Loading ...

The largest study to date of magnetic resonance-guided focused ultrasound (MRgFUS) thalamotomy for essential tremor identified a target area within the ventral intermediate nucleus (VIM) that could significantly differentiate patients with optimal vs suboptimal outcomes at 1 year.

“We now have a very clear map of where to sonicate for optimal outcomes and where not to sonicate to avoid side effects,” senior author Andreas Horn, MD, PhD, director of Deep Brain Stimulation Research, Brigham and Women’s Hospital, Boston, told Medscape Medical News. “We also have a better understanding of the structural connection in both tremor control and the occurrence of side effects.”

Studies show unilateral MRgFUS thalamotomy can result in about a 40% improvement in the Clinical Rating Scale for Tremor at 3 months, but transient or permanent adverse sensory, motor, speech, and gait effects can occur in up to 40% of patients.

Several novel imaging studies have investigated optimal targets, or sweet spots, but to date, there is no optimal or standardized protocol. Recent work using probabilistic mapping of ablative effects, or sweet spot mapping, supports the VIM proper or its ventral border as the optimal target, but little quantitative evidence exists on sites associated with side effects, and large cohort studies are lacking, the authors noted.

The findings were published online on May 14 in Science Advances.

Discovery and Validation Cohorts 

For the study, the researchers calculated sweet spots and side effects, or sour spots, using prospective data from 200 patients (mean age, 75 years; 68% men) who underwent unilateral MRgFUS thalamotomy for medically refractory essential tremor at Brigham and Women’s Hospital. Their total Fahn-Tolosa-Marin (FTM) Tremor Rating Scale score was 7.1, and the average disease duration was 27.6 years.

The main tremor sweet spot result was then validated in an additional 102 patients from the center in Boston and 49 patients who underwent MRgFUS thalamotomy in Toronto, Ontario, Canada.

Tremor was reduced by 94.9% at postoperative day 1 and by 91.6% at 3 months. At 1 year, the average overall FTM improvement was 86% and remained significant (P < .001).

The area of optimal tremor control at 1 year was identified at the center towards the anterior portion of the VIM and intersected with the cerebellothalamic tract (CTT) as defined by the Deep Brain Stimulation (DBS) Tractography Atlas and the Basal Ganglia Pathway Atlas.

Fiber filtering analysis revealed that 70% of tracts in the DBS atlas and 100% in the Basal atlas associated with optimal tumor control corresponded to the CTT and remained significant after Bonferroni correction (P < .05).

The team then tested whether the sweet spot could be used as a model to predict outcomes in the two validation cohorts and found that weighted average lesion overlaps significantly correlated with the percentage of tremor improvements at 3 months and 1 year.

“Indeed, the more this optimal site was lesioned, the better the tumor response was in these independent cohorts,” Horn said. “This has so far never been done in the field of MRgFUS.”

The findings were further validated in a cohort of six cases that underwent a second treatment, which adds unique insights into cases that had unsuccessful and successful treatment sessions, he added.

‘Sour Spots’ and Lesion Location

Side effects occurred in 75.88% of patients on day 1 post-thalamotomy, 52.44% at 3 months, and 37.5% at 1 year. The most common side effect on postoperative day 1 was gait imbalance (63.3%), followed by sensory deficits (24.6%), dysarthria (16.6%), and weakness (10.6%). Gait imbalance remained the most common side effect at 3 months (26.7%) and 1 year (17.1%). All side effects were mild.

Lesions extending ventrolaterally toward the inferior region of the VIM were associated with an increased risk for gait imbalance. Dysarthria was associated with lesions extending superomedially.

Sensory side effects and dysgeusia were associated with lesions extending posteriorly and medially into the ventral posterolateral and ventral posteromedial nuclei. Lesions extending laterally into the internal capsule were associated with weakness.

All sour spots and tracts were significant after Bonferroni correction (P < .05), except dysarthria and dysgeusia, which were significant at the uncorrected level.

Quantitative analysis of tracts revealed 100% correspondence to the medial lemniscus for sensory deficits, 52% to the corticospinal tract (CST) for weakness, 44% to the CTT and 16% to the CST for imbalance, and 55% correspondence to the trigeminothalamic tract and medial lemniscus for dysgeusia.

Larger lesion sizes were associated with stronger tremor improvement but also with side effects, “suggesting a cost-benefit trade-off,” Horn and colleagues noted.

Next Steps, Open Questions

“The present work identifies a target site that may be critical to sonicate for optimal response and can be transferred into individual patient anatomy using the methods developed in the present article,” Horn said.

“A next step will be to prospectively test results in future cohorts. If successful, this could help streamline and somewhat standardize the process of FUS [focused ultrasound] surgeries for thalamotomies,” he added.

He noted that it remains unclear whether the identified sweet spot is the total volume needed to sonicate, in which case lesions could be rather small, or whether, more likely, it is the differentiating portion of the tissue that needs to be lesioned in addition to adjacent regions.

Reached for comment, Bhavya R. Shah, MD, director of the Transcranial-FUS Lab, UT Southwestern Medical Center, Dallas, told Medscape Medical News that the retrospective analysis “supports the premise that targeting discrete and specific brain regions while avoiding others can improve tremor response and reduce adverse effects.”

The challenge with the paper, he said, is that it uses indirect or landmark-based targeting to treat patients and a combination of postoperative imaging, an atlas of population-based white matter tracts, and image registration approaches to determine where the final lesion was, what tracts were in the target zone, and clinical outcomes associated with the affected tracts.

“Indirect targeting methods suffer from inaccuracies based on human anatomic variation, subjectivity, experience, as well as underlying pathological derangement of the thalamus and surrounding structures,” Shah said.

“This work and others like it support the premise of prospectively using this anatomic knowledge to perform targeting instead of indirect targeting,” he added.

In 2022, his team reported initial results in 18 patients with essential tremor using four-tract tractography, which uses advanced imaging techniques to map discrete patient-specific anatomy to prospectively identify brain regions for treatment. This overcomes some of the concerns with indirect targeting and has shown benefit in a recent case study and independent retrospective analysis of 43 patients, Shah noted.

Asked about the technique, Horn said, “we see value in novel approaches such as four-tract tractography but see issues in the diffusion MRI sequences used in such approaches, which are typically prone to distortion artifacts and of poor resolution, ie, are not perfectly suited for stereotactic surgery, in which millimeters matter.”

“The use of specialized sequences such as the white-matter-nulled or FGATIR (fast gray matter acquisition T1 inversion recovery) sequences, as used here, have less of these limitations and still beautifully visualizes the key tracts such as the cerebellothalamic connections,” he said.

The study was funded by National Institutes of Health grants, a Neuronentics grant, the Kaye Family Research Endowment, the Ellison/Baszucki Family Foundation, the Manley Family, a FUS Foundation grant, and the Schilling Foundation. Horn reported receiving lecture fees from Boston Scientific, consultancy fees for FxNeuromodulation and Abbott, and serving as co-inventor of a patent granted to Charité University Medicine, Berlin, Germany, which covers multisystem DBS fiber filtering and an automated DBS algorithm unrelated to this work. Co-author disclosures are listed in the paper. Shah reported no relevant financial disclosures.