Abstract #2839

The utility of a convolutional neural network for generating a myelin volume index map from rapid simultaneous relaxometry imaging

Yasuhiko Tachibana^1,2, Akifumi Hagiwara^2,3, Masaaki Hori², Jeff Kershaw¹, Misaki Nakazawa², Tokuhiko Omatsu¹, Riwa Kishimoto¹, Shigeki Aoki², Tatsuya Higashi⁴, and Takayuki Obata^1,4

¹Applied MRI Research, Department of Molecular imaging and Theranostics, National Institute of Radiological Sciences, QST, Chiba, Japan, ²Department of Radiology, Juntendo University School of Medicine, Tokyo, Japan, ³Department of Radiology, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan, ⁴Department of Molecular imaging and Theranostics, National Institute of Radiological Sciences, QST, Chiba, Japan

Myelin volume measurement based on rapid-simultaneous-relaxometry-imaging (RSRI) is useful for the clinic. However, the software that translates RSRI into the index map contains a potential weak point: the index is obtained using a pixelwise table-lookup that doesn't incorporate information from surrounding pixels. A novel deep-learning-based method was developed to overcome this problem. A myelin volume index based on magnetization-transfer saturation imaging was used as a reference for training and testing. The proposed method was evaluated by leave-one-out cross-validation using images from 20 healthy volunteers. Correlation with the reference was significantly higher for the proposed method.

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

Join Here