Using a multi-decoder water-fat separation neural network for liver PDFF estimation

The authors of this study proposed a multi-task U-Net-based architecture to jointly estimate water-only and fat-only images. This approach allowed for the improvement in the estimation of water-fat images, enabling a reduction of the necessary echoes to achieve an accurate proton density fat fraction (PDFF) quantification. The proposed method was shown to be a reliable liver fat quantification tool for clinical use. Key points Novel water-fat separation neural network allows for liver PDFF estimation by using multi-echo MR images with a reduced number of echoes. Prospective single-center validation demonstrated that echo reduction leads to a significant shortening of the scan time, compared to standard 6-echo acquisition. Qualitative and quantitative performance of the proposed method showed no significant differences in PDFF estimation with respect to the reference technique. Article: Liver PDFF estimation using a multi-decoder water-fat separation neural network with a reduced number of echoes Authors: Juan Pablo Meneses, Cristobal Arrieta, Gabriel della Maggiora, Cecilia Besa, Jesús Urbina, Marco Arrese, Juan Cristóbal Gana, Jose E. Galgani, Cristian Tejos & Sergio Uribe