MRI follow-up after magnetic resonance-guided focused ultrasound for non-invasive thalamotomy: the neuroradiologist's perspective

Abstract

Purpose: Magnetic resonance-guided focused ultrasound (MRgFUS) systems are increasingly used to non-invasively treat tremor; consensus on imaging follow-up is poor in these patients. This study aims to elucidate how MRgFUS lesions evolve for a radiological readership with regard to clinical outcome.

Methods: MRgFUS-induced lesions and oedema were retrospectively evaluated based on DWI, SWI, T2-weighted and T1-weighted 3-T MRI data acquired 30 min and 3, 30 and 180 days after MRgFUS (n = 9 essential tremor, n = 1 Parkinson's patients). Lesions were assessed volumetrically, visually and by ADC measurements and compared with clinical effects using non-parametric testing.

Results: Thirty minutes after treatment, all lesions could be identified on T2-weighted images. Immediate oedema was rare (n = 1). Lesion volume as well as oedema reached a maximum on day 3 with a mean lesion size of 0.4 ± 0.2 cm³ and an oedema volume 3.7 ± 1.2 times the lesion volume. On day 3, a distinct diffusion-restricted rim was noted that corresponded well with SWI. Lesion shrinkage after day 3 was observed in all sequences. Lesions were no longer detectable on DWI in n = 7/10, on T2-weighted images in n = 4/10 and on T1-weighted images in n = 4/10 on day 180. No infarcts or haemorrhage were observed. There was no correlation between lesion size and initial motor skill improvement (p = 0.99). Tremor reduction dynamics correlated strongly with lesion shrinkage between days 3 and 180 (p = 0.01, R = 0.76).

Conclusion: In conclusion, cerebral MRgFUS lesions variably shrink over months. SWI is the sequence of choice to identify lesions after 6 months. Lesion volume is arguably associated with intermediate-term outcome.

Keywords: Essential tremor; High-intensity focused ultrasound ablation; Magnetic resonance imaging; Parkinson disease.

Fig. 1

T1-weighted Vim lesion image on day 3. The left-hemispherical ventral intermediate nucleus (Vim) lesion was expected about 14 mm left and centrally 1 mm on the anterior-posterior commissure line (AC-PC line) approximately 25% of the total distance ventral of the PC. The enlarged scheme illustrates the expected necrotic core (N), the ring-shaped cytotoxic oedema zone (C) and the blurred vasogenic oedema (V)

Fig. 2

Lesion dynamics at 3-T on T2-weighted and susceptibility-weighted images over time. a–d T2-weighted images and e–g susceptibility-weighted images. a Thirty minutes after therapy, b 3 days, c 30 days and d 180 days after therapy. e Three days, f 30 days and g 180 days after therapy. Note how the lesion first increases in size during the first 3 days (a vs. b), while it has completely vanished half a year after therapy on T2-weighted images and yet remains distinctly visible on susceptibility-weighted images. Image b also depicts the oedema surrounding the core lesion separated by a fine hypointense dark rim. The noisy aspect of image a is due to the distant MRI-integrated body coil used, while the patient was still wearing the treatment helmet

Fig. 3

Temporal evolution of ventral intermediate nucleus lesion volumes and corresponding apparent diffusion coefficients (ADCs). Dimensionless relative values are defined as relative to the volume measured on day 3 by building the ration (x_t / x_d = 3). a Relative lesion volume dynamics measured on T2-weighted images in 10 patients. b Absolute lesion volumes on T2-weighted images marked by mean and 5th to 95th percentiles. c Relative lesion volume dynamics measured on susceptibility-weighted images. d Absolute lesion volumes on susceptibility images marked by mean and 5^th to 95^th percentiles. e Absolute ADC values at different time points stated as mean and 5^th to 95^th percentiles. f The span of ADC values describes the difference between the minimum and maximum ADC measured in a lesion region of interest at a point of time. D0: day 0 MRI 30 min after therapy, D3: MRI on day 3 after therapy, D30: MRI 30 days after therapy, D180: MRI 180 days after therapy

Fig. 4

Fading of the MRgFUS lesion on T1-weighted images. While on SWI and T2-weighted images shrinkage of the lesion was noted, the lesion rather faded in signal intensity compared with the surrounding brain tissue. a Day 3, b day 30 (black arrows indicating lesion margins) and c “vanished” lesion on day 180

Fig. 5

Lesion presentation on diffusion-weighted images on day 3 after therapy and corresponding susceptibility-weighted image. ab = 1000 s/mm² diffusion-weighted image depicting a characteristic bright ring formation inside the lesion. b According to the low intensity on the apparent diffusion coefficient map, the lesion has restricted diffusivity. c The ring shape corresponds well to the speckled ring shape of the lesion observed on susceptibility-weighted images of the same person

Fig. 6

a Reduction of motor symptoms in general motor task testing as opposed to baseline (baseline score minus post-treatment score/baseline score) correlated with lesion volume both on day 3. b Reduction of tremor symptoms on the treated side as opposed to baseline (baseline score minus post-treatment score/baseline score) correlated with lesion volume both on day 3. c Relative reduction of motor symptoms in general motor task testing over time. d Relative reduction of tremor symptoms on the treated side over time. e Dynamics of relative changes in tremor reduction (day 180 versus day 3) as opposed to dynamics in T2 volume between day 180 and day 3. A − 1.00 on the x-axis represents a lesion that is no longer discernable. D3: MRI on day 3 after therapy, D30: MRI 30 days after therapy, D180: MRI 180 days after therapy, T2w: T2-weighted volumetry