Abstract #1149

Fast Longitudinal Image REgistration (FLIRE) for Breast MRI

Michelle W Tong¹, Maren M. Sjaastad Andreassen², Ana Rodríguez-Soto¹, Christopher C Conlin¹, Tyler Seibert³, Michael Hahn¹, Rebecca Rakow-Penner¹, and Anders M Dale¹

¹Radiology, University of California, San Diego, La Jolla, CA, United States, ²Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, ³Radiation Oncology, University of California, San Diego, La Jolla, CA, United States

MRI is commonly used to evaluate breast cancer response to neoadjuvant chemotherapy. Image registration algorithms aid diagnostic accuracy by correcting changes in patient position across visits. A novel technique, Fast Longitudinal Image REgistration (FLIRE), was evaluated by comparison to a standard algorithm optimized for breast tissue, DRAMMS. Breast images acquired over the course of neoadjuvant treatment were registered to baseline images. FLIRE runtime was 5 minutes (~20× faster). With this fast processing time, FLIRE performed within 2.5-10.2% of optimized DRAMMS in similarity measures. Overall accuracy and computation speed for FLIRE with default arguments demonstrates promise for clinical time-sensitive applications.

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

Join Here