Dataset Information

Robust Multiple Sclerosis Lesion Inpainting with Edge Prior.

ABSTRACT: Inpainting lesions is an important preprocessing task for algorithms analyzing brain MRIs of multiple sclerosis (MS) patients, such as tissue segmentation and cortical surface reconstruction. We propose a new deep learning approach for this task. Unlike existing inpainting approaches which ignore the lesion areas of the input image, we leverage the edge information around the lesions as a prior to help the inpainting process. Thus, the input of this network includes the T1-w image, lesion mask and the edge map computed from the T1-w image, and the output is the lesion-free image. The introduction of the edge prior is based on our observation that the edge detection results of the MRI scans will usually contain the contour of white matter (WM) and grey matter (GM), even though some undesired edges appear near the lesions. Instead of losing all the information around the neighborhood of lesions, our approach preserves the local tissue shape (brain/WM/GM) with the guidance of the input edges. The qualitative results show that our pipeline inpaints the lesion areas in a realistic and shape-consistent way. Our quantitative evaluation shows that our approach outperforms the existing state-of-the-art inpainting methods in both image-based metrics and in FreeSurfer segmentation accuracy. Furthermore, our approach demonstrates robustness to inaccurate lesion mask inputs. This is important for practical usability, because it allows for a generous over-segmentation of lesions instead of requiring precise boundaries, while still yielding accurate results.

SUBMITTER: Zhang H

PROVIDER: S-EPMC8692168 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Quantitative MRI (qMRI) probes the microstructural properties of the central nervous system (CNS) by providing biophysical measures of tissue characteristics. In this work, we aimed to (i) identify qMRI measures that distinguish histological lesion types in postmortem multiple sclerosis (MS) brains, especially the remyelinated ones; and to (ii) investigate the relationship between those measures and quantitative histological markers of myelin, axons, and astrocytes in the same experimental setting. Three fixed MS whole brains were imaged with qMRI at 3T to obtain magnetization transfer ratio (MTR), myelin water fraction (MWF), quantitative T1 (qT1), quantitative susceptibility mapping (QSM), fractional anisotropy (FA) and radial diffusivity (RD) maps. The identification of lesion types (active, inactive, chronic active, or remyelinated) and quantification of tissue components were performed using histological staining methods as well as immunohistochemistry and immunofluorescence. Pairwise logistic and LASSO regression models were used to identify the best qMRI discriminators of lesion types. The association between qMRI and quantitative histological measures was performed using Spearman's correlations and linear mixed-effect models. We identified a total of 65 lesions. MTR and MWF best predicted the chance of a lesion to be remyelinated, whereas RD and QSM were useful in the discrimination of active lesions. The measurement of microstructural properties through qMRI did not show any difference between chronic active and inactive lesions. MWF and RD were associated with myelin content in both lesions and normal-appearing white matter (NAWM), FA was the measure most associated with axon content in both locations, while MWF was associated with astrocyte immunoreactivity only in lesions. Moreover, we provided evidence of extensive astrogliosis in remyelinated lesions. Our study provides new information on the discriminative power of qMRI in differentiating MS lesions -especially remyelinated ones- as well as on the relative association between multiple qMRI measures and myelin, axon and astrocytes.

Dataset Information

Robust Multiple Sclerosis Lesion Inpainting with Edge Prior.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets