Abstract #1269

A neural network approach for estimating muscle perfusion from DCE-MRI data

Christopher C Conlin¹, Xiaowan Li¹, Stephen Decker², Christopher J Hanrahan¹, Gwenael Layec², Nan Hu³, Vivian S Lee⁴, and Jeff L Zhang¹

¹Department of Radiology and Imaging Sciences, University of Utah, Salt Lake City, UT, United States, ²School of Public Health and Health Sciences, University of Massachusetts Amherst, Amherst, MA, United States, ³Division of Biostatistics, Department of Internal Medicine, University of Utah, Salt Lake City, UT, United States, ⁴Verily Life Sciences, Cambridge, MA, United States

Perfusion is an important aspect of calf muscle function that can be measured with dynamic contrast-enhanced (DCE) MRI. However, conventional methods for quantifying perfusion from DCE-MRI data require an appropriate tracer-kinetic model, which may not be available clinically. In this study, we examined the feasibility of neural networks (NNs) for quantifying calf-muscle perfusion from DCE-MRI data. We found that NNs estimate perfusion with accuracy comparable to conventional methods, without the need for a tracer-kinetic model. NNs like those developed in this study can be readily incorporated into ordinary MRI scanner software, facilitating routine quantitative perfusion analysis with DCE-MRI.

This abstract and the presentation materials are available to members only; a login is required.

Join Here