Rao P Gullapalli1, Jiachen Zhuo1, Dheeraj Gandhi1, Charlene Aldrich2, Erma Owens1, John Hebel1, Paul Fishman3, Howard Eisenberg2, and Elias Melhem1
1Diagnostic Radiology & Nuclear Medicine, University of Maryland School of Medicine, Baltimore, MD, United States, 2Neurosurgery, University of Maryland School of Medicine, Baltimore, MD, United States, 3Neurology, University of Maryland School of Medicine, Baltimore, MD, United States
Synopsis
In the context of the remarkable reduction in tremors and improvement
in quality of life at one year following MRgFUS thalamotomy procedure to treat
Essential Tremors in a recently concluded multi-center trial, we examined pre-
and post-imaging data including an assessment of the accuracy of MRgFUS
targeting of the VIM nuclei and lesion evolution over 12 months. Lesions generated by this procedure were accurately
placed, and matched well with the known location of VIM nucleus based on
anatomical atlases. The lesions appear to regress in size over a 12 month
period but the therapeutic effect is maintained.Essential tremor (ET) is the most common movement disorder with an
estimated prevalence of about 0.3% and 5.55% and increases with age. Both RF and deep brain stimulation techniques
have proven to be effective in treating ET. Recent advances in focused
ultrasound technology with MR guidance make it possible to target regions in
the brain with great accuracy. Recently
concluded Pivotal Study for thalamotomy treatment of medication refractory
essential tremor subjects demonstrated remarkable reduction in tremors and
improved quality of life for the patients immediately following treatment that persisted
up to one year. In the context of the
remarkable improvement in quality of life at one year after the MRgFUS
thalamotomy procedure, we performed preliminary examination of the pre- and
post-imaging data including an assessment of the accuracy of MRgFUS targeting
of the VIM nuclei.
Methods
In the multi-center trial, Seventy-six patients from
multiple centers with severe, medication-refractory essential tremor were
randomized 3:1 to receive a unilateral MR-guided focused ultrasound thalamotomy
or a sham procedure.
Combined data from the multi-center trial revealed that contralateral hand
tremor, the primary endpoint, improved 49% at 3 months (p<0.001) with the
treatment (18.09 +4.81 to 9.55 + 5.06) compared to no change in sham
procedures (16.00 + 4.42 to 15.75 + 4.90), and the effect was maintained
at one year (10.89+4.86). Similarly,
functional measures of disability and quality of life reported by the patients were
statistically improved compared to sham cohort.
In the Maryland subset of patients (n=9) we examined the pre- and post imaging
data to asses targeting accuracy with respect to stereotactic location of VIM
and lesion evolution over 12 months. MRImages
were obtained on a 3T system before the HIFU treatment (V1), post-treatment (within
24hr) (V2) and at 12 months (V3) using a 12-channel head coil. 3D T1-weighted-MPRAGE (TE = 2.91ms, TR =
2300ms, TI = 900ms, flip angle = 9⁰) and 3D T2-weighted SPACE (TE = 222ms, TR =
3000ms, ETL = 133) were both acquired in the sagittal plane with an isotropic
resolution of 1x1x1 mm3 (matrix size: 256x256x176) pre-procedure, one day post
procedure and at 12 months. Accuracy of the HIFU ablation location was
estimated by measuring the distance from individual lesion center to the
atlas-based VIM target (marked at x=-14mm, y=-19mm, z=-2mm on a Montreal Neurological
Institute.1 Both T1 and
T2-weighted images from each time point were
registered to the MNI template using SPM8. Lesion volume was measured on
T2-weighted images by drawing a region of interest over the 3D-volume with
hyperintensity. Center of the lesion for each patient was then identified from
this volume and marked on the template space.
In addition, we performed minimum intensity projection (MIP) on the T1-weighted
images from V2 and V3 to measure the overall extent of the lesions.
Results
In the immediate post-operative period, thalamotomy lesions demonstrated restricted
diffusion and a variable region of surrounding vasogenic edema on T2W images.
The edema subsided completely and the T2W core lesions demonstrated regression
in the volume at 1 year follow-up (Figure 1). The average lesion size at V2 was
82 ± 41mm3 (range: 37mm3 - 174mm3) which
reduced dramatically to 13 ± 18mm3 (range: 0 - 48mm3). For two patients; there was a planned shift
in the target during the treatment resulting in extended lesions but their
overall lesion size was comparable to average lesion size.Lesions were close to
the VIM target, with an average distance of 2.82±1.04mm (1.78±1.03mm more
medial, and 1.28±1.54mm more superior) from the stereotactically defined VIM
target.
In addition, the lesions from each subjects were mapped onto
Oxford thalamic connectivity atlas
1 (FMRIB), where 7 sub-thalamic regions were
identified, each with anatomical connections to one of the 7 cortical zones,
based on the probabilistic diffusion tractography from multiple subjects. HIFU lesions were centered mostly at the
inferior end of the sub-region connected to the pre-motor cortex (PM). The
lesions were shown to be connected to the primary motor cortex (M1) and/or the
frontal pole region (FP) on probabilistic diffusion map (Figure 3).
Conclusion
High intensity MR-guided focused ultrasound can be delivered
through the intact skull to make precise ablations deep in the brain. MR guided focused ultrasound thalamotomy
improves hand tremor in ET and was well tolerated. Lesions generated by this
procedure are accurately placed, and match well with the known location of VIM
nucleus based on anatomical atlases. The lesions appear to regress in size over
a 12 month period but the therapeutic effect is maintained. Additional studies
will be need to determine the long term durability of the treatment.
Acknowledgements
Focused Ultrasound Foundation, BIRD Foundation, InsightecReferences
[1] Grabner et al., Symmetric atlasing and model based
segmentation: an application to the hippocampus in older adults”, Med Image
ComputComput Assist IntervIntConf Med Image ComputComput Assist Interv, vol. 9,
pp. 58–66, 2006.
[2] Behrens et al.,
Non-invasive mapping of connections between human thalamus and cortex using
diffusion imaging. Nature Neuroscience 6(7):750-757 (2003)