FAT1 weighted MRI: Diffusion meets anatomical imaging and application in thalamic surgery for tremor

Patient-specific targeting of the Ventral intermediate nucleus (Vim) of the thalamus can be achieved with MR connectivity. Nevertheless, there are several drawbacks to using tractography-based targeting methods to visualise distinct thalamic nuclei (e.g., subjective region of interest selection, and thresholding of resulting tracts and clusters). Fractional anisotropy (FA) mapping, another product of diffusion MRI (dMRI), does not rely on tractography, and could thus be clinically more viable for discerning thalamic anatomy for stereotactic surgery. The aim of this study is to develop and present a hybrid, high-resolution, and high-fidelity imaging modality that combines contrast from FA maps as well as anatomical T1 sequences (FAT1 imaging); and to evaluate FAT1 based Vim-target definition. Imaging and outcome data of 35 consecutive refractory tremor patients who had undergone 43 connectivity guided deep brain stimulation (DBS) and/or radiofrequency thermocoagulation (RF-T) between 2013 and 2021 were included. First, the pre-operatively acquired dMRI and MPRAGE sequences were used to create FAT1 maps in retrospect. Then, an FAT1 based Vim-target was planned by an experienced functional neurosurgeon who was blinded for patient outcome. Finally, to investigate FAT1 based targeting, a post-hoc analysis was carried out of the degree of overlap between the newly created FAT1 based Vim-target, and the volume of tissue activation (VTA, in case of DBS) or lesion volume (in case of RF-T). This degree of overlap was compared between favourable and unfavourable outcome groups: outcomes were measured by experts blinded for imaging data at the last follow-up using a Clinical Global Impression-Improvement score (CGI-I), where a CGI-I score of 1-2 (i.e., FTMTRS improvement of ≥50%) was considered favourable. In 36 of the 43 (84%) performed surgeries (24 DBS and 19 RF-T), FAT1 based Vim-targeting was possible. For the group showing favourable outcome (71% of the patients at a median follow-up of 13 months), the mean amount of overlap between the FAT1 based Vim-target and the VTA or lesion was 42% (±13), versus 17% (±15) for patients with an unfavourable outcome (MD 25%, 95% CI 14-35, p < 0.0001). Retrospective use of FAT1 based Vim-targeting as a tool to predict outcome had a sensitivity of 90%, specificity of 80%, positive predictive value of 90%, and negative predictive value of 80%. In conclusion, FAT1 imaging is a new, high-resolution, and high-fidelity modality that combines diffusion and anatomical MRI. It provides a fast and efficacious way of targeting the ventral intermediate nucleus of the thalamus. In this study, FAT1 based targeting was highly accurate in predicting outcomes after deep brain stimulation and radiofrequency thalamotomy.