Over 1.6 million Americans suffer from significant tricuspid valve leakage. In most cases this leakage is designated as secondary. Thus, valve dysfunction is assumed to be due to valve-extrinsic factors. We challenge this paradigm and hypothesize that the tricuspid valve maladapts in those patients rendering the valve at least partially culpable for its dysfunction. As a first step in testing this hypothesis, we set out to demonstrate that the tricuspid valve maladapts in disease. To this end, we induced biventricular heart failure in sheep that developed tricuspid valve leakage. In the anterior leaflets of those animals, we investigated maladaptation on multiple scales. We demonstrated alterations on the protein and cell-level, leading to tissue growth, thickening, and stiffening. These data provide a new perspective on a poorly understood, yet highly prevalent disease. Our findings may motivate novel therapy options for many currently untreated patients with leaky tricuspid valves.
Cardiac MRI allows for a comprehensive assessment of myocardial structure, function and tissue characteristics. Here we describe a foundational vision system for cardiac MRI, capable of representing the breadth of human cardiovascular disease and health. Our deep-learning model is trained via self-supervised contrastive learning, in which visual concepts in cine-sequence cardiac MRI scans are learned from the raw text of the accompanying radiology reports. We train and evaluate our model on data from four large academic clinical institutions in the United States. We additionally showcase the performance of our models on the UK BioBank and two additional publicly available external datasets. We explore emergent capabilities of our system and demonstrate remarkable performance across a range of tasks, including the problem of left-ventricular ejection fraction regression and the diagnosis of 39 different conditions such as cardiac amyloidosis and hypertrophic cardiomyopathy. We show that our deep-learning system is capable of not only contextualizing the staggering complexity of human cardiovascular disease but can be directed towards clinical problems of interest, yielding impressive, clinical-grade diagnostic accuracy with a fraction of the training data typically required for such tasks.
2 February 2024 in Journal of the Mechanical Behavior of Biomedical Materials