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

Magnetic resonance imaging (MRI) is a principal radiological modality thatprovides radiation-free, abundant, and diverse information about the wholehuman body for medical diagnosis, but suffers from prolonged scan time. Thescan time can be significantly reduced through k-space undersampling but theintroduced artifacts need to be removed in image reconstruction. Although deeplearning (DL) has emerged as a powerful tool for image reconstruction in fastMRI, its potential in multiple imaging scenarios remains largely untapped. Thisis because not only collecting large-scale and diverse realistic training datais generally costly and privacy-restricted, but also existing DL methods arehard to handle the practically inevitable mismatch between training and targetdata. Here, we present a Physics-Informed Synthetic data learning framework forFast MRI, called PISF, which is the first to enable generalizable DL formulti-scenario MRI reconstruction using solely one trained model. For a 2Dimage, the reconstruction is separated into many 1D basic problems and startswith the 1D data synthesis, to facilitate generalization. We demonstrate thattraining DL models on synthetic data, integrated with enhanced learningtechniques, can achieve comparable or even better in vivo MRI reconstructioncompared to models trained on a matched realistic dataset, reducing the demandfor real-world MRI data by up to 96%. Moreover, our PISF shows impressivegeneralizability in multi-vendor multi-center imaging. Its excellentadaptability to patients has been verified through 10 experienced doctors'evaluations. PISF provides a feasible and cost-effective way to markedly boostthe widespread usage of DL in various fast MRI applications, while freeing fromthe intractable ethical and practical considerations of in vivo human dataacquisitions.

One for Multiple: Physics-informed Synthetic Data Boosts Generalizable Deep Learning for Fast MRI Reconstruction