Joachim Schmitt, Dominik Weidlich, Kilian Weiss, Jonathan Stelter, Federica Montagnese, Marcus Deschauer, Benedikt Schoser, Claus Zimmer, Dimitrios C. Karampinos, Jan S. Kirschke, Sarah Schlaeger
Abstract
Quantitative muscle water T2 (T2w) mapping is regarded as a biomarker for disease activity and response to treatment in neuromuscular diseases (NMD). However, the implementation in clinical settings is limited due to long scanning times and low resolution. Using artificial intelligence (AI) to accelerate MR image acquisition offers a possible solution.
Introduction
As active pathological changes, patients with neuromuscular diseases (NMD) show alterations of muscle water content caused by inflammation, denervation, or dystrophy. These precede long-term chronic degeneration, characterized by intramuscular fat accumulation and fibrosis [1–3].
Materials and methods
Ten patients (5 women; mean age 49.7 years (range: 33-64 years), bodyweight mean 76.9 kg (range: 61–98 kg), height mean 182.2 cm (range: 169–198 cm), BMI mean 23.1 (range: 19–27) with facioscapulohumeral muscular dystrophy (FSHD) were recruited between July 19, 2021 and November 27, 2021.
Results
As shown in Fig 2, a visual inspection of the first and last echo of the T2w mapping raw image data revealed no difference in image quality of SENSE versus CSAI 5x T2w mapping.
Discussion
In this prospective study, we investigated the performance of a novel DL-based image acceleration approach (CSAI) for quantitative imaging. Compared to the standard acceleration with SENSE, CSAI allowed for a scan time reduction of more than 50% for T2w mapping based on the T2-prepared 3D TSE with SPAIR fat suppression, while showing excellent agreement of T2w values and comparable intra-method reproducibility.
Conclusion
DL-based acceleration of CS data allows for scan time reduction of more than 50% for T2w mapping in the thigh muscles of NMD patients without compromising quantitative validity. CSAI allowed for a significant scan time reduction for T2w mapping based on the T2-prepared 3D TSE with SPAIR fat suppression.
Citation: Schmitt J, Weidlich D, Weiss K, Stelter J, Montagnese F, Deschauer M, et al. (2025) Deep learning-based acceleration of muscle water T2 mapping in patients with neuromuscular diseases by more than 50% - translating quantitative MRI from research to clinical routine. PLoS ONE 20(4): e0318599. https://doi.org/10.1371/journal.pone.0318599
Editor: Arka Bhowmik,, Memorial Sloan Kettering Cancer Center, UNITED STATES OF AMERICA
Received: August 20, 2024; Accepted: January 17, 2025; Published: April 16, 2025
Copyright: © 2025 Schmitt et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data cannot be shared publicly because of regulations from the local ethics commission and participant privacy policies. The data underlying the results presented in the study are available from Prof. Dr. Christine Preibisch, Klinikum rechts der Isar der TU München Abteilung für Neuroradiologie, Ismaninger Str. 22 81675 München; preibisch@tum.de. In line with the journal guidelines, the person in charge is no coauthor of the study. The sharing of acquired data will be considered upon reasonable request. Institutional policies would then require a formal data sharing agreement. Christine Preibisch is a longterm employee of the institute which grantees longterm data availability.
Funding: Sarah Schlaeger was supported by a faculty internal grant (Technical University of Munich, KKF H-09). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: KW: employee of Philips Healthcare. JSK: Co-founder of Bonescreen GmbH, Lecture honoraria from Novartis.
