A Multicenter, Scan-Rescan, Human and Machine Learning CMR Study to Test Generalizability and Precision in Imaging Biomarker Analysis.

Journal: Circulation. Cardiovascular imaging
Published Date: Sep 24, 2019

Abstract

BACKGROUND: Automated analysis of cardiac structure and function using machine learning (ML) has great potential, but is currently hindered by poor generalizability. Comparison is traditionally against clinicians as a reference, ignoring inherent human inter- and intraobserver error, and ensuring that ML cannot demonstrate superiority. Measuring precision (scan:rescan reproducibility) addresses this. We compared precision of ML and humans using a multicenter, multi-disease, scan:rescan cardiovascular magnetic resonance data set.

Authors

  • Anish Bhuva
    Institute for Cardiovascular Science, University College London, United Kingdom
  • Wenjia Bai
    Department of Computing Imperial College London London UK.
  • Clement Lau
    Cardiac Imaging Department, Barts Heart Centre, St Bartholomew's Hospital, London, UK; William Harvey Research Institute, NIHR Barts Biomedical Research Centre, Queen Mary University of London, London, UK.
  • Rhodri Davies
    Institute for Cardiovascular Science, University College London, United Kingdom
  • Yang Ye
    Wuxi Hospital of Traditional Chinese Medicine, Wuxi, China.
  • Heeraj Bulluck
    Institute for Cardiovascular Science, University College London, United Kingdom
  • Elisa McAlindon
    Institute for Cardiovascular Science, University College London, United Kingdom
  • Veronica Culotta
    Institute for Cardiovascular Science, University College London, United Kingdom
  • Peter Swoboda
    Institute for Cardiovascular Science, University College London, United Kingdom
  • Gabriella Captur
    Institute of Cardiovascular Science, University College London, London, UK.
  • Thomas Treibel
    Department of Cardiovascular Imaging, Barts Heart Centre, Barts Health NHS Trust, London, United Kingdom
  • João Augusto
    Department of Cardiology, Hospital Professor Fernando Fonseca, IC19, Amadora, 2720-276, Lisbon, Portugal. joao.augusto@hff.min-saude.pt.
  • Kristopher Knott
    Department of Cardiovascular Imaging, Barts Heart Centre, Barts Health NHS Trust, London, United Kingdom
  • Andreas Seraphim
    Cardiac Imaging Department, Barts Heart Centre, St Bartholomew's Hospital, London, UK; Institute of Cardiovascular Science, University College London, London, UK.
  • Graham Cole
    Department of Cardiovascular Imaging, Barts Heart Centre, Barts Health NHS Trust, London, United Kingdom
  • Steffen Petersen
  • Nicola Edwards
    Data Science Institute and Department of Medicine
  • John Greenwood
    Bristol Heart Institute, Bristol NIHR Biomedical Research Centre, University Hospitals Bristol NHS Trust and University of Bristol, United Kingdom
  • Chiara Bucciarelli-Ducci
    Bristol Heart Institute, Bristol National Institute of Health Research Biomedical Research Centre, University Hospitals Bristol NHS Trust and University of Bristol, Bristol, UK.
  • Alun Hughes
    Multidisciplinary Cardiovascular Research Centre and Division of Biomedical Imaging, Leeds Institute of Cardiovascular and Metabolic Medicine, University of Leeds, United Kingdom
  • Daniel Rueckert
    Biomedical Image Analysis (BioMedIA) Group, Department of Computing, Imperial College London, UK. Electronic address: d.rueckert@imperial.ac.uk.
  • James Moon
    Imperial College London, National Heart and Lung Institute, Hammersmith Hospital, United Kingdom
  • Charlotte Manisty
    Cardiac Imaging Department, Barts Heart Centre, St Bartholomew's Hospital, London, UK; Institute of Cardiovascular Science, University College London, London, UK.

Keywords

External Resources

Popular Topics
Recent Journals