Supervised deep learning (SDL) methodology holds promise for accelerated magnetic resonance imaging (AMRI) but is hampered by the reliance on extensive training data. Some self-supervised frameworks, such as deep image prior (DIP), have emerged, eliminating the explicit training procedure but often struggling to remove noise and artifacts under significant degradation. This work introduces a novel self-supervised accelerated parallel MRI approach called PEARL, leveraging a multiple-stream joint deep decoder with two cross-fusion schemes to accurately reconstruct one or more target images from compressively sampled k-space. Each stream comprises cascaded cross-fusion sub-block networks (SBNs) that sequentially perform combined upsampling, 2D convolution, joint attention, ReLU activation and batch normalization (BN). Among them, combined upsampling and joint attention facilitate mutual learning between multiple-stream networks by integrating multi-parameter priors in both additive and multiplicative manners. Long-range unified skip connections within SBNs ensure effective information propagation between distant cross-fusion layers. Additionally, incorporating dual-normalized edge-orientation similarity regularization into the training loss enhances detail reconstruction and prevents overfitting. Experimental results consistently demonstrate that PEARL outperforms the existing state-of-the-art (SOTA) self-supervised AMRI technologies in various MRI cases. Notably, 5-fold  ∼ 6-fold accelerated acquisition yields a 1 %  ∼  2 % improvement in SSIM and a 3 %  ∼  6 % improvement in PSNR , along with a significant 15 %  ∼  20 % reduction in RLNE .

Download full-text PDF

Source
http://dx.doi.org/10.1109/JBHI.2023.3347355DOI Listing

Publication Analysis

Top Keywords

parallel mri
8
combined upsampling
8
joint attention
8
 ∼  improvement
8
pearl cascaded
4
self-supervised
4
cascaded self-supervised
4
cross-fusion
4
self-supervised cross-fusion
4
cross-fusion learning
4

Similar Publications

Background And Purpose: Asymptomatic carotid stenosis (ACS) is an independent risk factor for ischemic stroke and vascular cognitive impairment, affecting cognitive function across multiple domains. This study aimed to explore differences in static and dynamic intrinsic functional connectivity and temporal dynamics between patients with ACS and those without carotid stenosis.

Methods: We recruited 30 patients with unilateral moderate-to-severe (stenosis ≥ 50%) ACS and 30 demographically-matched healthy controls.

View Article and Find Full Text PDF

Background: Whole-body magnetic resonance imaging (wbMRI) allows general assessment of systemic cancers including lymphomas without radiation burden.

Aim: To evaluate the diagnostic performance of wbMRI in the staging of diffuse large B-cell lymphoma (DLBCL), determine the value of individual MRI sequences, and assess patients' concerns with wbMRI.

Methods: In this single-center prospective study, adult patients newly diagnosed with systemic DLBCL underwent wbMRI on a 3T scanner [diffusion weighted images with background suppression (DWIBS), T2, short tau inversion recovery (STIR), contrast-enhanced T1] and fluorodeoxyglucose (F-FDG) positron emission tomography/computed tomography (PET/CT) (reference standard).

View Article and Find Full Text PDF

Background: Tumour hypoxia resulting from inadequate perfusion is common in many solid tumours, including prostate cancer, and constitutes a major limiting factor in radiation therapy that contributes to treatment resistance. Emerging research in preclinical animal models indicates that exercise has the potential to enhance the efficacy of cancer treatment by modulating tumour perfusion and reducing hypoxia; however, evidence from randomised controlled trials is currently lacking. The 'Exercise medicine as adjunct therapy during RADIation for CAncer of the prostaTE' (ERADICATE) study is designed to investigate the impact of exercise on treatment response, tumour physiology, and adverse effects of treatment in prostate cancer patients undergoing external beam radiation therapy (EBRT).

View Article and Find Full Text PDF

Deep Learning Superresolution for Simultaneous Multislice Parallel Imaging-Accelerated Knee MRI Using Arthroscopy Validation.

Radiology

January 2025

From the Department of Radiology, Division of Musculoskeletal Radiology, NYU Grossman School of Medicine, 660 1st Ave, 3rd Fl, Rm 313, New York, NY 10016 (S.S.W., J.V., R.K., E.H.P., J.F.); Department for Diagnostic and Interventional Radiology, Eberhard Karls University Tübingen, University Hospital Tübingen, Tübingen, Germany (S.S.W.); Department of Radiology, University Hospital Basel, Basel, Switzerland (J.V.); Department of Radiology, Hospital do Coraçao, São Paulo, Brazil (T.C.R.); Academic Surgical Unit, South West London Elective Orthopaedic Centre (SWLEOC), London, United Kingdom (D.D.); Department of Radiology, Balgrist University Hospital, Zurich, Switzerland (B.F.); Department of Radiology, Jeonbuk National University Hospital, Jeonju, Republic of Korea (E.H.P.); Research Institute of Clinical Medicine of Jeonbuk National University, Biomedical Research Institute of Jeonbuk National University Hospital, Jeonju, Republic of Korea (E.H.P.); Medscanlagos Radiology, Cabo Frio, Brazil (A.S.); Centre for Data Analytics, Bond University, Gold Coast, Australia (S.E.S.); Siemens Healthineers AG, Erlangen, Germany (I.B.); and Siemens Medical Solutions USA, Malvern, Pa (G.K.).

Background Deep learning (DL) methods can improve accelerated MRI but require validation against an independent reference standard to ensure robustness and accuracy. Purpose To validate the diagnostic performance of twofold-simultaneous-multislice (SMSx2) twofold-parallel-imaging (PIx2)-accelerated DL superresolution MRI in the knee against conventional SMSx2-PIx2-accelerated MRI using arthroscopy as the reference standard. Materials and Methods Adults with painful knee conditions were prospectively enrolled from December 2021 to October 2022.

View Article and Find Full Text PDF

Purpose: Knee osteoarthritis (KOA) is a prevalent degenerative bone and joint disease observed in clinical practice. While acupuncture has demonstrated efficacy in treating KOA, the central mechanisms underlying its effects remain ambiguous. Recently, functional magnetic resonance imaging (fMRI) has been extensively applied in studying the brain mechanisms of acupuncture analgesia.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!