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.
Highlights
Biomedical repositories such as the UK Biobank provide increasing access to prospectively collected cardiac imaging, these data are unlabeled, which creates barriers to their use in supervised machine learning
These findings demonstrate how weakly supervised methods help mitigate the lack of expert-labeled training data in cardiac imaging settings, and how real-world health outcomes can be learned directly from large-scale, unlabeled medical imaging data
We evaluate the impact of training set size on weak supervision performance
Summary
Biomedical repositories such as the UK Biobank provide increasing access to prospectively collected cardiac imaging, these data are unlabeled, which creates barriers to their use in supervised machine learning. The highdimensionality and overall complexity of these images make them appealing candidates for use with deep learning[8] These prospectively collected MRIs are unlabeled, and the low prevalence of malformations such as aortic valve disease introduces considerable challenges in building labeled datasets at the scale required to train deep learning models. Instead of requiring hand-labeled examples from cardiologists, we use new methods[9,10] to encode domain knowledge in the form of multiple, noisy heuristics or labeling functions which are applied to unlabeled data to generate imperfect training labels. In patients identified by our classifier as having BAV, we find a 1.8-fold increase in risk of a major adverse cardiac event These findings demonstrate how weakly supervised methods help mitigate the lack of expert-labeled training data in cardiac imaging settings, and how real-world health outcomes can be learned directly from large-scale, unlabeled medical imaging data
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
