F012 A machine learning model for improving the AAO estimation for HD

Abstract

The Enroll-HD study is an integrated clinical research platform operating in the field of neurology, and is a large effort by the HD community to address the research and clinical needs of HD. It encompasses more than 25,000 participants, including controls, manifest HD and pre-manifest HD. The size of Enroll-HD makes it amenable to machine learning (ML) algorithms for discovering patterns and making predictions that pertain to several aspects of the disease, such as the prediction of the age of onset (AAO). The AAO is one of the most important variables in the Enroll-HD, and it describes the point in time that the first symptoms of HD occur. In this project, we developed a prediction model for improving the estimation of the AAO. We carefully selected the study population, and we included participants that have entered the Enroll-HD as pre-manifest and progressed to manifest. We performed feature selection to select the most critical variables for our prediction models. Our ML model achieved a performance of around 70% with an average error of 5 years. The model was composed by variables such as the CAG repeat, parent age of onset, marital status and years of smoking. Our results demonstrate that we are able to improve the estimation of the AAO for HD gene carriers. This is very important for selecting populations for clinical trials that are focusing on neuroprotective treatments. Timely treatments can result in a better quality of life not only for HD patients but also for their caregivers.

Authors 5

1. Kasper F. van der Zwaan Aachen Department of Neurology, Leiden University Medical Center (LUMC)

   Leiden University Medical Center

   Affiliation as printed

   Department of Neurology, Leiden University Medical Center, Leiden, The Netherlands

2. Stephanie Feleus Aachen Department of Epidemiology, Leiden University Medical Center Department of Neurology, Leiden University Medical Center (LUMC)

   Leiden University Medical Center

   Affiliation as printed

   Department of Epidemiology, Leiden University Medical Center, Leiden, The Netherlands

   Department of Neurology, Leiden University Medical Center, Leiden, The Netherlands

3. Susanne Tamara de Bot Aachen Department of Neurology, Leiden University Medical Center (LUMC)

   Leiden University Medical Center

   Affiliation as printed

   Department of Neurology, Leiden University Medical Center, Leiden, The Netherlands

4. Katherine Wolstencroft Aachen Leiden Institute of Advanced Computer Science

   Leiden University

   Affiliation as printed

   Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands

5. Eleni Mina Aachen Department of Human Genetics

   Leiden University Medical Center

   Affiliation as printed

   Department of Human Genetics, Leiden University Medical Center, Leiden, The Netherlands