Longitudinal MR imaging after unilateral MR-guided focused ultrasound thalamotomy: clinical and radiological correlation.

Abstract

Magnetic-resonance-guided focused ultrasound (MRgFUS) thalamotomy uses multiple converging high-energy ultrasonic beams to produce thermal lesions in the thalamus. Early postoperative MR imaging demonstrates the location and extent of the lesion, but there is no consensus on the utility or frequency of postoperative imaging. We aimed to evaluate the evolution of MRgFUS lesions and describe the incidence, predictors, and clinical effects of lesion persistence in a large patient cohort. A total of 215 unilateral MRgFUS thalamotomy procedures for essential tremor (ET) by a single surgeon were retrospectively analyzed. All patients had MR imaging 1 day postoperatively; 106 had imaging at 3 months and 32 had imaging at 1 year. Thin cut (2 mm) axial and coronal T2-weighted MRIs at these timepoints were analyzed visually on a binary scale for lesion presence and when visible, lesion volumes were measured. SWI and DWI sequences were also analyzed when available. Clinical outcomes including tremor scores and side effects were recorded at these same time points. We analyzed if patient characteristics (age, skull density ratio), preoperative tremor score, and sonication parameters influenced lesion evolution and if imaging characteristics correlated with clinical outcomes. Visible lesions were present in all patients 1 day post- MRgFUS and measured 307.4 ± 128.7 mm3. At 3 months, residual lesions (excluding patients where lesions were not visible) were 83.6% smaller and detectable in only 54.7% of patients (n = 58). At 1 year, residual lesions were detected in 50.0% of patients (n = 16) and were 90.7% smaller than 24 h and 46.5% smaller than 3 months. Lesions were more frequently visible on SWI (100%, n = 17), DWI (n = 38, 97.4%) and ADC (n = 36, 92.3%). At 3 months, fewer treatment sonications, higher maximum power, and greater distance between individual sonications led to larger lesion volumes. Volume at 24 h did not predict if a lesion was visible later. Lesion visibility at 3 months predicted sensory side effects but was not correlated with tremor outcomes. Overall, lesions are visible on T2-weighted MRI in about half of patients at both 3 months and 1 year post-MRgFUS thalamotomy. Certain sonication parameters significantly predicted persistent volume, but residual lesions did not correlate with tremor outcomes.