P 23 The overlap of lesions placed by MRIgFUS with the cerebellothalamic tract does not adequately explain symptom control in ET

Abstract

Background: The thalamic ventral intermediate (Vim) nucleus is part of a network that appears to be disrupted in essential tremor (ET) and is the most relevant target of functional neurosurgery to alleviate ET symptoms. As the most important afferent of the VIM, the cerebellothalamic tract (CTT) transmits information from cerebellar nuclei. Disruption of this tract - in the past via subthalatomies and in the present via deep brain stimulation is considered the best option of tremor treatment, partly because of the high anatomical density of CTT fibers before entering the VIM, and partly because of the greater distance to sensory thalamic nuclei. Recently, initial pilot studies have also documented transcranial MRI-guided high-intensity focused ultrasound (MRIgFUS) as an effective treatment modality. Unilateral MRIgFUS lesioning was highly efficacious in reducing contralateral hand tremor in ET with mild and transient adverse effects. However, targeting of MRIgFUS has so far been based on MRI imaging without dedicated evidence of the underlying connectome. Here, we test a method for quantitative analysis of the involved fiber tracts and test whether the measured clinical effect correlates with overlap of the set lesion with individual fiber tracts, such as CTT. Materials and Methods: We revisited pre- and post-procedural MRIs of 6 ET patients (Øage 70,7 (± 8,5) years , Ø Tremor rating (TR) (pre): 43,8 (± 9,8)) who underwent unilateral ablation of the CTT by MRIgFUS. The images, obtained 48 hours (48H) and 3-6 months (3M+) after the procedure, were used to create 3D-models of the lesions which were localized within MNI-space with the pre-procedural MRI and atlas registration using Suretune®. The models were subsequently used to calculate interferences with fiber tracts. A linear model was employed to assess the correlation between the quantitative interference and improvement of the clinical score at two timepoints after the procedure (48H, n=6: TR: 16,7 (±5,5) ; 3M+, n= 5: TR: 19,2 (±7,2)). Results: The lesions volume differed between the 48H (Ø 281.78mm3(±77.77), n=6) and 3M+(Ø 53.82mm3(±22), n=5) conditions (p<0.01). After 48H primarily the CTT (Ø72,4 % (±34)) but also the M1 (Ø36,05 % (±43)) interfered, while after 3M+ the overlap reduced tremendously for the CTT (Ø1% (±1,6), p=0.017) and slightly for the M1(Ø26,7% (±28,4), p=0.67). Preliminary results could not show a correlation between the observed clinical improvement with the interference of the lesions with a specific tract. Discussion: Interestingly, there is no clear correlation between clinical outcome and the overlap between the lesion placed by MRIgFUS and the fibers of the CTT. This might indicate that the observed clinical effect would have to be explained only partially by the targeted CTT but also by other fiber tracts. The data set was too small for a meaningful multivariate regression analysis, but pooling data sets from different centers may mean a solution to this problem in the future.