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

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 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.May 2, 2019 2/26 apply the CNN-LSTM to a cohort of 9,230 new patients with long-term outcome and MRI data from the UK May 2, 2019 3/26Biobank. 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.May 2, 2019 4/26

Automated detection of left ventricle in arterial input function images for inline perfusion mapping using deep learning: A study of 15,000 patients

“…fewer-shot learning 46 ) and effective methods to improve data labeling process (e.g. weak supervision 47 ). Their applications to clinical imaging AI are of high interest.…”

Challenges of developing a digital scribe to reduce clinical documentation burden

“…Crowd-sourcing may be used to annotate large quantities of data from other domains, but it is less suitable for healthcare scenarios because of the need for domain knowledge to guarantee data quality. 14,42 In general, digital scribe research is [43][44][45] A dataset of medical conversations along with the corresponding summaries would allow far-reaching advances in the digital scribe and clinical documentation space. Weak supervision has the potential to maximize the use of unlabeled medical data which is costly to annotate.…”

“…Weak supervision has the potential to maximize the use of unlabeled medical data which is costly to annotate. 42 Data trusts have also been proposed as a way of sharing medical data for research while giving users power over how their data is used. 46 It remains to be seen how the implementation of data trusts affects the advancement of medical and AI research.…”

Breaking Medical Data Sharing Boundaries by Employing Artificial Radiographs

“…The application of Artificial Intelligence (AI) in medicine promises to personalize diagnosis, decision management and therapy based on the combination of patient information with knowledge of thousands of experts and the outcome of billions of patient [1][2][3][4]. In recent years, a lot of scientific effort has focused on applications of AI in medicine with a particularly strong focus on radiology [5][6][7][8][9][10]. Whenever there has been progress towards this vision of an omniscient radiological AI, it has mostly been anticipated by corresponding technical advances in the field of Computer Vision on natural images.…”

“…However, human performance in computer vision on medical images was only achieved but not surpassed. Whenever human performance in computer vision on medical images has been reached, large datasets have been used -oftentimes pooled from many sites, containing thousands of images [5,17,18]. Going a step further and surpassing human performance in computer vision on natural images, however, required even larger databases containing up to billions of natural images [19].…”