Stanford University - Customer Case Study

Predicting near-term risk of stroke in patients with atrial fibrillation, a common heart rhythm disorder, remains a major clinical challenge and unmet need. Conventional frequentist and Bayesian approaches have failed to provide meaningful classification of near-term risk. Using Ayasdi's IRIS platform and topological data analysis, we are exploring methods to "phenotype" dynamic patterns of atrial fibrillation using raw data from implanted medical devices. Our goal is to identify patterns of heart rhythm disturbances that could be used to predict near-term risk of events in real-time.