Comparison of Anatomical Landmarks and Atlas- and Automated Atlas Segmentation Application-Based Targeting in...

MRgFUS Vim thalamotomy has recently been developed as a novel treatment for various movement disorders. An automated atlas segmentation application (BRAINLAB® Elements) enables individualized planning by registering atlas-based Vim segmentation to each patient's MRI. We investigated whether automated atlas segmentation application-based preoperative planning affects the treatment outcomes of MRgFUS Vim thalamotomy.

Between January 2019 and May 2022, patients with refractory essential tremor or tremor-dominant Parkinson disease underwent MRgFUS Vim thalamotomy at Ohnishi Neurological Center group, planning was performed according to anatomical landmarks and atlas, and in the segmentation-based group, planning was conducted using automated atlas segmentation application. We compared distributions of the center of the target at preoperative planning and coordinate deviation between preoperative planning and final lesion. Treatment outcome was also evaluated.

A total of 80 consecutive patients were included, with 40 patients in the conventional group and 40 patients in the segmentation-based group. The clinical improvement ratio of the Clinical Rating Scale for Tremor A + B subscore did not differ significantly between the 2 groups. The number of intraoperative target adjustments was significantly lower in the segmentation-based group (1.3 ± 0.9 vs 0.9 ± 0.7, 95% CI -0.8 to -0.1; P = .011). The number of sonications was significantly lower in the segmentation-based group (9.0 ± 2.5 vs 6.7 ± 1.8 times, 95% CI -3.3 to -1.3; P < .001).

Although it is difficult to eliminate the influence of the learning curve, our retrospective analysis revealed that automated atlas segmentation application-based preoperative planning and visualization requires fewer intraoperative target adjustments and sonications than anatomical landmarks and atlas-based preoperative planning in MRgFUS Vim thalamotomy.