501 Patient-Specific Targeting Method for Improved Quality-of-Life Outcomes from MRgFUS Treatment of Essential Tremor

Abstract

INTRODUCTION: MRI-guided-focused-ultrasound (MRgFUS) ablation of the ventral intermediate nucleus (VIM) is an FDA-approved treatment for essential tremor (ET). However, the VIM cannot be directly visualized on standard imaging, and the ideal ablation zone within the putative VIM varies inter-individually. Suboptimal lesion location can lead to side effects that impair quality of life (QoL). Patient-specific methods to locate the “go-zone” for MRgFUS ablation for ET that can predict QoL outcome could provide significant benefit over current methods. METHODS: We retrospectively studied 26 ET patients treated with MRgFUS at Stanford. Probabilistic tracking from the thalamus to M1 using each subject’s standard, low-resolution preoperative diffusion MRI generated patient-specific cerebello-thalamic-cortical tracts. Within the thalamus, the personalized hotspot with high probability of being connected to M1 defined the “go-zone.” The normative DISTAL-atlas VIM, which some centers use to target deep brain stimulation, provided a comparison. The lesion was segmented in postoperative MRIs and transformed to the preoperative-MRI space. Similarity/overlap between the “go-zone” and lesion was quantified using the Sorensen-Dice coefficient. Postoperative QoL ("better," "same," or "worse") was determined from each subject’s last-available follow-up. RESULTS: Probabilistic thalamus-to-M1 tracking robustly identified cerebello-thalamo-cortical tracts using standard, clinical diffusion MRIs. Lesion similarity to the patient-specific tractographic "go-zone" was associated with better QoL outcomes. Dice coefficients of the "worse" group were significantly lower than of the combined “same”/”better” group (p = 0.014). This was not seen with the DISTAL-atlas VIM. CONCLUSIONS: We proposed a clinically-feasible probabilistic-tractography-based targeting method for MRgFUS treatment of ET. Lesion similarity to the personalized “go-zone” is associated with better QoL outcome, unlike similarity to the normative DISTAL-atlas VIM. This illustrates the value of our patient-specific targeting approach.