English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Automatic identification of brain lesions from magnetic resonance imaging(MRI) scans of stroke survivors would be a useful aid in patient diagnosis andtreatment planning. We propose a multi-modal multi-path convolutional neuralnetwork system for automating stroke lesion segmentation. Our system has nineend-to-end UNets that take as input 2-dimensional (2D) slices and examines allthree planes with three different normalizations. Outputs from these nine totalpaths are concatenated into a 3D volume that is then passed to a 3Dconvolutional neural network to output a final lesion mask. We trained andtested our method on datasets from three sources: Medical College of Wisconsin(MCW), Kessler Foundation (KF), and the publicly available Anatomical Tracingsof Lesions After Stroke (ATLAS) dataset. Cross-study validation results (withindependent training and validation datasets) were obtained to compare withprevious methods based on naive Bayes, random forests, and three recentlypublished convolutional neural networks. Model performance was quantified interms of the Dice coefficient. Training on the KF and MCW images and testing onthe ATLAS images yielded a mean Dice coefficient of 0.54. This was reliablybetter than the next best previous model, UNet, at 0.47. Reversing the trainand test datasets yields a mean Dice of 0.47 on KF and MCW images, whereas thenext best UNet reaches 0.45. With all three datasets combined, the currentsystem compared to previous methods also attained a reliably highercross-validation accuracy. It also achieved high Dice values for many smallerlesions that existing methods have difficulty identifying. Overall, our systemis a clear improvement over previous methods for automating stroke lesionsegmentation, bringing us an important step closer to the inter-rater accuracylevel of human experts.

A multi-path 2.5 dimensional convolutional neural network system for segmenting stroke lesions in brain MRI images