Weakly supervised classification of aortic valve malformations using unlabeled cardiac MRI sequences.

Journal: Nature communications
PMID: 31308376

Abstract

Biomedical repositories such as the UK Biobank provide increasing access to prospectively collected cardiac imaging, however these data are unlabeled, which creates barriers to their use in supervised machine learning. We develop a weakly supervised deep learning model for classification of aortic valve malformations using up to 4,000 unlabeled cardiac MRI sequences. Instead of requiring highly curated training data, weak supervision relies on noisy heuristics defined by domain experts to programmatically generate large-scale, imperfect training labels. For aortic valve classification, models trained with imperfect labels substantially outperform a supervised model trained on hand-labeled MRIs. In an orthogonal validation experiment using health outcomes data, our model identifies individuals with a 1.8-fold increase in risk of a major adverse cardiac event. This work formalizes a deep learning baseline for aortic valve classification and outlines a general strategy for using weak supervision to train machine learning models using unlabeled medical images at scale.

Authors

  • Jason A Fries
    Department of Computer Science, Stanford University, Stanford, CA, 94305, USA. jason-fries@stanford.edu.
  • Paroma Varma
    Stanford University.
  • Vincent S Chen
    Department of Computer Science, Stanford University, Stanford, CA, 94305, USA.
  • Ke Xiao
    Department of Pediatrics, Stanford University, Stanford, CA, 94304, USA.
  • Heliodoro Tejeda
    Department of Pediatrics, Stanford University, Stanford, CA, 94304, USA.
  • Priyanka Saha
    Department of Pediatrics, Stanford University, Stanford, CA, 94304, USA.
  • Jared Dunnmon
    Stanford University, Stanford, CA.
  • Henry Chubb
    Department of Pediatrics, Stanford University, Stanford, CA, 94304, USA.
  • Shiraz Maskatia
    Department of Pediatrics, Stanford University, Stanford, CA, 94304, USA.
  • Madalina Fiterau
  • Scott Delp
    Stanford University Department of Bioengineering, Stanford, CA, United States of America.
  • Euan Ashley
    Division of Cardiovascular Medicine, Stanford University, Palo Alto, California; Departments of Medicine, Genetics, and Biomedical Data Science, Stanford University, Palo Alto, California; Center for Inherited Cardiovascular Disease, Stanford University, Palo Alto, California.
  • Christopher Ré
    1Stanford University, Stanford, CA USA.
  • James R Priest
    Department of Pediatrics, Stanford University, Stanford, CA, 94304, USA.

Keywords

External Resources

Popular Topics
Recent Journals