Meeting Banner
Abstract #1589

Towards Automated Scan Volume Placement for Breast MRI Using a Deep Neural Network

Timothy J Allen1, Leah C Henze Bancroft2, Ersin Bayram3, Lloyd Estkowski4, Pingni Wang5, Ty A Cashen6, and James H Holmes7
1Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, 2Radiology, University of Wisconsin-Madison, Madison, WI, United States, 3GE Healthcare, Houston, TX, United States, 4GE Healthcare, Waukesha, WI, United States, 5GE Healthcare, Menlo Park, CA, United States, 6GE Healthcare, Madison, WI, United States, 7Radiology, University of Iowa, Iowa City, IA, United States


We investigate the use of a deep neural network to automatically position imaging volumes for breast MR exams. The axial localizer images and scan volume information from a variety of MR scanners were used to train a deep neural network to replicate the clinical technologists’ placement. The average intersection over union between clinical placement and neural network predicted placement was 0.46 ± 0.21. The distance between volume centers was 7.4 cm on average and as low as 1.1 cm. These results show promise for improving consistency of imaging volume placement in breast MRI.

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

Join Here