Brown University

Artificial Intelligence for Targeting the Ventral Intermediate Nucleus in Focused Ultrasound Thalamotomy: A Step Toward Safer and More Precise Treatment of Essential Tremor

Description

Abstract:
"Introduction Focused ultrasound (FUS) thalamotomy is an effective, minimally invasive surgery for essential tremor (ET). However, this procedure causes chronic sensory deficits in approximately 10% of patients, likely due to involvement of posterior sensory thalamic nuclei. This study compared thalamic targets predicted by an artificial intelligence (AI) to the locations chosen by neurosurgeons, and whether AI-lesion vector offsets were associated with persistent sensory deficits. Methods Fourteen ET patients who successfully underwent FUS thalamotomy but developed chronic sensory deficits (median follow-up 12 months; IQR 11.3-14.7 months) were matched to fourteen similar patients without deficits (median follow-up 11 months; IQR 9.9-12.5 months). All patients were treated at a single institution between the years 2011 and 2025. MRI-derived coordinates of lesion locations were retrospectively compared to AI-predicted locations (RebrAIn OptimMRI software). Differences in lesion characteristics and tremor outcomes were analyzed using appropriate statistical and regression methods. Results Tremor was less improved in patients with sensory deficits (58.2 ± 45.5%) than in those without (84.5 ± 29.1%, p = 0.035). Total recurrence was also more frequent in this group (5/14 vs 1/14), though not significantly (p = 0.16). AI-predicted lesions were significantly more anterior in sensory deficit patients (mean offset 1.09 ± 0.62 mm) compared to the control group (0.43 ± 0.57 mm, p = 0.007). Lesions were also positioned more laterally in the sensory deficit group (14.7 ± 0.8mm vs 13.87 ± 0.54mm from midline, p = 0.005), although the AI-lesion offset along this axis did not differ between groups. Regression analyses showed no association between these findings and lesion volume. There were no significant differences in lesion characteristics between groups along the superior–inferior axis. Conclusions AI correctly predicted more anterior targets in patients with chronic sensory deficits and worse tremor outcomes following FUS thalamotomy. These findings suggest that AI-assisted targeting could improve FUS thalamotomy accuracy and reduce sensory complications caused by suboptimal lesion placement. "

Access Conditions

Use and Reproduction
All rights reserved
Rights
In Copyright
Restrictions on Use
All Rights Reserved

Citation

Mancini, Evan , Beck, Daniel , Grogan, Dayton , Dominguez, Martin , Gassa, Narimane , Moreno, Antoine , Zemzemi, Nejib , Cuny, Emmanuel , Moosa, Shayan, "Artificial Intelligence for Targeting the Ventral Intermediate Nucleus in Focused Ultrasound Thalamotomy: A Step Toward Safer and More Precise Treatment of Essential Tremor" (2025). Student Neurosurgery/Neurology Research Conference. Brown Digital Repository. Brown University Library. https://doi.org/10.26300/2345-a539

Relations

Collection:

  • Student Neurosurgery/Neurology Research Conference

    Brown University Department of Neurosurgery and Department of Neurology and Warren Alpert Medical School annually hosts the Student Neurosurgery/Neurology Research Conference, a national conference for medical students conducting research in neurosurgery/neurology or related neuroscience fields.
    ...