Session Time: 1:45pm-3:15pm
Location: Exhibit Hall C
Objective: To evaluate the prognostic accuracy of structural magnetic resonance imaging and genetic-based measures in premanifest Huntington’s disease for estimating future motor impairment.
Background: Despite extensive efforts to find the underlying relationship between brain structure atrophy and clinical decline in Huntington’s disease (HD), we still lack tools to provide a reliable estimation of a patient’s prognosis using imaging data. Availability of such methods in the premanifest HD (pre-HD) subpopulation, would allow discrimination of individuals at high risk of pronounced motor impairment.
Methods: 80 pre-HD individuals were evaluated at baseline and a follow-up visit (3 years between them). Data from different domains at baseline such as imaging (gray matter concentration), genetics (CAG repeats, CAG age product), age and the Unified Huntington’s disease rating scale (UHDRS) total motor score (TMS) were used to estimate the risk of pronounced motor impairment at the follow-up visit. The subjects were assigned a high or low risk based on their TMS at the follow-up visit. A linear support vector machine was used to learn the association between baseline data and future motor impairment. The performance of the learned classifiers was estimated using leave-one-out cross-validation. Both TMS and voxel-level gray matter concentration were corrected for ‘normal aging’ based on a linear fit estimated on 85 roughly age- and sex-matched healthy controls.
Results: We obtained an 83% accuracy rate to predict the risk level of future motor impairment based on genetics, demographics and the initial motor assessment. By adding brain imaging features, the classification accuracy was further increased to 93%, achieving a specificity of 100%. Within the imaging data domain, strong effects are detected in a region that spans the right caudate nucleus and the adjacent Nucleus Accumbens.
Conclusions: The analysis of gray matter integrity along with genetic-based measures is capable of detecting the degree of future motor deterioration of each individual. This implies that it is possible to perform a stratification of pre-HD based on distinct levels of future impairment with these two data modalities. We argue that the proposed analysis has the potential to detect subpopulations at higher risk of severe impairment for future clinical trials.
To cite this abstract in AMA style:E. Castro, P. Polosecki, I. Rish, G. Cecchi. Baseline multimodal information predicts future motor impairment in premanifest Huntington’s disease [abstract]. Mov Disord. 2017; 32 (suppl 2). https://www.mdsabstracts.org/abstract/baseline-multimodal-information-predicts-future-motor-impairment-in-premanifest-huntingtons-disease/. Accessed December 9, 2023.
« Back to 2017 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/baseline-multimodal-information-predicts-future-motor-impairment-in-premanifest-huntingtons-disease/