in addition to broadening the regions in the brain accessible to lesion creation.

Recent studies have used computer simulations to better predict the temperature increase in targets in individual skulls by modeling the skull efficiency with properties extracted from CT, in particular the HU (36, 37), an arbitrary unit of radio density. Although the HU has widespread applicability, it is still dependent on various other factors (38), and so standardization of CT parameters should optimize the use of HU as a rigorous diagnostic tool for evaluating skull adequacy for MRgFUS lesioning. Recently, researchers reported the use of a novel method employing microbubbles as an ultrasound contrast agent; this technique allowed acoustic echoes to modify phase corrections and thereby narrow the acoustic focus. This method, called “echo focusing,” provided sonication efficiency for lesion formation that was superior to that obtained with CT-based aberration correction (24, 39). In these studies, an echo-focusing phase aberration correction technique was incorporated by measuring returning acoustic signals from intravenously injected microbubbles around the intended target region during sonication (40, 41). With echo focusing, successful lesion formation was achieved in 12 patients with ET, including 3 patients in whom MRgFUS thalamotomy treatment using CT-based aberration correction had failed (24). In another study, 8 patients with low SDR (mean SDR = 0.35) were successfully treated using the echo-focusing method by raising the temperature to >54°C in patients with ET and to >52°C in patients with PD; these temperatures were sufficient for lesion formation (39). This echo-focusing technique could be particularly beneficial for patients with low SDR and for those with a target that is more lateral than the thalamus, as this research demonstrated permanent lesion formation in cases of pallidotomy in patients with an SDR <0.4.

Similar to deep brain stimulation (DBS), FUS has features that confer “closed-loop” status. Specifically, during the staged procedure, repeated examinations of the effect of increasing sublesional temperatures on the patient's tremor (as measured with continuous MR thermometry; Figure 5) (35) provide near–real-time feedback to verify targeting, monitor outcome, and update the treatment plan. Additionally, because the patient does not need to be placed under general anesthesia for FUS, the effect of treatment on tremor can be observed immediately after each sonication both from the accelerometer and the patient's handwriting (Figure 6). Approaches such as the use of an intraoperative accelerometer to quantify the tremor response in real time are necessary and will help to complete closed-loop feedback procedures in a patient-specific manner (Figure 7).

Research should continue to focus on developing a reliable method to identify the target, reduce lasting side effects, and enhance durability. For instance, during the course of the MRgFUS thalamotomy, different MR sequences can provide information about lesion volume and diameter changes over time (Figure 8). Only T2-weighted sequences can show the lesion shortly after it is created; however, the lesion may not be apparent on T2-weighted images obtained up to 180 days after the procedure. Susceptibility-weighted images, on the other hand, can demonstrate the lesion up to 180 days after treatment (42). Fast gray matter acquisition T1 inversion recovery imaging can be used for surgical planning, as this method offers superior visualization of the target and is especially effective in differentiating between the internal capsule and thalamus (Figure 9). Furthermore, T1-weighted 7T images can depict lesion shrinkage and shifting up to 65 days after treatment (Figure 10).

Disabling movement disorders such as ET and PD are often diagnosed in patients of advanced age, and the incidence of these disorders is increasing due to a growing and aging world population. The effect of movement disorders on daily life is considerable, impairing routine functions such as holding a glass of water, writing a check, or using a hand-held device. Safe and effective treatment options are therefore needed so that patients can maintain independent function. Although medications such as beta-blockers may adequately control mild upper extremity tremor, they are practically inconsequential for slowing moderate to severe tremor. Propranolol and primidone (beta-Blockers) are the most common choice of drugs in medical therapy for treating moderate to severe functional disability in ET. However, if the patient has a contraindications to beta-blockers or inadequate tremor control, other drugs such as Mysoline, Benzodiazepines, gabapentin, topiramate, zonisamide can be used as add-on therapy or monotherapy. Surgeries such as FUS-mediated Vim thalamotomy or DBS are considered last when medical treatment does not help suppress tremor. On the other hand, PD has more complex features encompassing both motor and non-motor disabilities. Most PD patients initiate treatment with levodopa therapy, the most effective drug in treating PD. However, the long-term use of levodopa frequently leads to dyskinesia (43) and wearing-off phenomenon. Patients with levodopa resistance or with rapid progression motor symptoms seek surgical treatment such as bilateral DBS, or unilateral FUS-mediated Vim thalamotomy. They could also benefit from recently approved FUS-mediated pallidotomy for bradykinesia and drug-induced dyskinesia, etc.

DBS of targets such as the ventral intermediate nucleus (Vim) of the thalamus (44), subthalamic nucleus (45), and internal globus pallidus (46) is a well-established option for the treatment of movement disorders. The benefits of this treatment are long-lasting and the risk profile is low; however, the procedure is costly and invasive and requires permanently implanted hardware. Because FUS circumvents these surgical complications, there was early enthusiasm that this technique might appeal to patients reluctant to undergo DBS. Radiosurgery using Gamma Knife had a similar appeal, but this procedure was limited by an absence of intraoperative validation (47) and a lack of reliable methods for targeting.

The overall outcomes of FUS and radiosurgical thalamotomy for tremor are related to the size and location of the lesion; these factors also govern potential side effects. While the durability and efficacy of treatment may be greater with larger lesions, so too is the probability of adverse effects (10). Radiosurgical thalamotomy has a shallower temperature gradient than HIFU (48), which consequently increases the uncertainty of the lesion's margins and increases the risk of extending the lesion beyond the planned target. Although the temperature range with HIFU may be easier to control, there is a small lag (2–3 s) in MRI thermometry maps, which may partially offset this advantage.

In 2016, based on data from a clinical trial by Elias et al. (14), the FDA approved the use of HIFU for ablation of the Vim in patients with ET, and this indication was expanded in 2018 to include ablation of the Vim in patients with tremor-dominant PD. With this technique, high-intensity ultrasound waves irreversibly create lesions in the target structures via coagulative necrosis in the tissue secondary to the heat resulting from frictional forces (49, 50). The temperatures typically need to exceed 55°C for this treatment to be effective (51). In the first proof-of-concept clinical study testing HIFU in a randomized controlled trial, ablation of the thalamic Vim with MRgFUS significantly suppressed tremor, with patients demonstrating improvements in finger-to-nose pointing tasks (14). In addition, scores on the CRST were reduced by 81.3% at 3 months after HIFU treatment compared with baseline scores. A follow-up study published in 2018 (52) demonstrated a sustained effect at 2 years in most patients with ET, and another follow-up study highlighted continued benefit from the unilateral thalamotomy after 3 years (53). It is also noteworthy that patients who benefited from the unilateral MRgFUS thalamotomy without experiencing any side effects wished to extend the lesioning procedure for bilateral thalamotomy to improve on the other side. Although a few case reports of bilateral thalamotomy state that staged second treatment were successful without severe adverse events (54–57), bilateral thalamotomy is not currently approved as standard treatment as the research of bilateral thalamotomy in relation to the tremor etiology, adjustment of the second lesion, and overall incidence of adverse events are still under investigation (NCT04112381, https://clinicaltrials.gov/ct2/show/NCT04112381).

Since this initial research was published, multiple other studies have demonstrated similar findings. For instance, Bond et al. (9) used MRgFUS to perform unilateral Vim thalamotomy in patients with PD refractory to levodopa and reported a 62% improvement in CRST scores for hand tremor contralateral to the treatment side. In another study, MRgFUS of the Vim led to a 55.9% improvement in tremor score at 3 months after treatment in 6 patients with diverse tremor types including PD, dystonia, dystonia gene–associated tremor, and writer's cramp (58). Two patients experienced lasting side effects of hemitongue numbness and hemiparesis with hemihypoesthesia, suggesting the need for long-term safety evaluation with a larger sample size. In 2020, researchers assessed the feasibility of using MRgFUS for unilateral pallidotomy in PD patients with dyskinesia induced by the long-term use of levodopa (34). The study demonstrated a 43% improvement from baseline in Unified Dyskinesia Rating Scale score, an effect that persisted through 12 months. As the pathophysiology of PD is different from that of ET, patients show distinct cardinal signs such as rest tremor, rigidity, and bradykinesia, which has prompted researchers to explore novel therapeutic targets for PD. In November 2021, FDA extended the therapeutic target for PD and approved MRgFUS for pallidotomy. Additionally, a clinical trial assessing bilateral ablation of the pallidothalamic tract for PD is currently ongoing (NCT04728295, https://clinicaltrials.gov/ct2/show/NCT04728295). Furthermore, researchers in Madrid assessed lesioning the subthalamic nucleus for PD with markedly asymmetric signs and found that the Movement Disorder Society-Unified Parkinson's Disease Rating Scale motor score (i.e., part III) was decreased by approximately 50% from baseline to 4 months after treatment (59).

Although early research studied the response of tremor to HIFU targeting of the Vim, other targets are also under investigation for the treatment of disabling movement disorders. For example, the internal globus pallidus is a preferred target when patients' symptoms are dominated by dyskinesia and dystonia. Previous research has demonstrated that pallidotomy and pallidal DBS lead to marked improvements in the symptoms and motor dysfunctions of PD (60).

The current evidence suggests that HIFU is a safe and effective option for patients with disabling ET or PD who are not candidates for DBS or are reluctant to undergo surgery. However, more studies are needed to address the nontremor motor symptoms of movement disorders. Additionally, questions regarding durability, the safety of bilateral treatment, and novel therapeutic targets for tremor are currently being investigated in clinical trials.

OCD is a common psychiatric disorder characterized by repetitive behaviors, compulsions, and urges detrimental to health and quality of life (61). Selective serotonin reuptake inhibitors are currently the first-line pharmacotherapy for management of OCD. Because of the chronic nature of this disease, medical therapy is often combined with cognitive and behavior therapy to increase the durability of treatment. However, 20% to 30% of patients do not respond to medication and could potentially benefit from neurosurgical options such as DBS (62, 63), radiosurgery (Gamma Knife) (64), and MRgFUS (33). Of these techniques, MRgFUS has the advantage of being noninvasive, with the added benefit of lack of general anesthesia and associated surgical complications. In addition, with MRgFUS, the lesion size and location can be controlled in real time.

The study of MRgFUS in psychiatric disorders dates back to the 1950s, when neurosurgeon Petter Aron Lindstrom first removed brain tissue using MRgFUS as an alternative to prefrontal invasive craniotomy and lobotomy (65). Lindstrom introduced the concept of MRgFUS-mediated lobotomy to his colleague Lars Leksell, who then set out to study the use of MRgFUS for treating psychiatric disorders (Steiner L. Personal communication. 2007). Leksell designed a custom stereotactic headframe as a precise lesioning tool but was unable to complete his MRgFUS investigations because of the challenges imposed by the transmission of ultrasound through the skull. However, his contributions to developing a noninvasive modality for the treatment of functional brain disorders laid the foundation for the development of the first Gamma Knife model.

Research into the pathological brain networks responsible for OCD has focused on the cortico-striatal-thalamo-cortical pathway (33), and recent neuroimaging studies suggest involvement of the orbitofrontal cortex, the dorsal anterior cingulate cortex, and the amygdalo-cortical circuit (66, 67). In a study of 4 patients with OCD, bilateral MRgFUS was used to create lesions in the anterior limb of the internal capsule. This dense white matter tract consists of afferent and efferent fibers of the affect network and reward network that run their course up to the orbitofrontal cortex (68) and form a target for medication-resistant OCD. The patients treated with this procedure demonstrated a gradual improvement in the Yale-Brown Obsessive-Compulsive Scale (Y-BOCS), with a mean improvement of 33% over a 6-month follow-up (33). In addition, depression and anxiety levels after MRgFUS capsulotomy were almost immediately improved (mean reductions of 61.1 and 69.4%, respectively). Similarly, Kim et al. (69) used MRgFUS bilateral capsulotomy to treat 7 patients with medication-refractory OCD and measured the patients' Y-BOCS scores at 1 week and at 1, 3, 6, 12, and 24 months. Significant improvement (38%) from baseline was seen at the 24-month follow-up, without any severe adverse events. In addition, OCD symptoms started to improve as early as 1 week after MRgFUS capsulotomy, and depression and anxiety levels were also reduced at 1-week follow-up (−47.4% on the Hamilton Rating Scale for Depression and −53.6 % on the Hamilton Rating Scale for Anxiety). The improvement in Y-BOCS score seen in this study was comparable to the improvement reported in a meta-analysis of DBS studies (45.1% reduction in Y-BOCS score among 116 patients) (70). In another recent study, 10 patients (5 with refractory OCD, 5 with MDD) underwent MRgFUS anterior capsulotomy (71). In these patients, reduced symptoms of OCD (measured with Y-BOCS scores) and MDD (measured with Hamilton Rating Scale for Depression scores) were highly correlated with self-reported frontal function measured with the Frontal Systems Behavior Scale (72), thus indicating successful disruption of pathological function within the frontal-striatal networks by both up-regulating frontal-executive function and down-regulating OCD symptoms. However, this study reported no cognitive impairment after treatment, and therefore could not provide information regarding the upper limit of safety for lesion size. The results from these studies warrant large-scale and sham-controlled clinical trials to broaden our understanding of MRgFUS for the treatment of OCD.

Several clinical trials are ongoing to assess the use of FUS for the treatment of epilepsy, with some using low-intensity focused ultrasound (LIFU) to induce a neuromodulatory effect on the area with the highest epileptogenic activity within the temporal lobe and others using HIFU for thermoablation of a hypothesized epileptogenic focus using MRgFUS. Previous studies regarding FUS neuromodulation have led to the establishment of safe sonication parameters for reversible mechanical disruption of the neural circuit (73–75), and these induced neuromodulatory effects could be stimulatory and inhibitory depending on the targeted brain regions and pulsing schemes. In one clinical trial (NCT03657056, https://clinicaltrials.gov/ct2/show/NCT03657056), the BX Pulsar 1002 is being used to precisely target the epileptic focus in patients who are scheduled for temporal lobe resection. The feasibility and safety of the treatment with both excitatory and inhibitory LIFU parameters will be examined by assessing BOLD functional MRI signal changes throughout the LIFU procedure. Investigators in another clinical trial (NCT03868293, https://clinicaltrials.gov/ct2/show/NCT03868293) are also using the LIFU neuromodulatory effect to treat drug-resistant temporal lobe epilepsy. In this study, patients are undergoing a total of 8 LIFU treatment sessions within 1 month, with researchers assessing treatment efficacy by comparing the number of seizure episodes during and after treatment with the number of episodes in the pretreatment period. Investigators also hope to identify the electrophysiological changes in the epileptic tissue after LIFU neuromodulation and expect to reduce the frequency and/or attenuate the amplitude of epileptiform discharges recorded in electroencephalography data. The unique bimodal modulatory effect of LIFU intervention on the neuronal circuits that may initiate seizure activity may provide an important mapping strategy to identify the seizure focus when combined with electrophysiology or brain imaging readout. In another clinical trial assessing LIFU (NCT 02151175, https://clinicaltrials.gov/ct2/show/NCT02151175), investigators are again using the BX Pulsar 1002 device to study the excitatory and inhibitory effects of stimulation on patients with nondominant mesial temporal lobe epilepsy. The primary endpoint in this study is the safety of the device, which will be assessed by identifying any histological tissue changes. Secondary outcomes include changes in seizure frequency, neurological status, neuropsychological profile, and psychological profile. An initial publication from this group reported that among 8 patients who underwent LIFU sonication of the temporal lobe followed by resection of the affected side, no abnormal histological or neuropsychological changes were observed (76).

The results of another clinical study regarding the use of LIFU in the seizure onset zone were recently published (77). In this study, seizure focus was determined once patients had experienced at least 3 confirmed seizures after stereo-electroencephalography (SEEG) implantation. Patients then underwent LIFU using burst tone and nonthermal parameters with a 10-min exposure time. Sonication occurred while the SEEG electrodes were still implanted; they were removed 3 days later. Of the six patients who underwent treatment, three had no change in seizure frequency, two had a decrease in seizure frequency, and one had an increase in seizure frequency. More cases will need to be evaluated to determine the efficacy of LIFU as a treatment for patients with epilepsy. However, the use of LIFU should not be limited to treatment alone; variations in seizure activity can also aid in diagnosis and confirm the epileptogenic focus.

HIFU is also being investigated for the treatment of epilepsy. Patients with focal epilepsy are currently being recruited for a multicenter clinical trial (NCT02804230, https://clinicaltrials.gov/ct2/show/NCT02804230) that aims to evaluate the feasibility, safety, and initial efficacy of MRgFUS ablation of epileptic foci (defined in this study as temporal sclerosis, dysplasias, and heterotopias). Two other trials (NCT03417297, https://clinicaltrials.gov/ct2/show/NCT03417297 and NCT05032105, https://clinicaltrials.gov/ct2/show/NCT05032105) are assessing the feasibility and safety of using HIFU to ablate the anterior nucleus of the thalamus. One of the trials is focused on determining whether this ablation will help to prevent secondary generalization. The other is assessing the effect of ablation on focal seizure-related anxiety, using functional MRI to evaluate the reactivity of the amygdala to threat.

One of the challenges involved in treating epileptogenic lesions with HIFU is the size limitation of the ablation. Usually <1 cm in any diameter, the convergence of ultrasound waves in HIFU cannot completely ablate a large epileptogenic lesion in a single session. For instance, Yamaguchi et al. (78) described the use of HIFU to ablate a hypothalamic hamartoma in a 26-year-old man with medically refractory epilepsy and gelastic seizures. The hamartoma was too large for ablation, and so the case report details the authors' approach to achieve disconnection. First, electroencephalography was used to identify the location of the patient's seizure activity (right frontal lobe). Diffusion tensor tractography identified connectivity between the hypothalamus and right frontal lobe in the right posterior portion of the hamartoma. This boundary area was subsequently ablated with HIFU, creating a lesion with dimensions of 4.73 mm by 6.46 mm by SI (superior-inferior) 7.73 mm. The patient had no seizures after the ablation and remained seizure free over 1 year of follow-up. This case demonstrates the limitation of HIFU ablation size in epilepsy. However, with the integration of structural connectivity imaging, a disconnective approach could be the optimal treatment strategy.

Another challenge is that the most common type of epilepsy, mesial temporal lobe epilepsy, typically requires ablation of the anterior hippocampus and/or amygdala. Due to the incident angles of the ultrasound beam and the skull at this location, it is very difficult to achieve a high enough treatment efficiency to cause thermal ablation. One group in Japan published the first case report in thermal lesioning at the hippocampus for treating mesial temporal lobe epilepsy (12). Even with the maximum energy and high SDR (0.56), the temperature did not exceed ablative level, and postoperative MRI did not indicate any viable lesion. The patient remained seizure free after 12 months, and the authors theorized that there was a potential neuromodulatory effect due to the subablative temperature.

The International Association for the Study of Pain defines the chronic pain indication as “persistent or recurrent pain lasting longer than 3 months” (79), at which point the pain network will no longer serve as a protective and healing mechanism but will be a pathological condition. Acute pain can occur in any part of the body, but as the pain evolves into a chronic state, pain information from the periphery to the thalamus will drive changes in higher-order brain areas, including reward, motivation, and cognition (80). Altering these widespread brain networks will change the biochemistry of pain transduction and affect the patient's cognitive and emotional experience in pain perception. Unfortunately, the current status of pain management using pharmacotherapy alone is limited to achieving satisfactory pain relief, and the conventional noninvasive brain stimulation modality is still controversial and lacks good scientific data to prove the effectiveness of the treatment (80).

MRgFUS has therapeutic potential for pain management using ablative therapy. In one sham-controlled randomized clinical study (NCT05122403), patients with medication-refractory neuropathic pain are undergoing MRgFUS central lateral thalamotomy followed by a double-blinded assessment regarding treatment effects and adverse events 3 months after treatment. Another ongoing clinical trial (NCT03309813) is targeting bilateral thalamic nuclei with MRgFUS to reduce pain and increase quality of life in patients with chronic trigeminal neuropathic pain. The goal of this randomized, sham-controlled, crossover study is to evaluate the safety and feasibility of treating chronic pain using the MRgFUS lesioning procedure.

Among brain tumors, glioblastoma multiforme (GBM) is the most aggressive and is known to respond poorly to conventional chemotherapies. Contributing to this difficulty is the presence of the BBB, the tight junctions of which impair access of the macromolecular agents into the cellular environment. Additionally, the integrity of the BBB is highly heterogeneous due to the tumor microenvironment (81). Researchers initially found that HIFU could modify the BBB through the use of high intensities similar to those used to create lesions in normal tissue; however, this also led to an increased frequency of hemorrhage and edema (82, 83). Later research demonstrated that a reduction in the acoustic intensity could still achieve BBB opening with a lower incidence of unwanted side effects (84). In 2001, Hynynen et al. (85) found that with the intravenous injection of microbubbles, low-power FUS (ie, within the range of diagnostic ultrasound level, <1.5–2 MPa) could induce transient, reproducible, and localized BBB opening in rabbits without producing any associated neuronal damage. This study opened the door to improved chemotherapy delivery for patients with malignant brain tumors, with this technique allowing successful delivery of chemotherapeutic agents and thereby improving on the low (5%) efficacy of conventional systemic administration (86, 87). The FDA-approved use of concurrent microbubbles as contrast agents in diagnostic imaging thus permits safe doses of FUS intensity to disrupt the BBB (88, 89).

The first-in-human BBB opening was achieved using a transducer surgically implanted into the epidural space superficial to the tumor in patients with recurrent GBM (Table 1) (90). Patients received monthly FUS treatments coupled with intravenous injection of microbubbles for BBB opening. The pressure amplitude began at 0.5 MPa and increased to 1.1 MPa through 5 different doses (0.5, 0.65, 0.8, 0.95, 1.1 MPa), and BBB disruption was found to occur at pressure amplitudes >0.8 MPa. Disruption of the BBB was quantified with gadolinium-enhanced T1-weighted MR images; this technique was chosen based on data from a previous BBB opening study in nonhuman primates (91). Carboplatin, a common chemotherapeutic agent, was used in this study to control the recurrent GBM. This study was limited by the need to surgically implant the ultrasound transducer. Additionally, the transducer was unfocused and unable to electronically steer the beam after surgical implantation.

A few years later, researchers performed a first-in-human trial of noninvasive MRgFUS BBB opening in patients with malignant glioma, using concurrent systemic administration of temozolomide chemotherapy (Table 1) (92). T1-weighted MR images demonstrated a 15% to 50% increase in signal enhancement, indicating transient BBB opening in the target tissue (Figure 11). Approximately 24 h after FUS and chemotherapeutic administration, the patients underwent craniotomy and tumor resection. Sonicated and unsonicated peritumor tissue samples were collected and the tissue chemotherapy concentrations were measured. Note that during the trial, the chemotherapy agent was switched from liposomal doxorubicin to temozolomide, and limited resectable tumor volume in three of five patients prevented statistical analysis of the tumor samples. Nevertheless, the researchers observed a chemotherapy concentration that was 7.7 times higher in the sonicated peritumor tissue than in the unsonicated peritumor tissue in one patient.

Another group of researchers subsequently tried to enhance the treatment effect by creating multiple BBB openings with MRgFUS (Table 1) (93). In this study, 6 patients who underwent a gross total resection of malignant glioma received 6 cycles of temozolomide with associated FUS BBB opening performed at the beginning of each 4-week cycle. Patients underwent follow-up MRI 1 year after the first chemotherapy cycle (6 months after the last chemotherapy cycle), and there was no evidence of any FUS-related adverse effects.

In 2021, another study demonstrated the feasibility and safety of using NaviFUS, a frameless novel device that integrates neuronavigation and an FUS system, in patients with GBM (94). Six patients were assigned to one of three different ultrasound doses in the mechanical index (0.48, 0.58, or 0.68) to temporarily open the BBB. The lowest dose used (0.48) was previously identified as the threshold of BBB opening in nonclinical studies (96); the maximum dose of 0.68 was chosen based on Good Laboratory Practice safety tests (97). Dynamic contrast-enhanced MRI was performed immediately after and 24 h after the BBB opening procedure and demonstrated the efficacy of NaviFUS BBB opening. T2-weighted images were obtained to evaluate any hemorrhages associated with BBB opening. All patients were scheduled for tumor resection surgery within 2 weeks after the FUS BBB opening, and clinical visits for follow-up were performed routinely until the third week after BBB opening to assess physical and neurological functions.

AD represents an enormous societal and healthcare burden as the population ages. Still, the development of new pharmacotherapeutics provides diminishing returns, as these drugs are restricted from entering the brain by the BBB. FUS temporarily loosens the BBB tight junction, allowing the delivery of therapeutic agents to the sonicated brain area. However, even without these therapeutics, studies have reported that BBB opening alone triggers a significant reduction of Aβ deposition through microglia activation (98, 99). Therefore, researchers have assessed the use of potential therapeutics such as a GSK-3 inhibitor and RN2N as an additive strategy that can further increase the therapeutic benefit of BBB opening (100). These preclinical studies using transgenic mouse models of AD have demonstrated improvements in Aβ plaque clearance and up-regulation of cognitive function in AD pathology after opening of the BBB. Based on these findings, researchers assessed the use of FUS in 4 patients with AD, and clinical and radiographic evaluations in these patients demonstrated reversible, repeatable, and safe noninvasive opening of the BBB with FUS (Table 1) (95). The researchers in this study targeted the superior frontal gyrus white matter of the dorsolateral prefrontal cortex to reduce the risk of adverse events. Note that a [18F]-florbetaben positron emission tomography scan performed 1 week after both the first and second sonication could not demonstrate a clear effect of BBB opening on Aβ clearance. Additionally, the small sample size limited the conclusions that could be drawn regarding the safety and efficacy of this treatment, as well as whether FUS BBB opening affected the clinical and pathological symptoms of AD.

Another research group developed a novel strategy in which FUS was delivered to deep brain regions without tissue ablation or BBB opening (101). The researchers used single ultra-short pulses (3 μs) instead of conventional pulses (100 ms) (75, 102) at 5-Hz pulse repetition frequency, and used 6,000 pulse numbers per session. This approach was found to be safe and effective in a preclinical study using an energy level of 0.3 mJ mm⁻²; the energy level was decreased to 0.2 mJ mm⁻² for clinical purposes. In this study, 35 patients with mild AD treated at 2 separate clinics underwent FUS. In patients from clinic 1, researchers targeted the AD brain network, which included the dorsolateral prefrontal cortex, inferior frontal cortex, and language areas extending to Broca's and Wernicke's areas. In patients from clinic 2, researchers performed global cortical stimulation by distributing the total sonication energy over all brain areas by moving the headpiece probe over the scalp in a circular trajectory. Neuropsychological changes after therapy were evaluated with Consortium to Establish a Registry for Alzheimer's disease (CERAD) scores. Study patients demonstrated significant improvements in CERAD corrected total scores and logistic regression scores after treatment, and these improvements remained consistent over 3 months. Principal component analysis was also performed to assess CERAD-derived cognitive measurements of learning and memory, verbal skills, and visuospatial processing. Patients from clinic 1 demonstrated improvements in learning/memory and verbal skills lasting up to 3 months, and showed a decline in visuospatial processing. However, patients from clinic 2 demonstrated no significant change in visuospatial processing. This absence of stimulation effect could be due to the lack of stimulation of the occipito-parietal region in patients from clinic 2.

FUS also holds potential for the treatment of ALS. ALS is a devastating and incurable neurological illness, and medical advances have been incremental. As with brain tumors, the BBB is a pharmacologic barrier to potentially effective treatments for ALS. To this end, researchers assessed the use of MRgFUS to open the BBB in patients with ALS and demonstrated successful results (103). In this research, the brain region targeted for BBB opening was the eloquent primary motor cortex, and the process was found to be safe, feasible, and reversible. For patients with ALS, BBB opening is used to introduce agents such as nonviral vectors that transport therapeutic genetic elements into neurons; it is therefore essential that these agents are not damaged as they travel through the BBB (104).