Segmenting white matter hyperintensities on isotropic three-dimensional Fluid Attenuated Inversion Recovery magnetic resonance images: Assessing deep learning tools on norwegian imaging database

TL;DR

This study evaluates deep learning models for automated segmentation of white matter hyperintensities in 3D FLAIR MRI images, demonstrating that 3D models can achieve high accuracy without T1-weighted images.

Contribution

It introduces and validates 3D deep learning models, especially 3D nnU-Net, for WMH segmentation on 3D FLAIR images, outperforming previous models and state-of-the-art methods.

Findings

3D nnU-Net achieved a dice score of 0.76.

Models trained on 3D data perform comparably or better than existing methods.

Segmentation accuracy is maintained without T1-weighted images.

Abstract

Automated segmentation of white matter hyperintensities (WMHs) is an essential step in neuroimaging analysis of Magnetic Resonance Imaging (MRI). Fluid Attenuated Inversion Recovery (FLAIR-weighted) is an MRI contrast that is particularly useful to visualize and quantify WMHs, a hallmark of cerebral small vessel disease and Alzheimer's disease (AD). Clinical MRI protocols migrate to a three-dimensional (3D) FLAIR-weighted acquisition to enable high spatial resolution in all three voxel dimensions. The current study details the deployment of deep learning tools to enable automated WMH segmentation and characterization from 3D FLAIR-weighted images acquired as part of a national AD imaging initiative. Among 441 participants (194 male, mean age: (64.91 +/- 9.32) years) from the DDI study, two in-house networks were trained and validated across five national collection sites. Three models were tested on a held-out subset of the internal data from the 441 participants and an external dataset with 29 cases from an international collaborator. These test sets were evaluated independently. Five established WMH performance metrics were used for comparison against ground truth human-in-the-loop segmentation. Results of the three networks tested, the 3D nnU-Net had the best performance with an average dice similarity coefficient score of 0.76 +/- 0.16, performing better than both the in-house developed 2.5D model and the SOTA Deep Bayesian network. With the increasing use of 3D FLAIR-weighted images in MRI protocols, our results suggest that WMH segmentation models can be trained on 3D data and yield WMH segmentation performance that is comparable to or better than state-of-the-art without the need for including T1-weighted image series.