Visualisation of the zona incerta for deep brain stimulation at 3.0 Tesla. Clin Neuroradiol. 2012;22:55-68. [PMID: 22349435 DOI: 10.1007/s00062-012-0136-3]

An Updated Overview of the Magnetic Resonance Imaging of Brain Iron in Movement Disorders

Brain iron load is one of the most important neuropathological hallmarks in movement disorders. Specifically, the iron provides most of the paramagnetic metal signals in the brain and its accumulation seems to play a key role, although not completely explained, in the degeneration of the basal ganglia, as well as other brain structures. Moreover, iron distribution patterns have been implicated in depicting different movement disorders. This work reviewed current literature on Magnetic Resonance Imaging for Brain Iron Detection and Quantification (MRI-BIDQ) in neurodegenerative processes underlying movement disorders.

Coronal Gradient Echo MRI to Visualize the Zona Incerta for Deep Brain Stimulation Targeting in Parkinson's Disease

INTRODUCTION

Deep brain stimulation of the zona incerta is effective at treating tremor and other forms of parkinsonism. However, the structure is not well visualized with standard MRI protocols making direct surgical targeting unfeasible and contributing to inconsistent clinical outcomes. In this study, we applied coronal gradient echo MRI to directly visualize the rostral zona incerta in Parkinson's disease patients to improve targeting for deep brain stimulation.

METHODS

We conducted a prospective study to optimize and evaluate an MRI sequence to visualize the rostral zona incerta in patients with Parkinson's disease (n = 31) and other movement disorders (n = 13). We performed a contrast-to-noise ratio analysis of specific regions of interest to quantitatively assess visual discrimination of relevant deep brain structures in the optimized MRI sequence. Regions of interest were independently assessed by 2 neuroradiologists, and interrater reliability was assessed.

RESULTS

Rostral zona incerta and subthalamic nucleus were well delineated in our 5.5-min MRI sequence, indicated by excellent interrater agreement between neuroradiologists for region-of-interest measurements (>0.90 intraclass coefficient). Mean contrast-to-noise ratio was high for both rostral zona incerta (6.39 ± 3.37) and subthalamic nucleus (17.27 ± 5.61) relative to adjacent white matter. There was no significant difference between mean signal intensities or contrast-to-noise ratio for Parkinson's and non-Parkinson's patients for either structure.

DISCUSSION/CONCLUSION

Our optimized coronal gradient echo MRI sequence delineates subcortical structures relevant to traditional and novel deep brain stimulation targets, including the zona incerta, with high contrast-to-noise. Future studies will prospectively apply this sequence to surgical planning and postimplantation outcomes.

Asleep Deep Brain Stimulation of the Nucleus Ventralis Intermedius for Essential Tremor Using Indirect Targeting and Interventional Magnetic Resonance Imaging: Single-Institution Case Series

BACKGROUND

Literature on asleep deep brain stimulation (DBS) of the ventralis intermedius (Vim) nucleus in essential tremor is relatively sparse. Furthermore, controversy exists as to whether indirect ("consensus" or "atlas-based") targeting of the Vim requires physiologic adjustment for effective clinical outcomes in DBS surgery.

OBJECTIVES

The objective of this study was to evaluate the clinical results of asleep Vim DBS using indirect coordinates and real-time interventional magnetic resonance imaging guidance.

METHODS

Retrospective review of a prospectively collected database was performed to identify patients with essential tremor undergoing asleep Vim DBS using interventional magnetic resonance imaging guidance. Stereotactic and clinical outcomes were abstracted and analyzed using descriptive statistics.

RESULTS

A total of 12 consecutive patients were identified, all of whom were available for 6-month clinical follow-up. Stereotactic (radial) error was 0.5 ± 0.2 mm on the left and 0.5 ± 0.3 mm on the right. Fahn-Tolosa-Marin tremor scores in the treated limb(s) decreased by 71.2% ± 31.0% (P = 0.0088), The Essential Tremor Rating Assessment Scale activities of daily living improved by 74.9% ± 23.7% (P < 0.0001), and The Essential Tremor Rating Assessment Scale performance improved by 64.3 ± 16.2% (P = 0.0004). Surgical complications were mild and generally transient. Stimulation-related side effects were similar to those reported in historical series of awake Vim DBS.

CONCLUSIONS

Asleep Vim DBS using indirect targeting and interventional magnetic resonance imaging-guided placement is safe and effective, with 6-month clinical results similar to those achieved with awake placement. These data support the use of asleep surgery in essential tremor and represent a baseline for comparison with future studies using more advanced targeting techniques.

Deep Learning-Based Deep Brain Stimulation Targeting and Clinical Applications

Background

The purpose of the present study was to evaluate deep learning-based image-guided surgical planning for deep brain stimulation (DBS). We developed deep learning semantic segmentation-based DBS targeting and prospectively applied the method clinically.

Methods

T2^∗ fast gradient-echo images from 102 patients were used for training and validation. Manually drawn ground truth information was prepared for the subthalamic and red nuclei with an axial cut ∼4 mm below the anterior–posterior commissure line. A fully convolutional neural network (FCN-VGG-16) was used to ensure margin identification by semantic segmentation. Image contrast augmentation was performed nine times. Up to 102 original images and 918 augmented images were used for training and validation. The accuracy of semantic segmentation was measured in terms of mean accuracy and mean intersection over the union. Targets were calculated based on their relative distance from these segmented anatomical structures considering the Bejjani target.

Results

Mean accuracies and mean intersection over the union values were high: 0.904 and 0.813, respectively, for the 62 training images, and 0.911 and 0.821, respectively, for the 558 augmented training images when 360 augmented validation images were used. The Dice coefficient converted from the intersection over the union was 0.902 when 720 training and 198 validation images were used. Semantic segmentation was adaptive to high anatomical variations in size, shape, and asymmetry. For clinical application, two patients were assessed: one with essential tremor and another with bradykinesia and gait disturbance due to Parkinson’s disease. Both improved without complications after surgery, and microelectrode recordings showed subthalamic nuclei signals in the latter patient.

Conclusion

The accuracy of deep learning-based semantic segmentation may surpass that of previous methods. DBS targeting and its clinical application were made possible using accurate deep learning-based semantic segmentation, which is adaptive to anatomical variations.

Interleaving Stimulation in Parkinson's Disease, Tremor, and Dystonia

BACKGROUND/AIMS

Interleaving stimulation (ILS) in deep brain stimulation (DBS) provides individualized stimulation of 2 contacts delivered in alternating order. Currently, limited information on the utility of ILS exists. The aims of this study were to determine the practical applications and outcomes of ILS DBS in Parkinson's disease (PD), tremor, and dystonia.

METHODS

We performed a single-center, unblinded, retrospective chart review of all patients with DBS attempted on ILS at our referral center assessing for rationale and outcomes.

RESULTS

Fifty patients (PD, n = 27; tremor, n = 7; dystonia, n = 16 patients) tried ILS for 2 rationales: management of adverse effects (n = 29) and to improve clinical efficacy (n = 21). A total of 19 patients demonstrated improvement with ILS for adverse effect management predominately for the treatment of dyskinesias (n = 12). In the vast majority of dyskinetic patients, a contact added into the rostral zona incerta with ILS was performed. Nine out of 21 patients demonstrated improved clinical efficacy with ILS with all 6 PD patients who tried ILS for this rationale demonstrating benefit.

CONCLUSIONS

In PD, ILS provided benefits for dyskinesias and parkinsonism, with minimal improvement of other adverse effects. In tremor and dystonia, marginal effects in terms of mitigation of adverse effects and improvement of clinical outcomes were evident.

Deep Brain Stimulation of the H Fields of Forel Alleviates Tics in Tourette Syndrome

The current rationale for target selection in Tourette syndrome revolves around the notion of cortico-basal ganglia circuit involvement in the pathophysiology of the disease. However, despite extensive research, the ideal target for deep brain stimulation (DBS) is still under debate, with many structures being neglected and underexplored. Based on clinical observations and taking into account the prevailing hypotheses of network processing in Tourette syndrome, we chose the fields of Forel, namely field H1, as a target for DBS. The fields of Forel constitute the main link between the striatopallidal system and the thalamocortical network, relaying pallidothalamic projections from core anatomical structures to the thalamic ventral nuclear group. In a retrospective study we investigated two patients suffering from chronic, medically intractable Tourette syndrome who underwent bilateral lead implantation in field H1 of Forel. Clinical scales revealed significant alleviation of tics and comorbid symptoms, namely depression and anxiety, in the postoperative course in both patients.

Efficient imaging of midbrain nuclei using inverse double-echo steady-state acquisition

PURPOSE

Imaging of midbrain nuclei using T2- or T2*-weighted MRI often entails long echo time, leading to long scan time. In this study, an inverse double-echo steady-state (iDESS) technique is proposed for efficiently depicting midbrain nuclei.

METHODS

Thirteen healthy subjects participated in this study. iDESS was performed along with two sets of T2*-weighted spoiled gradient-echo images (SPGR1, with scan time identical to iDESS and SPGR2, using clinical scanning parameters as a reference standard) for comparison. Generation of iDESS composite images combining two echo signals was optimized for maximal contrast-to-noise ratio (CNR) between the red nuclei and surrounding tissues. Signal-to-noise ratios (SNRs) were calculated from the occipital lobe. Comparison was also made using phase-enhanced images as in standard susceptibility-weighted imaging (SWI).

RESULTS

The iDESS images present significantly higher SNR efficiency (171.3) than SPGR1 (158.7, p = 0.013) and SPGR2 (95.5, p < 10(-8)). iDESS CNR efficiency (19.2) is also significantly greater than SPGR1 (6.9, p < 10(-6)) and SPGR2 (14.3, p = 0.0016). Compared with DESS, iDESS provides further advantage on enhanced phase information and hence improved contrast on SWI-processed images.

CONCLUSIONS

iDESS efficiently depicts midbrain nuclei with improved CNR efficiency, increased SNR efficiency, and reduced scan time and is less prone to susceptibility signal loss from air-tissue interfaces.

Anatomic characterization of prelemniscal radiations by probabilistic tractography: implications in Parkinson's disease

To characterize the anatomical connectivity of the prelemniscal radiations (Raprl), a white matter region within the posterior subthalamic area (PSA) that is an effective neurosurgical target for treating motor symptoms of Parkinson's disease (PD). Diffusion-weighted images were acquired from twelve healthy subjects using a 3T scanner. Constrained spherical deconvolution, a method that allows the distinction of crossing fibers within a voxel, was used to compute track-density images with sufficient resolution to accurately delineate distinct PSA regions and probabilistic tractography of Raprl in both hemispheres. Raprl connectivity was reproducible across all subjects and showed fibers traversing through this region towards primary and supplementary motor cortices, the orbitofrontal cortex, ventrolateral thalamus, and the globus pallidus, cerebellum and dorsal brainstem. All brain regions reached by Raprl fibers are part of motor circuits involved in the pathophysiology of PD; while these fiber systems converge at the level of the PSA, they can be spatially segregated. Fibers of distinct and specific motor control networks are identified within Raprl. The description of this anatomical crossroad suggests that, in the future, tractography could allow deep brain stimulation or lesional therapies in white matter targets according to individual patient's symptoms.

Ultra-high field magnetic resonance imaging of the basal ganglia and related structures

Deep brain stimulation is a treatment for Parkinson's disease and other related disorders, involving the surgical placement of electrodes in the deeply situated basal ganglia or thalamic structures. Good clinical outcome requires accurate targeting. However, due to limited visibility of the target structures on routine clinical MR images, direct targeting of structures can be challenging. Non-clinical MR scanners with ultra-high magnetic field (7T or higher) have the potential to improve the quality of these images. This technology report provides an overview of the current possibilities of visualizing deep brain stimulation targets and their related structures with the aid of ultra-high field MRI. Reviewed studies showed improved resolution, contrast- and signal-to-noise ratios at ultra-high field. Sequences sensitive to magnetic susceptibility such as T2^* and susceptibility weighted imaging and their maps in general showed the best visualization of target structures, including a separation between the subthalamic nucleus and the substantia nigra, the lamina pallidi medialis and lamina pallidi incompleta within the globus pallidus and substructures of the thalamus, including the ventral intermediate nucleus (Vim). This shows that the visibility, identification, and even subdivision of the small deep brain stimulation targets benefit from increased field strength. Although ultra-high field MR imaging is associated with increased risk of geometrical distortions, it has been shown that these distortions can be avoided or corrected to the extent where the effects are limited. The availability of ultra-high field MR scanners for humans seems to provide opportunities for a more accurate targeting for deep brain stimulation in patients with Parkinson's disease and related disorders.

Imaging for deep brain stimulation: The zona incerta at 7 Tesla

AIM: To evaluate different promising magnetic resonance imaging (MRI) methods at 7.0 Tesla (T) for the pre-stereotactic visualization of the zona incerta (ZI).

METHODS: Two neuroradiologists qualitatively and quantitatively examined T2-turbo spin-echo (T2-TSE), T1-weighted gradient-echo, as well as FLASH2D-T2Star and susceptibility-weighted imaging (SWI) for the visualization of the ZI at 7.0 T MRI. Delineation and image quality for the ZI were independently examined using a 6-scale grading system. Inter-rater reliability using Cohen’s kappa coefficient (κ) were assessed. Contrast-to-noise ratios (CNR), and signal-to-noise ratios (SNR) for the ZI were calculated for all sequences. Differences in delineation, SNR, and CNR between the sequences were statistically assessed using a paired t-test. For the anatomic validation the coronal FLASH2D-T2Star images were co-registered with a stereotactic atlas (Schaltenbrand-Wahren).

RESULTS: The rostral part of the ZI (rZI) could easily be identified and was best and reliably visualized in the coronal FLASH2D-T2Star images. The caudal part was not definable in any of the sequences. No major artifacts in the rZI were observed in any of the scans. FLASH2D-T2Star and SWI imaging offered significant higher CNR values for the rZI compared to T2-TSE images (P > 0.05). The co-registration of the coronal FLASH2D-T2Star images with the stereotactic atlas schema (Schaltenbrand-Wahren) confirmed the correct localization of the ZI in all cases.

CONCLUSION: FLASH2D-T2Star imaging (particularly coronal view) provides the reliable and currently optimal visualization of the rZI at 7.0 T. These results can facilitate a better and more precise targeting of the caudal part of the ZI than ever before.

The subthalamic nucleus at 3.0 Tesla: choice of optimal sequence and orientation for deep brain stimulation using a standard installation protocol: clinical article

OBJECT

Reliable visualization of the subthalamic nucleus (STN) is indispensable for accurate placement of electrodes in deep brain stimulation (DBS) surgery for patients with Parkinson disease (PD). The aim of the study was to evaluate different promising new MRI methods at 3.0 T for preoperative visualization of the STN using a standard installation protocol.

METHODS

Magnetic resonance imaging studies (T2-FLAIR, T1-MPRAGE, T2*-FLASH2D, T2-SPACE, and susceptibility-weighted imaging sequences) obtained in 9 healthy volunteers and in 1 patient with PD were acquired. Two neuroradiologists independently analyzed image quality and visualization of the STN using a 6-point scale. Interrater reliability, contrast-to-noise ratios, and signal-to-noise ratios for the STN were calculated. For illustration of the anatomical accuracy, coronal T2*-FLASH2D images were fused with the corresponding coronal section schema of the Schaltenbrand and Wahren stereotactic atlas.

RESULTS

The STN was best and reliably visualized on T2*-FLASH2D imaging (in particular, the coronal view). No major artifacts in the STN were observed in any of the sequences. Susceptibility-weighted, T2-SPACE, and T2*-FLASH2D imaging provided significantly higher contrast-to-noise ratio values for the STN than standard T2-weighted imaging. Fusion of the coronal T2*-FLASH2D and the digitized coronal atlas view projected the STN clearly within the boundaries of the STN found in anatomical sections.

CONCLUSIONS

For 3.0-T MRI, T2*-FLASH2D (particularly the coronal view) provides optimal delineation of the STN using a standard installation protocol.