Stratification of amyotrophic lateral sclerosis patients: a crowdsourcing approach

Robert Kueffner,Venkat Balagurusamy,Hsih-Te Yang,Donna Dillenberger,Nazem Atassi,Raquel Norel,Javier Garcı́a-Garcı́a ,Jian Peng,Thea Norman,Guang Li,Merit Cudkowicz,Maya Bronfeld,Adriano Chiò ,Wen-Chieh Fang,Jinfeng Xiao,Orla Hardiman,Huan-Jui Chang,Liuxia Wang,Barbara Di Camillo,Bruce Hoff,Joshua Knight,Neta Zach,Melanie Leitner ,Lara M Mangravite ,Gustavo Stolovitzky

doi:10.1038/s41598-018-36873-4

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease where substantial heterogeneity in clinical presentation urgently requires a better stratification of patients for the development of drug trials and clinical care. In this study we explored stratification through a crowdsourcing approach, the DREAM Prize4Life ALS Stratification Challenge. Using data from >10,000 patients from ALS clinical trials and 1479 patients from community-based patient registers, more than 30 teams developed new approaches for machine learning and clustering, outperforming the best current predictions of disease outcome. We propose a new method to integrate and analyze patient clusters across methods, showing a clear pattern of consistent and clinically relevant sub-groups of patients that also enabled the reliable classification of new patients. Our analyses reveal novel insights in ALS and describe for the first time the potential of a crowdsourcing to uncover hidden patient sub-populations, and to accelerate disease understanding and therapeutic development.

Highlights

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder which causes the death of motor neurons that control voluntary muscles, leading to progressive muscle weakening and paralysis and death within an average of only 3–5 years from symptom onset[1]
This Challenge invited participants to develop computational algorithms that could predict ALS disease progression, with the best performing algorithms achieving a prediction accuracy that would allow a 20% reduction in the number of patients needed for a trial[16] and are currently being tested in real-world clinical trial applications[17,18]
The challenge was based on two datasets: (1) ALS clinical trials data collected through the Pooled Resource Open-Access ALS Clinical Trials (PRO-ACT) database, and (2) community-based ALS clinical data collected through ALS registries

Summary

Introduction

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder which causes the death of motor neurons that control voluntary muscles, leading to progressive muscle weakening and paralysis and death within an average of only 3–5 years from symptom onset[1]. One of the biggest challenges today is the well-established heterogeneity of ALS1,4, with patients displaying widely different patterns of disease manifestation and progression, and genetic analyses suggesting heterogeneity of the underlying biological mechanisms[5,6,7,8] This heterogeneity has detrimental effects on clinical trial planning and interpretation[3], on attempts to understand disease mechanisms, and on clinical care, as it increases uncertainty about prognosis and optimal treatment. The PRO-ACT database was previously used for a crowdsourcing computational challenge: The 2012 DREAM-Phil Bowen ALS Prediction Prize4Life challenge (The ALS Prediction Challenge)[16] This Challenge invited participants to develop computational algorithms that could predict ALS disease progression, with the best performing algorithms achieving a prediction accuracy that would allow a 20% reduction in the number of patients needed for a trial[16] and are currently being tested in real-world clinical trial applications[17,18]

Objectives

Methods

Results

Conclusion