AUTHOR=Chan Karissa , Maralani Pejman Jabehdar , Moody Alan R. , Khademi April 

TITLE=Synthesis of diffusion-weighted MRI scalar maps from FLAIR volumes using generative adversarial networks

JOURNAL=Frontiers in Neuroinformatics

VOLUME=Volume 17 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2023.1197330

DOI=10.3389/fninf.2023.1197330

ISSN=1662-5196

ABSTRACT=Acquisition and pre-processing pipelines for diffusion-weighted imaging (DWI) volumes are resource- and time-consuming. Generating synthetic DWI scalar maps from commonly acquired brain MRI sequences such as Fluid-attenuated inversion recovery (FLAIR) could be useful for supplementing datasets. In this work we design and compare GAN-based image translation models for generating DWI scalar maps from FLAIR MRI for the first time. We employ a pix2pix model, two modified CycleGANs using paired and unpaired data, and a convolutional autoencoder to synthesize DWI fractional anisotropy (FA) and mean diffusivity (MD) from whole FLAIR volumes. In total, 420 FLAIR and DWI volumes (11,957 images) from multi-center dementia and vascular disease cohorts were used for training/testing. Pix2pix demonstrated the best performance both quantitatively and qualitatively, with evaluation metrics comparable to existing literature, followed by the paired CycleGAN. A new evaluation metric based on histogram similarity shows pix2pix’s FA and MD models had significantly better structural similarity of tissue structures and fine details, including WM tracts and CSF spaces, between real and generated images. Detailed regional analysis of synthetic volumes also demonstrated that synthetic DWI images can not only be used to supplement clinical datasets, but demonstrates potential utility in bypassing or correcting registration in data pre-processing.