Dataset Information

Application of Machine Learning Using Decision Trees for Prognosis of Deep Brain Stimulation of Globus Pallidus Internus for Children With Dystonia.

ABSTRACT:

Background

While Deep Brain Stimulation (DBS) of the Globus pallidus internus is a well-established therapy for idiopathic/genetic dystonia, benefits for acquired dystonia are varied, ranging from modest improvement to deterioration. Predictive biomarkers to aid DBS prognosis for children are lacking, especially in acquired dystonias, such as dystonic Cerebral Palsy. We explored the potential role of machine learning techniques to identify parameters that could help predict DBS outcome.

Methods

We conducted a retrospective study of 244 children attending King's College Hospital between September 2007 and June 2018 for neurophysiological tests as part of their assessment for possible DBS at Evelina London Children's Hospital. For the 133 individuals who underwent DBS and had 1-year outcome data available, we assessed the potential predictive value of six patient parameters: sex, etiology (including cerebral palsy), baseline severity (Burke-Fahn-Marsden Dystonia Rating Scale-motor score), cranial MRI and two neurophysiological tests, Central Motor Conduction Time (CMCT) and Somatosensory Evoked Potential (SEP). We applied machine learning analysis to determine the best combination of these features to aid DBS prognosis. We developed a classification algorithm based on Decision Trees (DTs) with k-fold cross validation for independent testing. We analyzed all possible combinations of the six features and focused on acquired dystonias.

Results

Several trees resulted in better accuracy than the majority class classifier. However, the two features that consistently appeared in top 10 DTs were CMCT and baseline dystonia severity. A decision tree based on CMCT and baseline severity provided a range of sensitivity and specificity, depending on the threshold chosen for baseline dystonia severity. In situations where CMCT was not available, a DT using SEP alone provided better than the majority class classifier accuracy.

Conclusion

The results suggest that neurophysiological parameters can help predict DBS outcomes, and DTs provide a data-driven, highly interpretable decision support tool that lends itself to being used in clinical practice to help predict potential benefit of DBS in dystonic children. Our results encourage the introduction of neurophysiological parameters in assessment pathways, and data collection to facilitate multi-center evaluation and validation of these potential predictive markers and of the illustrative decision support tools presented here.

SUBMITTER: Shah SA

PROVIDER: S-EPMC7115974 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:IntroductionGlobus pallidus internus (GPi) deep brain stimulation (DBS) is widely used in patients with dystonia. However, 10-20% of patients receive insufficient benefits. The objectives of this study are to evaluate the effectiveness of bilateral subthalamic nucleus (STN) DBS along with unilateral posteroventral pallidotomy (PVP) in patients with dystonia who experienced unsatisfactory GPi-DBS and to address the reported rescue procedures after suboptimal DBS or lesion surgery in dystonia patients.MethodsSix patients with isolated dystonia who had previously undergone bilateral GPi-DBS with suboptimal improvement were included. Standardized assessments of dystonia using the Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS) and quality of life using SF-36 were evaluated before surgery and 1, 6 months, and last follow-up (LFU) after surgery. STN bilateral OFF (bi-OFF), unilateral ON (uni-ON), and bilateral ON (bi-ON) states were recorded at LFU. Specific items were used to find publications published before 10 April 2022 regarding rescue procedures after suboptimal DBS or lesion surgery in patients with dystonia for reference. Eleven original studies including case reports/series were identified for discussion.ResultsSubstantial clinical benefits were achieved in all six patients. Significant amelioration was achieved during the 1-month (6.5 ± 7.45; p = 0.0049), 6-month (5.67 ± 6.3; p = 0.0056) follow-ups, and at LFU (4.67 ± 4.72; p = 0.0094) when compared with the baseline (LFU of GPi DBS with on status) (17.33 ± 11.79) assessed by BFMDRS. The percentage of improvement reached 70.6, 74.67, and 77.05%, respectively. At LFU, significant differences were found between the stimulation bi-OFF and uni-ON (11.08 ± 8.38 vs. 9 ± 8.52, p = 0.0191), and between the stimulation bi-OFF and bi-ON (11.08 ± 8.38 vs. 4.67 ± 4.72, p = 0.0164). Trends depicting a better improvement in stimulation bi-ON compared with uni-ON (4.67 ± 4.72 vs. 9 ± 8.52, p = 0.0538) were observed.ConclusionOur results suggest that bilateral STN-DBS plus unilateral PVP may be an effective rescue procedure for patients with isolated dystonia who experienced suboptimal movement improvement following GPi-DBS. However, given the heterogeneity of patients and the small sample size, these findings should be interpreted with caution.

Project description:IntroductionDystonic opisthotonus is defined as a backward arching of the neck and trunk, which ranges in severity from mild backward jerks to life-threatening prolonged severe muscular spasms. It can be associated with generalized dystonic syndromes or, rarely, present as a form of axial truncal dystonia. The etiologies vary from idiopathic, genetic, tardive, hereditary-degenerative, or associated with parkinsonism. We report clinical cases of dystonic opisthotonus associated with adult-onset dystonic syndromes, that benefitted from globus pallidus internus (GPi) deep brain stimulation (DBS).MethodsClinical data from patients with dystonic syndromes who underwent comprehensive medical review, multidisciplinary assessment, and tailored medical and neurosurgical managements were prospectively analyzed. Quantification of dystonia severity pre- and postoperatively was performed using the Burke-Fahn-Marsden Dystonia Rating Scale and quantification of overall pain severity was performed using the Visual Analog Scale.ResultsThree male patients, with age of onset of the dystonic symptoms ranging from 32 to 51 years old, were included. Tardive dystonia, adult-onset dystonia-parkinsonism and adult-onset idiopathic axial dystonia were the etiologies identified. Clinical investigation and management were tailored according to the complexity of the individual presentations. Although they shared common clinical features of adult-onset dystonia, disabling dystonic opisthotonus, refractory to medical management, was the main indication for GPi-DBS in all patients presented. The severity of axial dystonia ranged from disturbance of daily function to life-threatening truncal distortion. All three patients underwent bilateral GPi DBS at a mean age of 52 years (range 48-55 years), after mean duration of symptoms prior to DBS of 10.7 years (range 4-16 years). All patients showed a rapid and sustained clinical improvement of their symptoms, notably of the dystonic opisthotonos, at postoperative follow-up ranging from 20 to 175 months. In some, the ability to resume activities of daily living and reintegration into the society was remarkable.ConclusionAdult-onset dystonic syndromes predominantly presenting with dystonic opisthotonus are relatively rare. The specific nature of dystonic opisthotonus remains a treatment challenge, and thorough investigation of this highly disabling condition with varying etiologies is often necessary. Although patients may be refractory to medical management and botulinum toxin injection, Globus pallidus stimulation timed and tailored provided symptomatic control in this cohort and may be considered in other carefully selected cases.

Project description:Globus pallidus internus deep brain stimulation is an established therapy for patients with medication-refractory Parkinson's disease. Clinical outcomes are highly dependent on applying stimulation to precise locations in the brain. However, robust neurophysiological markers are needed to determine the optimal electrode location and to guide postoperative stimulation parameter selection. In this study, we evaluated evoked resonant neural activity in the pallidum as a potential intraoperative marker to optimize targeting and stimulation parameter selection to improve outcomes of deep brain stimulation for Parkinson's disease. Intraoperative local field potential recordings were acquired in 22 patients with Parkinson's disease undergoing globus pallidus internus deep brain stimulation implantation (N = 27 hemispheres). A control group of patients undergoing implantation in the subthalamic nucleus (N = 4 hemispheres) for Parkinson's disease or the thalamus for essential tremor (N = 9 patients) were included for comparison. High-frequency (135 Hz) stimulation was delivered from each electrode contact sequentially while recording the evoked response from the other contacts. Low-frequency stimulation (10 Hz) was also applied as a comparison. Evoked resonant neural activity features, including amplitude, frequency and localization were measured and analysed for correlation with empirically derived postoperative therapeutic stimulation parameters. Pallidal evoked resonant neural activity elicited by stimulation in the globus pallidus internus or externus was detected in 26 of 27 hemispheres and varied across hemispheres and across stimulating contacts within individual hemispheres. Bursts of high-frequency stimulation elicited evoked resonant neural activity with similar amplitudes (P = 0.9) but a higher frequency (P = 0.009) and a higher number of peaks (P = 0.004) than low-frequency stimulation. We identified a 'hotspot' in the postero-dorsal pallidum where stimulation elicited higher evoked resonant neural activity amplitudes (P < 0.001). In 69.6% of hemispheres, the contact that elicited the maximum amplitude intraoperatively matched the contact empirically selected for chronic therapeutic stimulation by an expert clinician after 4 months of programming sessions. Pallidal and subthalamic nucleus evoked resonant neural activity were similar except for lower pallidal amplitudes. No evoked resonant neural activity was detected in the essential tremor control group. Given its spatial topography and correlation with postoperative stimulation parameters empirically selected by expert clinicians, pallidal evoked resonant neural activity shows promise as a potential marker to guide intraoperative targeting and to assist the clinician with postoperative stimulation programming. Importantly, evoked resonant neural activity may also have the potential to guide directional and closed-loop deep brain stimulation programming for Parkinson's disease.

Project description:Background and purposeDeep brain stimulation is a well-established treatment for generalized dystonia, but outcomes remain variable. Establishment of an imaging marker to guide device targeting and programming could possibly impact the efficacy of deep brain stimulation in dystonia, particularly in the absence of acute clinical markers to indicate benefit. We hypothesize that the stimulation-based functional and structural connectivity using resting-state fMRI and DTI can predict therapeutic outcomes in patients with generalized dystonia and deep brain stimulation.Materials and methodsWe performed a retrospective analysis of 39 patients with inherited or idiopathic-isolated generalized dystonia who underwent bilateral globus pallidus internus deep brain stimulation. After electrode localization, the volumes of tissue activated were modeled and used as seed regions for functional and structural connectivity measures using a normative data base. Resulting connectivity maps were correlated with postoperative improvement in the Unified Dystonia Rating Scale score.ResultsStructural connectivity between the volumes of tissue activated and the primary sensorimotor cortex was correlated with Unified Dystonia Rating Scale improvement, while more anterior prefrontal connectivity was inversely correlated with Unified Dystonia Rating Scale improvement. Functional connectivity between the volumes of tissue activated and primary sensorimotor regions, motor thalamus, and cerebellum was most correlated with Unified Dystonia Rating Scale improvement; however, an inverse correlation with Unified Dystonia Rating Scale improvement was seen in the supplemental motor area and premotor cortex.ConclusionsFunctional and structural connectivity with multiple nodes of the motor network is associated with motor improvement in patients with generalized dystonia undergoing deep brain stimulation. Results from this study may serve as a basis for future development of clinical markers to guide deep brain stimulation targeting and programming in dystonia.

Project description:IntroductionThe globus pallidus internus (GPi) region has evolved as a potential target for deep brain stimulation (DBS) in Parkinson's disease (PD). DBS of the GPi (GPi DBS) is an established, safe and effective method for addressing many of the motor symptoms associated with advanced PD. It is important that clinicians fully understand this target when considering GPi DBS for individual patients.MethodsThe literature on GPi DBS in PD has been comprehensively reviewed, including the anatomy, physiology and potential pitfalls that may be encountered during surgical targeting and post-operative management. Here, we review and address the implications of lead location on GPi DBS outcomes. Additionally, we provide a summary of randomized controlled clinical trials conducted on DBS in PD, together with expert commentary on potential applications of the GPi as target. Finally, we highlight future technologies that will likely impact GPi DBS, including closed-loop adaptive approaches (e.g. sensing-stimulating capabilities), advanced methods for image-based targeting and advances in DBS programming, including directional leads and pulse shaping.ResultsThere are important disease characteristics and factors to consider prior to selecting the GPi as the DBS target of PD surgery. Prior to and during implantation of the leads it is critical to consider the neuroanatomy, which can be defined through the combination of image-based targeting and intraoperative microelectrode recording strategies. There is an increasing body of literature on GPi DBS in patients with PD suggesting both short- and long-term benefits. Understanding the GPi target can be useful in choosing between the subthalamic (STN), GPi and ventralis intermedius nucleus as lead locations to address the motor symptoms and complications of PD.ConclusionGPi DBS can be effectively used in select cases of PD. As the ongoing DBS target debate continues (GPi vs. STN as DBS target), clinicians should keep in mind that GPi DBS has been shown to be an effective treatment strategy for a variety of symptoms, including bradykinesia, rigidity and tremor control. GPi DBS also has an important, direct anti-dyskinetic effect. GPi DBS is easier to program in the outpatient setting and will allow for more flexibility in medication adjustments (e.g. levodopa). Emerging technologies, including GPi closed-loop systems, advanced tractography-based targeting and enhanced programming strategies, will likely be future areas of GPi DBS expansion. We conclude that although the GPi as DBS target may not be appropriate for all PD patients, it has specific clinical advantages.