Consistency is key: influence of skull density ratio distribution on the formation of clinically effective lesions and long-term tremor suppression following treatment with MR-guided focused ultrasound

Abstract

OBJECTIVE Skull density ratio (SDR) influences the permeability of the skull to the ultrasound waves used in magnetic resonance–guided focused ultrasound (MRgFUS) for the treatment of tremor. SDR values vary across the skull and the mean value is known to be predictive of sonication thermal increase. The aim of this investigation was to explore the effects of the SDR distribution on clinical outcomes following treatment with MRgFUS. METHODS Data from 61 patients with essential or dystonic tremor treated with MRgFUS targeting the ventral intermediate nucleus (Vim) were retrospectively analyzed. Tremor suppression was assessed using the Clinical Rating Scale for Tremor (CRST) and hand tremor score (HTS). Vim ablation volume was measured on the T1-weighted MR image acquired both at 1 day and 12 months after treatment. The numerical distribution of SDR values measured for each element in the ultrasound transducer was quantified by calculating the mean, standard deviation, skewness, entropy, and kurtosis of the SDR histogram. The effect of the SDR metrics on change in CRST and HTS was examined using a linear mixed-effects model. Additionally, the effect of the regional distribution of SDR values was explored in an element-wise analysis between patients with above- and below-average tremor suppression. RESULTS A significant positive effect was found between SDR kurtosis and improvement in CRST (β = 0.33, p = 0.004) and HTS (β = 0.38, p < 0.001). The effect was found to be significant at 1 month posttreatment (CRST: β = 0.415, p = 0.008; HTS: β = 0.369, p = 0.016), and at the most recent clinical follow-up (CRST: β = 0.395, p < 0.001; HTS: β = 0.386, p < 0.001). One hundred seventy-one significant elements were identified in the element-wise analysis. The mean percentage difference from the mean SDR in these elements was associated with improvement in CRST (β = 0.27, p < 0.008) and HTS (β = 0.27, p < 0.015). Higher SDR kurtosis was associated with increased lesion volume at 12 months (p = 0.040) and less reduction in volume relative to the day-1 lesion volume (p = 0.007) CONCLUSIONS Greater SDR kurtosis was associated with larger, more stable lesions at 12 months posttreatment and increased tremor suppression at long-term follow-up. SDR kurtosis may provide a more meaningful prognostic factor than the mean SDR.