Dataset Information

Automatic spatial estimation of white matter hyperintensities evolution in brain MRI using disease evolution predictor deep neural networks.

ABSTRACT: Previous studies have indicated that white matter hyperintensities (WMH), the main radiological feature of small vessel disease, may evolve (i.e., shrink, grow) or stay stable over a period of time. Predicting these changes are challenging because it involves some unknown clinical risk factors that leads to a non-deterministic prediction task. In this study, we propose a deep learning model to predict the evolution of WMH from baseline to follow-up (i.e., 1-year later), namely "Disease Evolution Predictor" (DEP) model, which can be adjusted to become a non-deterministic model. The DEP model receives a baseline image as input and produces a map called "Disease Evolution Map" (DEM), which represents the evolution of WMH from baseline to follow-up. Two DEP models are proposed, namely DEP-UResNet and DEP-GAN, which are representatives of the supervised (i.e., need expert-generated manual labels to generate the output) and unsupervised (i.e., do not require manual labels produced by experts) deep learning algorithms respectively. To simulate the non-deterministic and unknown parameters involved in WMH evolution, we modulate a Gaussian noise array to the DEP model as auxiliary input. This forces the DEP model to imitate a wider spectrum of alternatives in the prediction results. The alternatives of using other types of auxiliary input instead, such as baseline WMH and stroke lesion loads are also proposed and tested. Based on our experiments, the fully supervised machine learning scheme DEP-UResNet regularly performed better than the DEP-GAN which works in principle without using any expert-generated label (i.e., unsupervised). However, a semi-supervised DEP-GAN model, which uses probability maps produced by a supervised segmentation method in the learning process, yielded similar performances to the DEP-UResNet and performed best in the clinical evaluation. Furthermore, an ablation study showed that an auxiliary input, especially the Gaussian noise, improved the performance of DEP models compared to DEP models that lacked the auxiliary input regardless of the model's architecture. To the best of our knowledge, this is the first extensive study on modelling WMH evolution using deep learning algorithms, which deals with the non-deterministic nature of WMH evolution.

SUBMITTER: Rachmadi MF

PROVIDER: S-EPMC7294240 | biostudies-literature | 2020 Jul

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automatic spatial estimation of white matter hyperintensities evolution in brain MRI using disease evolution predictor deep neural networks.

Rachmadi Muhammad Febrian MF Valdés-Hernández Maria Del C MDC Makin Stephen S Wardlaw Joanna J Komura Taku T

Medical image analysis 20200426

Previous studies have indicated that white matter hyperintensities (WMH), the main radiological feature of small vessel disease, may evolve (i.e., shrink, grow) or stay stable over a period of time. Predicting these changes are challenging because it involves some unknown clinical risk factors that leads to a non-deterministic prediction task. In this study, we propose a deep learning model to predict the evolution of WMH from baseline to follow-up (i.e., 1-year later), namely "Disease Evolution ...[more]

PMID: 32428823

Similar Datasets

Project description:BackgroundWhite matter hyperintensities are an important marker of cerebral small vessel disease. This disease burden is commonly described as hyperintense areas in the cerebral white matter, as seen on T2-weighted fluid attenuated inversion recovery magnetic resonance imaging data. Studies have demonstrated associations with various cognitive impairments, neurological diseases, and neuropathologies, as well as clinical and risk factors, such as age, sex, and hypertension. Due to their heterogeneous appearance in location and size, studies have started to investigate spatial distributions and patterns, beyond summarizing this cerebrovascular disease burden in a single metric-its volume. Here, we review the evidence of association of white matter hyperintensity spatial patterns with its risk factors and clinical diagnoses.Design/methodsWe performed a systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Statement. We used the standards for reporting vascular changes on neuroimaging criteria to construct a search string for literature search on PubMed. Studies written in English from the earliest records available until January 31st, 2023, were eligible for inclusion if they reported on spatial patterns of white matter hyperintensities of presumed vascular origin.ResultsA total of 380 studies were identified by the initial literature search, of which 41 studies satisfied the inclusion criteria. These studies included cohorts based on mild cognitive impairment (15/41), Alzheimer's disease (14/41), Dementia (5/41), Parkinson's disease (3/41), and subjective cognitive decline (2/41). Additionally, 6 of 41 studies investigated cognitively normal, older cohorts, two of which were population-based, or other clinical findings such as acute ischemic stroke or reduced cardiac output. Cohorts ranged from 32 to 882 patients/participants [median cohort size 191.5 and 51.6% female (range: 17.9-81.3%)]. The studies included in this review have identified spatial heterogeneity of WMHs with various impairments, diseases, and pathologies as well as with sex and (cerebro)vascular risk factors.ConclusionThe results show that studying white matter hyperintensities on a more granular level might give a deeper understanding of the underlying neuropathology and their effects. This motivates further studies examining the spatial patterns of white matter hyperintensities.

Project description:Background: White matter hyperintensities of presumed vascular origin (WMH) are a common finding in elderly people and a growing social malady in the aging western societies. As a manifestation of cerebral small vessel disease, WMH are considered to be a vascular contributor to various sequelae such as cognitive decline, dementia, depression, stroke as well as gait and balance problems. While pathophysiology and therapeutical options remain unclear, large-scale studies have improved the understanding of WMH, particularly by quantitative assessment of WMH. In this review, we aimed to provide an overview of the characteristics, research subjects and segmentation techniques of these studies. Methods: We performed a systematic review according to the PRISMA statement. One thousand one hundred and ninety-six potentially relevant articles were identified via PubMed search. Six further articles classified as relevant were added manually. After applying a catalog of exclusion criteria, remaining articles were read full-text and the following information was extracted into a standardized form: year of publication, sample size, mean age of subjects in the study, the cohort included, and segmentation details like the definition of WMH, the segmentation method, reference to methods papers as well as validation measurements. Results: Our search resulted in the inclusion and full-text review of 137 articles. One hundred and thirty-four of them belonged to 37 prospective cohort studies. Median sample size was 1,030 with no increase over the covered years. Eighty studies investigated in the association of WMH and risk factors. Most of them focussed on arterial hypertension, diabetes mellitus type II and Apo E genotype and inflammatory markers. Sixty-three studies analyzed the association of WMH and secondary conditions like cognitive decline, mood disorder and brain atrophy. Studies applied various methods based on manual (3), semi-automated (57), and automated segmentation techniques (75). Only 18% of the articles referred to an explicit definition of WMH. Discussion: The review yielded a large number of studies engaged in WMH research. A remarkable variety of segmentation techniques was applied, and only a minority referred to a clear definition of WMH. Most addressed topics were risk factors and secondary clinical conditions. In conclusion, WMH research is a vivid field with a need for further standardization regarding definitions and used methods.

Dataset Information

Automatic spatial estimation of white matter hyperintensities evolution in brain MRI using disease evolution predictor deep neural networks.

Publications

Automatic spatial estimation of white matter hyperintensities evolution in brain MRI using disease evolution predictor deep neural networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets