MULTIMODAL MRI TO PREDICT DBS MOTOR AND COGNITIVE OUTCOMES IN PARKINSON’S DISEASE

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Implantable deep brain stimulation (DBS) is a second-line surgical neuromodulation for Parkinson's disease (PD) that can provide significant relief of motor symptoms when medications become less effective, however there are currently no reliable predictors of therapeutic efficacy. While the gold standard suggests that a patient will benefit from DBS if their motor symptoms respond to PD medications with at least 30% improvement, the pre- dictive accuracy of this criteria is variable across studies, and has been disproportionately evaluated in the con- text of only one of two common brain targets for PD. A lack of reliable prognostic criteria to predict overall out- comes with DBS, including risk for cognitive side-effects in balance with motor symptom improvement, has led to variable patient outcomes. Some not considered candidates by the gold standard have been reported to re- spond well to DBS, while others have experienced limited benefit despite strong candidacy and well positioned electrodes. With over 4000 DBS surgeries performed in the US for PD each year, there is an increasing demand for better prognostic tools and streamlined approaches to inform optimal candidate and brain target selection. We aim to address this unmet need by leveraging advanced MRI techniques for improved prediction of patient outcomes after one year of DBS. Previous studies have shown that measures of brain connectivity derived from functional MRI (fMRI) and diffusion tensor imaging (DTI), can be used to predict motor symptom response to DBS. Brain iron accumulation in the basal ganglia, a marker of PD severity derived from susceptibility contrast on T2* MRI, has also shown promise for predicting DBS motor outcomes. However, practical implementation of the results from previous studies in the pre-operative setting is limited by the use of normative connectomes, post-operative electrode coordinates, and less sensitive susceptibility techniques for prediction, along with out- come data from only one of two brain targets for PD. To overcome these limitations, we will use patient-specific pre-operative MRI data to predict outcomes for both PD targets. Specifically, we propose a novel multivariate approach that incorporates fMRI and DTI with quantitative susceptibility mapping (QSM), a superior susceptibility technique to T2* MRI, to enhance prediction accuracy. By using complimentary features of disease burden that are highly relevant to DBS effects on brain connectivity and individual basal ganglia structures, we expect that our approach will improve upon the current gold standard. In 100 patients with PD undergoing DBS, we aim to: 1) evaluate the impact of 3T MRI on clinical prediction of motor outcomes, 2) identify MR and clinical features most relevant for predicting overall versus individual motor and cognitive outcomes, and 3) investigate additional variance in patient outcomes explained by post-operative targeting accuracy. The results will provide a framework in which DBS outcomes can be reliably predicted at the patient and symptom level to inform candidate and target selection, and even therapeutic settings. In this way, we can ensure that resources are geared toward patients most likely to benefit from DBS.