Machine learning methods for optimal prediction of motor outcome in Parkinson's disease

Mohammad R Salmanpour; Mojtaba Shamsaei; Abdollah Saberi; Ivan S Klyuzhin; Jing Tang; Vesna Sossi; Arman Rahmim

doi:10.1016/j.ejmp.2019.12.022

Machine learning methods for optimal prediction of motor outcome in Parkinson's disease

Phys Med. 2020 Jan:69:233-240. doi: 10.1016/j.ejmp.2019.12.022. Epub 2020 Jan 7.

Authors

Mohammad R Salmanpour¹, Mojtaba Shamsaei¹, Abdollah Saberi², Ivan S Klyuzhin³, Jing Tang⁴, Vesna Sossi⁵, Arman Rahmim⁶

Affiliations

¹ Department of Energy Engineering and Physics, Amirkabir University of Technology, Tehran, Iran.
² Department of Computer Engineering, Islamic Azad University, Tehran, Iran.
³ Department of Medicine, University of British Columbia, Vancouver, BC, Canada.
⁴ Department of Electrical & Computer Engineering, Oakland University, Rochester, MI, USA.
⁵ Department of Physics & Astronomy, University of British Columbia, Vancouver, BC, Canada.
⁶ Department of Physics & Astronomy, University of British Columbia, Vancouver, BC, Canada; Department of Radiology, University of British Columbia, Vancouver, BC, Canada; Department of Radiology, Johns Hopkins University, Baltimore, MD, USA. Electronic address: arman.rahmim@ubc.ca.

PMID: 31918375
DOI: 10.1016/j.ejmp.2019.12.022

Abstract

Purpose: It is vital to appropriately power clinical trials towards discovery of novel disease-modifying therapies for Parkinson's disease (PD). Thus, it is critical to improve prediction of outcome in PD patients.

Methods: We systematically probed a range of robust predictor algorithms, aiming to find best combinations of features for significantly improved prediction of motor outcome (MDS-UPDRS-III) in PD. We analyzed 204 PD patients with 18 features (clinical measures; dopamine-transporter (DAT) SPECT imaging measures), performing different randomized arrangements and utilizing data from 64%/6%/30% of patients in each arrangement for training/training validation/final testing. We pursued 3 approaches: i) 10 predictor algorithms (accompanied with automated machine learning hyperparameter tuning) were first applied on 32 experimentally created combinations of 18 features, ii) we utilized Feature Subset Selector Algorithms (FSSAs) for more systematic initial feature selection, and iii) considered all possible combinations between 18 features (262,143 states) to assess contributions of individual features.

Results: A specific set (set 18) applied to the LOLIMOT (Local Linear Model Trees) predictor machine resulted in the lowest absolute error 4.32 ± 0.19, when we firstly experimentally created 32 combinations of 18 features. Subsequently, 2 FSSAs (Genetic Algorithm (GA) and Ant Colony Optimization (ACO)) selecting 5 features, combined with LOLIMOT, reached an error of 4.15 ± 0.46. Our final analysis indicated that longitudinal motor measures (MDS-UPDRS-III years 0 and 1) were highly significant predictors of motor outcome.

Conclusions: We demonstrate excellent prediction of motor outcome in PD patients by employing automated hyperparameter tuning and optimal utilization of FSSAs and predictor algorithms.

Keywords: Motor symptom (MDS-UPDRS-III); Outcome prediction; Parkinson’s disease; Predictor and feature subset selection algorithms.

MeSH terms

Adult
Aged
Aged, 80 and over
Algorithms
Computer Simulation
Dopamine Plasma Membrane Transport Proteins / chemistry
Female
Humans
Machine Learning*
Male
Middle Aged
Parkinson Disease / diagnostic imaging*
Parkinson Disease / physiopathology*
Pattern Recognition, Automated
Reproducibility of Results
Tomography, Emission-Computed, Single-Photon
Treatment Outcome

Substances

Dopamine Plasma Membrane Transport Proteins