Event Detail

Event Type: 
Geometry-Topology Seminar
Date/Time: 
Monday, April 12, 2021 - 12:00 to 12:50
Location: 
Zoom. Please contact Christine Escher for a link.

Speaker Info

Institution: 
EECS
Abstract: 

In this talk, I will present a novel application of topological data analysis (TDA): persistent homology of Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) data. Topological features of publically available DCE-MRI data are plotted in a persistence diagram. The bottleneck metric is used to measure each patient's persistence homology distance and construct a distance matrix that records the pairwise distance between each patient. Next, a clustering algorithm is applied to the distance matrix allowing us to assign each patient one of many distinct topological classes. The assigned class is used to predict the treatment outcome. We find that for the patient cohort used to develop these methods, the topological class is a statistically significant predictor of treatment outcomes. In this talk, I will introduce DCE-MRI, provide a brief introduction to TDA, and describe how TDA can be used on DCE-MRI data. I will describe how the persistence homology of DCE-MRI data is constructed, and will briefly describe our methods and results.