Introduction
2025 Scientific and Organising Committee
Linda Heskamp, Utrecht, NL (chair)
Jordi Diaz-Manera, Newcastle upon Tyne, UK (vice-chair)
Pierre Carlier, Liege, BE
Robert Carlier, Paris, FR
Simonetta Gerevini, Bergamo, IT
Hermien Kan, Leiden, NL
Ferdinand Knieling, Erlangen, DE
Afarin Madani, Brussels, BE
Valentina Mazzoli, New York, US
Anna Pichiecchio, Pavia, IT
Francesco Santini, Bazel, CH
Werner Stenzel, Berlin, DE
Volker Straub, Newcastle upon Tyne, UK
Jodi Warman-Chardon, Ottawa, CA
Dear colleagues,
We are happy to welcome you to the fifth International Imaging in Neuromuscular Disease Conference organised by the MYO-MRI+ consortium and held in Berlin from Sunday 9th to Tuesday 11th of November 2025.
This bi-annual event will bring together leading experts in the field of neuromuscular imaging and its application to neuromuscular disease, to share knowledge, network, and present cutting-edge research in the field. The Conference will not only highlight innovations in neuromuscular imaging technologies, but also their application in neuromuscular disease research and clinical practice. The main topics selected for the MYO-MRI+ 2025 Conference include: muscle MRI as an outcome measure in clinical trials, changing imaging diagnostic patterns in muscular dystrophies, and emerging imaging techniques.
More than 60 high quality abstracts have been submitted. The abstract book provides you with all abstracts ordered by oral and poster presenters.
We look forward to an interesting scientific programme and wonderful discussions at the conference!
Linda Heskamp and Jordi Diaz Manera
– chair and vice-chair of the Scientific and Organising Committee –
Invited speaker abstracts
INV01 Reproducible MRI protocols for multicentre studies
Donnie Cameron
1
1Radboud University Medical Center, Nijmegen, Netherlands
Background: MRI measures of fat replacement, in particular, are increasingly used as monitoring and response biomarkers for natural history studies and pharmaceutical trials in neuromuscular disorders. In order to deliver successful clinical trials in these conditions, multicentre studies are frequently required to increase or accelerate patient enrolment and satisfy trial endpoints. However, the reliability of pooled MRI data in a multicentre setting is influenced by differences in hardware, software, proprietary pulse sequences, and staff experience. In this invited talk, we explore some of these issues and consider how they can be addressed or minimised through, for example, MRI protocol harmonisation and ongoing quality assurance.
Aims: By the end of the session, attendees should be able to:
-Identify possible sources of inter-site error in multicentre studies;
-Identify key sequence parameters for creating reproducible MRI protocols for neuromuscular disorders;
-Set up suitable MRI quality assurance for maintaining their reproducible protocols; and
-Understand key considerations for structured data and data sharing.
Methods: This talk will cover the setup and execution of multicentre studies, from the planning phase, through image acquisition and quality control, to structured data organisation and data sharing considerations. Particular attention will be paid to neuromuscular-disorder-specific aspects. Namely, in terms of MRI methods, chemical-shift-based water-fat-separation, or ‘Dixon’, imaging will be a primary focus, with T2 mapping, diffusion tensor imaging, and proton MR spectroscopy also being considered. We will discuss how these methods can be harmonised or standardised between sites while ensuring acceptability for patients with neuromuscular disorders. Real-life examples will also be given from multicentre MRI studies past and present, including a large study currently running across three medical centres in the Netherlands.
Conclusion: MRI has a steadily increasing role in multicentre studies and clinical trials for neuromuscular disorders. This necessitates development of reproducible MRI protocols and quality assurance to facilitate pooling of data and minimisation of site errors. With this talk, we have equipped attendees with knowledge of potential sources of error and key areas for harmonisation or standardisation of MRI protocols to deliver high-quality results in their upcoming multicentre studies.
INV02 Vendor-independent programming to boost imaging use in multi-center studies
Andreia Sofia Gaspar
1Institute for Systems and Robotics - Lisboa and Department of Bioengineering, Instituto Superior Técnico – Universidade de Lisboa, Lisboa, Portugal
Background: Multicenter studies allow to evaluate the robustness of different treatments and diagnostic tools since it provides a way to test different clinical conditions and larger cohorts. In this context, protocol uniformization between centers is crucial to reduce confounding factors and increase the confidence of the studies' outcome and conclusions. Imaging techniques can give important information, but protocol standardization between centers is hindered by availability of specific methodologies that are vendor and package specific [1]. Even when available, matching the MRI sequences parameters can be out of reach due to different hardware and software available onsite. In this context, vendor-neutral framework allows to homogenize the methodology applied in different scanners and centers and improve the design of multi-centre studies [2].
Aim: In this presentation, I’ll talk about how vendor-independent MRI techniques can potentiate consistency and power of multicentre-studies, especially in the context of quantitative MRI.
Approach: The importance of quantitative MRI is grounded in its capability to have objective and reproducible measures of a specific parameter. To improve reproducibility and reduce variability in observed T1 values protocol matching is crucial [3,4].
The open-source myocardial T1 mapping (Open-MOLLI) sequence [5,6] was developed to enable reproducible cardiac T1-mapping within the Pulseq framework. Open-MOLLI has been tested in different Siemens systems and centers (1.5 T and 3.0T) and compared to the vendor MOLLI implementations. It was shown to provide comparable myocardial T1 values with equivalent same-scan repeatability [5,7].
Conclusion: This work demonstrates how vendor-neutral sequences can facilitate the inclusion of imaging techniques in future multi-center studies.
References
Karakuzu A. et al. Magn Reson Med. 2022; 88: 1212-1228.
Fujita S, et al. Magn Reson Med. 2025; 1-12.
Teixeira RPAG, et al. Magn Reson Med. 2020; 84: 221-236.
Kathryn E. et al. Magn Reson Med.2024; 1-10. 7.
Gaspar AS, et al. Magn Reson Med. 2024; 1-10. 7.
Messroghli DR, et al. J Magn Reson Imaging. 2007;26(4):1081-1086.
Gaspar AS, et al. 2025 ISMRM & ISMRT Annual Meeting & Exhibition, p.1665.
INV03 Tips and tricks for imaging processing
Francesco Santini
1Basel Muscle MRI, Department of Biomedical Engineering, University of Basel, Switzerland, 2Department of Radiology, University Hospital of Basel, Switzerland
Background: Image processing and analysis is a crucial step to convert the data acquired with an imaging system into quantitative biomarkers for objective interpretation, in the context of diagnosis or as endpoints in clinical trials. As with every step of the imaging pipeline, image processing can introduce bias, noise, and uncertainty in the final data. Thus, the characterization and standardization of image processing is of paramount importance to obtain data that are reproducible and comparable across sites and studies. Especially in the context of rare diseases, the development of generalizable and reproducible processing methods is important, since data from multiple centers often need to be pooled together to obtain statistical relevance.
Methods: In this talk, we will explore the common pitfalls that might hinder the wide applicability of image processing methods, and the possible solutions to these pitfalls. In particular, we will address the issues of data and metadata standardization, the advantages of open-source development with clear versioning control for reproducibility, and the challenges and opportunities presented by machine learning methods.
The community solution of ORMIR-MIDS as a standardized data format will be presented, and how it can be exploited to create image processing pipelines that, even when implementing different algorithms, can interoperate by standardizing the input and output conventions.
In the realm of machine learning, diagnostic-relevant metrics, and the necessity of finding a tradeoff between performance and generalizability will be discussed.
Conclusion: By designing an image processing pipeline with generalizability, interoperability, and reproducibility in mind, we can develop tools that are impactful both scientifically and practically, and can strengthen neuromuscular research as a community effort.
INV04 Ultrasound of brachial plexus and upper limb nerve
Afarin Madani
1
1Hopitaux Universitaires de Bruxelles H.U.B-Erasme Hospital, Brussels, Belgium
Background: High-resolution ultrasound (US) has emerged as a valuable noninvasive imaging tool for the evaluation of peripheral nerve disorders, including entrapments, traumatic injuries, tumors, and neuromuscular diseases. In recent years, US has gained recognition as a cost-effective complement to electrodiagnostic studies, offering high-resolution visualization of nerve structures along their course and providing diagnostic information often undetected by conventional techniques. Importantly, US assists in differentiating acute from chronic inflammatory polyneuropathies, hereditary from acquired forms, and contributes to the identification of treatment-responsive neuropathies.
Aims: The purpose of this session is twofold: (1) to outline the essential methodological and anatomical principles required for reliable interpretation of neuromuscular ultrasound, and (2) to demonstrate the practical application of these principles through a live examination of the brachial plexus and peripheral nerves of the upper limb.
Materials and Methods: Theoretical aspects include a systematic and quantitative approach to nerve imaging. Cross-sectional area (CSA) is emphasized as the most reproducible parameter for detecting neuropathies, while Doppler imaging is used to distinguish vascular structures from nerve fascicles. Muscle ultrasound is introduced as a complementary technique to assess echogenicity, thickness, and contractile dynamics. Practical application will be illustrated through live scanning of a volunteer, focusing on the brachial plexus and major peripheral nerves of the upper limb. Key landmarks will be demonstrated with attention to nerve size, echogenicity, vascularity, and surrounding perineural structures.
Results: Accumulated evidence and clinical practice indicate that neuromuscular US significantly improves diagnostic accuracy. The ability to measure CSA, assess echogenicity, and evaluate vascularity strengthens diagnostic accuracy in both focal and generalized neuropathies. Furthermore, real-time muscle ultrasound adds functional and structural insights not achievable by other imaging modalities.
Conclusion: High-resolution US is a powerful adjunct to electrodiagnosis in the assessment of peripheral nerve disorders. Its successful application requires a comprehensive knowledge of sonographic techniques and their limitations, as well as the ability to identify nerves based on anatomical features. By combining theoretical understanding with practical demonstration, this session aims to provide the skills necessary for confident and accurate interpretation of neuromuscular ultrasound.
INV05 Lower limb nerves
Marie Faruch Bilfeld
Chu Toulouse, Toulouse, France
Ultrasound is a precise and non-invasive tool to explore the nerves of the lower limb.
We will successively review the echo-anatomy of the main nerve pathways: the femoral nerve, the sciatic nerve in the gluteal region and along the posterior thigh, the tibial nerve, as well as the saphenous and sural nerves.
The goal is to show how ultrasound makes it possible to follow these nerves, identify them in their anatomical environment, and highlight its practical value in the assessment of neuropathies.
INV08 Morphological and molecular characterization of skeletal muscle and muscle imaging in late onset Pompe patients
Alexander Mensch, Lara Schlaffke,
Anne Schänzer
Background: Pompe disease is a lysosomal disorder caused by pathogenic bi-allelic variants in the GAA gene. Reduced activity of the acid alpha-glucosidase (GAA) enzyme leads to glycogen accumulation in various tissues. Striated muscle cells are most affected, and muscle biopsies reveal a vacuolar myopathy with lysosomal and extra-lysosomal glycogen deposits, as well as increased autophagy and disruption of the myofibrillar architecture. Patients with late-onset Pompe disease (LOPD) develop a proximal and axial muscle weakness, as well as respiratory impairment, with significant clinical variability. Enzyme replacement therapy (ERT) with homologous GAA slows disease progression, improves muscular function and therefore prolongs survival. Muscle imaging is a helpful non-invasive tool for monitoring disease progression under ERT.
Aims and Methods: In our overview, we present skeletal muscle pathology and molecular processes in LOPD and associate the finding with conventional MRI and diffusion MRI as well as Dixon-based imaging. For comparison, we present morphological and MR image data from patients with other non-metabolic myopathies to highlight differences in etiopathology.
Results: Morphological and molecular characterization of skeletal muscle biopsies from LOPD revealed disturbance of sarcomeric architecture and upregulation of autophagic processes. Conventional muscle MRI showed replacement with fat and collagen tissue which reflects later stages of the LOPD disease. In the GAA-/- Pompe mouse model and in LOPD patients, muscle diffusion MRI showed reduced diffusion in the early stages of Pompe disease, representing sarcomeric disruption and increased autophagy. In contrast, other genetic limb-girdle muscular dystrophies (LGMD) caused by mutations in membrane and membrane-associated proteins, showed decreased diffusion, indicating fiber atrophy in line with simulation experiments along with significantly higher water T2 values reflecting ongoing active muscle damage.
Conclusion: LOPD is a metabolic disorder with a distinct muscle morphology and molecular signature which correlates with MRI imaging and depends on disease progression. Detecting early stages of the disease with MRI is important for guiding treatment regimens.
INV09 Limb-Girdle Muscular Dystrophies: From Clinical Phenotypes to Muscle Imaging and Myopathology in the Next-Generation Sequencing Era
Willem De Ridder
2Translational Neurosciences and Peripheral Neuropathy Group UAntwerpen - Faculty of Medicine, Antwerp, Belgium, 3Institute Born-Bunge, University of Antwerp (CDE), Antwerp, Belgium
Background: Limb-girdle muscular dystrophies (LGMDs) represent a heterogeneous group of genetically inherited disorders primarily affecting the pelvic and shoulder girdle muscles. Over 30 causative genes have been identified, with autosomal recessive forms predominating. While next-generation sequencing (NGS) has revolutionized the diagnostic landscape, clinical evaluation and complementary tools such as muscle MRI and muscle biopsy remain critical in guiding diagnosis, especially in cases with variants of uncertain significance (VUS) or limited access to advanced genetic testing.
Aims: This lecture aims to provide an integrated overview of LGMDs, highlighting how clinical features, muscle imaging, and myopathological findings interact within the revolutionized diagnostic pathway. Special focus will be placed on the role of muscle MRI as a diagnostic and interpretative tool in the current era of genomic medicine.
Methods/Patients/Materials: The presentation will review LGMD subtypes through illustrative case series, integrating phenotype-genotype correlations with muscle MRI patterns and biopsy findings. Emphasis will be placed on the diagnostic utility of MRI in characterizing selective muscle involvement, guiding genetic prioritization, and evaluating disease progression.
Results: Muscle MRI shows distinctive—and in some subtypes, nearly pathognomonic—involvement patterns across specific LGMD subtypes, facilitating differential diagnosis. In many settings, MRI contributes to prioritizing gene testing and interpreting VUS. Furthermore, combining imaging with targeted histopathology can confirm pathogenicity, particularly in resource-limited contexts.
Conclusion: Despite the widespread adoption of NGS, the diagnosis of LGMDs still relies on a multidisciplinary approach. Muscle MRI not only complements clinical evaluation but also serves as a powerful tool for refining differential diagnoses, guiding biopsy, and supporting variant interpretation. Continued integration of imaging, clinicopathological data, and genomics is essential for optimal diagnosis and management of LGMDs.
INV10 Biophysical modelling of muscle microstructure
Martijn Froeling
Center for Image Sciences, Department of Imaging and Oncology, University Medical Center Utrecht, Utrecht, Netherlands
Understanding skeletal muscle microstructure is essential for assessing muscle physiology, function, and disease. While quantitative MRI (qMRI) provides sensitive markers such as water T2, proton density fat fraction (PDFF), and diffusion tensor imaging (DTI) parameters (FA, MD, RD), these measures are indirect and typically nonspecific. Biophysical modelling bridges the gap between MRI signals and the underlying tissue architecture and properties such as fiber diameter, extracellular matrix composition, fat infiltration, and inflammation.
At its core, modelling aims to both predict MRI signals from known tissue features (forward models) and infer microstructural parameters from measured data (inverse models). Relevant scales range from micrometer-level cells to whole muscles, requiring careful consideration of which aspects to capture. Model sensitivity and specificity depend on the chosen signal, acquisition parameters, and the physical assumptions underlying the model.
Modelling can be done in different ways and on different scales. For example, Diffusion-weighted imaging (DWI) can be modelled at multiple levels: Monte Carlo simulations of water diffusion can explore the influence of cell shape, the Random Permeable Barrier Model (RPBM) links diffusion-time dependence to cell size and permeability, DTI characterizes diffusion anisotropy and can be extended to multi-compartment fat–water models, and tractography quantifies how macroscopic architecture and muscle volume can influences microscopic metrics.
MRI signal interpretation faces challenges, including motion artefacts, noise, and acquisition constraints. Simulations help assess the impact of these factors and determine whether corrections are necessary. As such, modelling serves multiple purposes: improving acquisition and processing methods, linking MRI to histology and biomechanics, integrating longitudinal changes, and ultimately enabling “virtual biopsy” of muscle. Achieving this goal will require multi-parametric approaches that combine complementary qMRI contrasts across scales.
Understanding and correctly interpreting MRI-derived parameters so they can serve as a quantitative probe of muscle microstructure requires not only high-quality measured data from both patients and controls, but also biophysical modelling to reveal and explain the underlying mechanisms.
INV12 Treating NMD's - The future perspective on imaging
Andrew Blamire
1
1Translational and Clinical Research Institute, Faculty of Medical Sciences, Newcastle University, Newcastle upon Tyne, United Kingdom
Magnetic resonance imaging (MRI) is well known as a diagnostic tool providing high-definition anatomical detail, both within the central nervous and musculoskeletal systems. Its non-invasive nature makes MRI an ideal tool to follow disease progression (1, 2, 3) and response to intervention and consequently clinical trials in neuromuscular diseases have adopted MRI as secondary endpoints. Recent analysis however has shown that the number of trials incorporating imaging may have begun to decline (4). So, after several decades of imaging research, what is the future for the use of imaging in diagnosing, understanding and treating neuromuscular diseases (NMDs)?
The power of MRI rests with the capacity to link scan contrast to an array of underlying tissue characteristics, from microstructure to metabolism and function as well as gross anatomy. MR technology development continues to innovate and bring new insights into muscle disease in the research setting such as advanced in motor unit assessment (5), but translation of these advanced methods from research into endpoints or diagnostic tests takes time and investment.
Key factors for regulatory approval of imaging as endpoints for clinical trials are validation of the relationship between imaging measures and clinically meaningful patient outcomes; demonstration of cross-site standardisation of image acquisition; and use of objective analysis methods to extract the endpoint with minimal user input (6). Over the past 20 years, much work has focused on scan standardisation, and progress has been made for the simpler routine anatomical scans such as T1 weighted imaging but more complex quantitative methods such as Dixon, T2 mapping or diffusion imaging, which vary in implementation between scanners, still pose challenges.
In parallel with the work on standardisation of muscle scan acquisition, imaging has been deployed in other research domains to study population characteristics at very large (eg 100k) scale (7). Such studies provide sufficient data to train effective machine learning and artificial intelligence algorithms. These are driving ever-improving automated image segmentation and feature extraction which offer good precision (8) even from sequences previously considered to be too qualitative. Data sharing on a global scale is also allowing development of similar approaches for diagnosis in NMDs (9). These approaches perhaps bring into question the need for more advanced methodology in the routine setting. The combination of continued work on standardisation with advances in AI methods look certain to influence the way in which imaging is used in the future.
1) Veeger, TTJ et al, Baseline fat fraction is a strong predictor of disease progression in Becker muscular dystrophy. NMR in Biomedicine, 2022; 35: e4691.
2) Barnard, Amet al. MR biomarkers predict clinical function in Duchenne muscular dystrophy. Neurology, 2020; 94: E897-E909.
3) Reyngoudt, H et al. Three-year quantitative magnetic resonance imaging and phosphorus magnetic resonance spectroscopy study in lower limb muscle in dysferlinopathy. J Cachexia Sarcopenia and Muscle, 2022; 13:1850-1863.
4) Todd, M et al, Use of imaging biomarkers and ambulatory functional endpoints in Duchenne muscular dystrophy clinical trials: Systematic review and machine learning-driven trend analysis. J Neuromuscular Diseases, 2025 DOI: 10.1177/22143602251360664.
5) Heskamp, Let al. Motor Unit Magnetic Resonance Imaging (MUMRI) In Skeletal Muscle. J Magnetic Resonance Imaging, 2024; 60: 225302271.
6) FSE, Clinical Trial Imaging Endpoint Process Standards Guidance for Industry. 2018.
7) Littlejohns, TJ et al. The UK Biobank imaging enhancement of 100,000 participants: rationale, data collection, management and future directions. Nature Communications, 2020; 11:2624.
8) Thanaj, M et al. Precision MRI phenotyping of muscle volume and quality at a population scale. Frontiers in Physiology, 2024;15 :1288657.
9) Verdu-Diaz, J et al. Myo-Guide: A Machine Learning-Based Web Application for Neuromuscular Disease Diagnosis With MRI. J Cachexia Sarcopenia and Muscle, 2025;16:e13815.
INV13 Opportunities and challenges with sodium imaging
Teresa Gerhalter
1
1Medical University of Graz, Graz,
Sodium magnetic resonance imaging (23Na MRI) is an emerging modality that offers unique insights into cellular physiology and tissue viability, complementing conventional proton-based imaging. By directly quantifying sodium ion concentrations, this technique provides a non-invasive window into ion homeostasis, membrane integrity, and metabolic activity—parameters central to the pathophysiology of neuromuscular disorders.
In neuromuscular diseases such as Duchenne muscular dystrophy, disrupted sodium homeostasis often precedes overt structural damage. 23Na MRI can detect early physiological changes, offering a window into pathological changes before irreversible muscle loss occurs. This presents new opportunities for monitoring therapeutic response and understanding disease mechanisms in vivo.
Recent advances in hardware, pulse sequence design, and ultra-high-field MRI systems have significantly improved the sensitivity and spatial resolution of sodium imaging. These technological developments are enabling more reliable and clinically relevant assessments of tissue sodium concentration (TSC) in vivo, facilitating new applications in research and transition to clinics.
However, several challenges limit broader adoption. Sodium’s low natural abundance and low gyromagnetic ratio yield inherently low signal-to-noise ratio (SNR), requiring longer acquisition times or advanced reconstruction strategies. Moreover, standardization across platforms and validation against histological or biochemical gold standards remain key hurdles for quantitative sodium imaging.
This talk will explore the current capabilities and limitations of sodium MRI in neuromuscular disorders, highlighting case studies where it has provided critical physiological information beyond conventional imaging.
INV16 MRI as an outcome measure: correlation with function
Doris Leung
Kennedy Krieger Institute, Baltimore, USA
Imaging-based biomarkers have the potential to fill key gaps among the outcome measures that are currently available for clinical trials and observational studies of muscle disease. One of the important next steps in the validation of imaging-based measurements as biomarkers of disease severity will be establishing how these measurements correspond to muscle function. In this session, we will discuss ways in which investigators have evaluated the associations between muscle MRI measurements and clinical measurements of muscle strength and function in the muscular dystrophies. Data collected from multiple MRI studies of facioscapulohumeral muscular dystrophy (FSHD) will be presented to illustrate the unique challenges in predicting changes in muscle function from MRI data. We will also discuss the emerging role of artificial intelligence in modeling complex muscle function from imaging data. Ongoing longitudinal imaging studies and planned future studies in FSHD will also be presented.
INV17 MRI in myopathies: diagnosis, follow-up and opportunities for new techniques
Fengdan Wang
1
1Peking Union Medical College Hospital, Beijing, China
Background: Myopathies are disorders of skeletal muscle that affect either the muscle structure, channel, or metabolism. Significant challenges exist in the clinical evaluation of myopathies, primarily due to the absence of biomarkers capable of fulfilling the current requirements for disease staging, prognosis assessment, and therapeutic efficacy evaluation.
Aims: To investigate qMRI [fat fraction (FF), T2 map and diffusion tensor imaging (DTI)], acceleration techniques, and deep learning (DL)-based 3D automatic segmentation for rapid, accurate quantitative analysis in idiopathic inflammatory myopathy (IIM) and hereditary myopathy [Duchenne muscular dystrophy (DMD), Becker muscular dystrophy (BMD), and intermediate muscular dystrophy (IMD)].
Methods/Patients/Materials: In total, 64 patients with IIM, 21 patients with genetically confirmed hereditary myopathy, and 32 healthy controls were prospecitively included. They underwent qMRI with acceleration techniques on 3T MRI. A DL based software was used to perform muscle segmentation and evaluation.
Results: T2 values accurately differentiated patients from those of controls (p< 0.001) with a cut-off value of 36.4 ms (sensitivity 96.9%, and specificity 100%). In patients with IIM, muscle T2 values positively correlated with all the serum muscle enzymes (all p < 0.05). Compared to controls, DMD patients showed significantly higher FF, prolonged T2, reduced ADC, and decreased λ2 (p<0.05) in various muscle groups. The BMD+IMD group exhibited milder changes limited to specific parameters and muscles. Significant correlations were observed between qMRI indices and clinical variables such as Gower’s sign. Traditional enzymatic biomarkers had weak correlations with qMRI values, aligning with clinical observations that they do not reliably reflect muscle involvement or prognosis.
Conclusion: MRI advancements including qMRI, acceleration techniques, and DL-based autosegmentation offer myopathies new biomarkers to be used in diagnosis and follow-up.
INV18 Muscle and nerve MRI, from research to clinical application
Simonetta Gerevini
Neuroradiology Dept Asst Papa Giovanni XXIII Hospital, Bergamo, Italy
Background: There is not a shared and worldwide accepted MRI protocol for brachial plexus or muscle MRI from a clinical perspective. The most diffuse protocols make use of surface coils such as neurovascular coils for brachial plexus, and no coils (only the magnet one) or flex coils for muscle MRI.
Aims: To show MRI of nerve and muscle imaging in terms of clinical perspective; it means MRI protocol (techniques ie how to do things, with limitations and pitfalls), nerve and muscle imaging on the major clinical issues, so brachial and lumbar plexus acquisitions (in inflammatory disease, genetic disease and so on).
Methods/Patients/Materials: I will show a brief review of BP anatomy and MRI anatomy of muscles in terms of signals and sequences. First how to define normal from pathology: presence of fine signal alteration in supraclavicular region such as loss of fat planes, or for the presence of true brachial plexus signal alteration that can be focal or diffuse to a single root or trunk or to all the components. This can be mono lateral or bilateral, symmetric or not. We have then to look for oedema that can be intraneural or perineural as far as muscular oedema, muscular signal alteration due to denervation in all the muscular group innervated from the involved part of brachial plexus and or for specific modification of signal dimensions or aspect of single or more muscles. In some cases we can also add contrast media
The most frequent clinical issues about brachial plexus will be shown: a) different types of post traumatic lesions and how to use sequences to better define that lesions, b) most common neurogenic tumors and how to discriminate between them, c) post treatment modifications; d) TOS, e inflammatory and genetic disorders. There will be a global overview on nerve from both diagnostic and advanced techniques point of view.
INV19 Disease monitoring with whole body MR neurography
Hans Katzberg
University of Toronto, Toronto, Canada
Magnetic resonance neurography (MRN) has emerged as a valuable adjunct to electrodiagnostic testing in the evaluation of peripheral nerve disorders, providing high-resolution visualization of nerve structure and microarchitecture. In chronic inflammatory demyelinating polyneuropathy (CIDP), current EAN/PNS guidelines acknowledge the role of imaging as supportive evidence, yet robust imaging biomarkers remain under investigation.
This lecture will review the principles, acquisition strategies, and clinical applications of MRN, with a focus on diffusion tensor imaging (DTI) metrics such as fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), and axial diffusivity (AD). Earlier investigations have shown that DTI can capture microstructural alterations in peripheral nerves, offering potential diagnostic value. We will present new findings from a prospective study of treatment-naïve CIDP patients and matched healthy controls undergoing 3.0 Tesla whole-body MRN. Manual segmentation of brachial and lumbosacral plexus regions revealed significantly reduced FA in CIDP compared with controls, with the greatest differences in the brachial and sacral plexuses. Other DTI metrics showed less discriminatory value. Subgroup analysis suggested FA differences between typical and multifocal CIDP phenotypes, though FA remained stable following short-term immunotherapy.
The session will place these results in the context of the broader MRN literature, addressing technical considerations, reproducibility, and integration with other modalities such as ultrasound. Limitations, including small sample size and the need for multicenter validation, will be discussed alongside future directions for MRN as a diagnostic and potentially prognostic biomarker in inflammatory neuropathies.
INV20 Identifying muscles conducive to monitoring the progression of oculopharyngeal muscular dystrophy (OPMD) with quantitative 2-point DIXON techniques
Ian C. Smith1,2, Fadi Estaffio3, Tanay Sarkar3, Sono Khan3, Kaitlynn Meier-Ross3, Gerd Melkus1,5, Marcos L Sampaio1,5, Jodi Warman-Chardon1,3,5
1Ottawa Hospital Research Institute, Ottawa, Canada, 2Eric Poulin Centre for Neuromuscular Disease, The Ottawa Hospital, Ottowa, Canada, 3Faculty of Medicine, University of Ottawa, Canada, 4Department of Radiology, Radiation Oncology and Medical Physics, University of Ottawa, Canada, 5Department of Physics, Carleton University, Canada, 6Department of Medicine - Division of Neurology, University of Ottawa, Canada
Background: Oculopharyngeal muscular dystrophy (OPMD) is a late-onset neuromuscular disorder characterized by eyelid ptosis, dysphagia, and proximal limb weakness. The slow and variable clinical progression of OPMD necessitates development of sensitive and objective tools that are suitable for use in clinical trials on treatments in development.
Methods/Patients/Materials: We assessed fatty replacement in 97 muscles (47 muscles assessed bilaterally, plus tongue) using 2-point DIXON techniques in patients with OPMD (n=22) and in controls (CON; n=4). Muscle cross-sections were traced using ITK Snap, and muscle fat fraction (FF) was calculated using fat-only and water-only images. The OPMD group was subdivided into participants with the lowest average FFs (OPMDLow) and highest average FFs (OPMDHigh). Muscles were then ranked on 5 criteria: i) mean difference in FF between CON and OPMDLow, ii) mean difference in FF between OPMDHigh and OPMDLow, iii) mean difference in FF between OPMDHigh and CON, iv) change in FF in longitudinal scans, and v) correlation of single muscle FF with mean FF of all 97 muscles in OPMD. Ranks were averaged bilaterally, with the former 3 criteria establishing disease relevance (single weighted), and the latter two criteria serving as sensitivity indicators (double weighted).
Results: The top 7 ranked muscles were biceps femoris long head, adductor magnus, gluteus medius, semimembranosus, gluteus maximus, soleus, and semitendinosus. Across these 7 muscles, the mean difference in FF between CON and OPMDLow ranged from 4.1%-12.3%, and mean change in FF over 2 years ranged from 1.5% to 5.7%. Several muscles (e.g, serratus anterior, tongue, lumbar paraspinals) offered excellent distinction between OPMD and CON but ranked poorly on one or both sensitivity indicators.
Conclusion: Muscles appearing most suitable as outcome measures for clinical trials were concentrated to the posterior aspect of the thigh and pelvis. Applied strategically, this information can reduce total scan time and total costs, and obviate positional dysphagia experienced by many participants with OPMD during whole body MRI.
Oral Presentations
O1.1 Non-invasive assessment of histological changes in dystrophic and developing skeletal muscles in GRMD and Control Dogs using Bi-Component T2 Relaxometry Mapping
Ericky Caldas de Almeida Araujo1, Inès Barthelemy2, Yves Fromes1, Stéphane Blot2, Benjamin Marty1,
Harmen Reyngoudt1
1Institut de Myologie, Centre d’Evaluation et d’Exploration Neuromusculaire, Laboratoire de RMN, Paris, France, 2Université Paris Est Créteil, INSERM, IMRB, Créteil, France; EnvA, IMRB, Maisons-Alfort, France
Background: There is an urgent need in clinical neuromuscular research for novel imaging biomarkers capable of detecting pathophysiological tissue alterations across entire muscles, non-invasively and with high specificity.
Aim: To investigate the feasibility of evaluating histological alterations in dystrophic and developing skeletal muscle using bi-component T2 water relaxometry mapping.
Methodology: Twelve Golden Retriever dogs were examined longitudinally at three time points between 2 to 12 months of age. Eight were affected by Golden Retriever Muscular Dystrophy (GRMD), and four served as healthy controls. MRI was performed at 3 Tesla, targeting the distal pelvic limbs. Fat fraction (FF) maps were generated using the Dixon method, and bi-component T2 maps were derived from multi-slice, multi-echo spin echo (MSME) data by fitting the signal to a bi-component dictionary based on the Extended Phase Graph (EPG) formalism. In a second, comparable cohort (12 dogs: 8 GRMD and 4 controls), histological analysis of the Tibialis cranialis muscle was conducted longitudinally over the same age range (2–12 months).
Results: The mean FF remained below 5% in all animals, with no significant differences between GRMD and control dogs, nor age-dependent changes. In contrast, the short T2 (T21) and the relative fraction of the long-T2 component (A2) decreased significantly with age in both groups but were significantly higher in GRMD dogs across all age ranges.
Histological analysis confirmed the absence of adipocytes in both groups. In GRMD dogs, signs of necrosis and inflammation were observed, along with interstitial connective tissue volume at 12 months of age. In both groups, muscle fiber diameter increased progressively with age, accompanied by a corresponding decrease in the relative volume of extracellular space. In GRMD dogs, inflammatory and necrotic processes decreased over time.
Conclusions: These findings highlight the sensitivity of bi-component water T2 relaxometry metrics to specific pathological and morphological alterations in both GRMD and control dogs during muscle development. Unlike GRMD dogs, boys with Duchenne muscular dystrophy typically present with substantial fat infiltration, making lipid-related signal contamination a significant source of bias. Future work should address this limitation to support the clinical translation of this methodology.
O1.2 Conventional and in-magnet cardiopulmonary exercise testing of patients with neuromuscular disease to investigate peripheral causes of exercise intolerance
Melissa Hooijmans1,2, Joost Raaphorst3, Nicole Voermans4, Fieke Koopman5, Aart Nederveen2, Gustav Strijkers6, Jeroen Jeneson1,7, Eric Voorn5
1Department of Human Movement Sciences, Faculty of Behavioral and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, Netherlands, 2 Department of Radiology and Nuclear Medicine, Amsterdam UMC, University of Amsterdam, Amsterdam Movement Sciences, Amsterdam, Netherlands, 3Department of Neurology, Amsterdam UMC, University of Amsterdam, Amsterdam Neuroscience, Amsterdam, Netherlands, 4Department of Neurology, Radboud University Medical Center, Nijmegen, Netherlands, 5Department of Rehabilitation Medicine, Amsterdam UMC, University of Amsterdam, Amsterdam Movement Sciences, Amsterdam, Netherlands, 6Department of Biomedical Engineering and Physics, Amsterdam UMC, University of Amsterdam, Amsterdam Movement Sciences, Amsterdam, Netherlands, 7Center for Child Development and Exercise, Wilhelmina Children's Hospital, Division of Child Health, University Medical Center, Utrecht, Netherlands
Background: Many patients with neuromuscular disease (NMD) experience exercise intolerance which significantly impacts daily-life-activities. The contribution of other factors than neurological control of muscle function, such as microvascular and mitochondrial function, is in many cases poorly understood. Clarification of their contribution is essential to develop targeted interventions. Interleaved dynamic 1H-MR imaging and 31P-MR spectroscopy (31P-MRS) enable simultaneous investigation of muscle microvascular responsiveness and mitochondrial function.
Aim: Investigate microvascular and mitochondrial function in upper leg muscles in patients with NMD versus controls performing a maximal in-magnet cycling test.
Methods/patients/materials: Nineteen NMD patients (Chronic Progressive External Ophthalmoplegia (CPEO); n=9; PMMSA-4FS (range=6-13), Inclusion Body Myositis (IBM); n=10; IBMFRS (range:19-39)) and ten controls underwent a 3T MRI examination. We performed interleaved qT2* mapping (15 echoes; TR/TE/dTE: 27/1.1/1.78ms) and 31P MRS (pulse acquire; TR/TE 4000/0.1ms) in the upper leg muscles before, during and after maximal leg-cycling test with a 4-second temporal resolution. From these datasets we derived three MRI parameters (MAX hyperaemia (%), time-to-peak (TTP; s) and initial slope hyperaemic response (IS; %/s)) to assess microvascular responsiveness and three MRS parameters (change in phosphocreatine level (ΔPCr; %), end-exercise pH drop during exercise and time to 95% recovery of PCr (95PCrRT; s) post-exercise) to evaluate mitochondrial function. All NMD patients underwent a cardiopulmonary exercise test (CPET) on a bicycle ergometer outside the MRI to assess the level of exercise intolerance. Differences between-groups were assessed using a Kruskal-Wallis test.
Results: All participants successfully completed both exercise protocols. Median percentage predicted maximal oxygen uptake during CPET was 66.9% in CPEO and 70,1% in IBM. Similar levels of median ΔPCr and end-exercise pH were found in participants with CPEO (81.1%; pH:6.8), IBM (79.9%; pH:6.8) and controls (80.2%; pH:6.8). Median 95%PCrRT was significantly longer in CPEO (214s; p=0.003) compared to controls (86s) and showed a wide range in IBM (56- 206s). The capillary measures showed considerable variation within-groups, but no differences were detected at the group-level.
Conclusion: Interleaved T2*-mapping and 31P-MRS during and post dynamic in-magnet leg exercise revealed mitochondrial dysfunction in CPEO patients and, to a lesser degree, in IBM, potentially underlying their exercise intolerance.
O1.3 Mapping skeletal muscle mitochondrial oxidative phosphorylation in health and SMA using a novel technique OXCEST MRI
Ritambhar Burman1, Samuel Hughes1, Lauren Wooten1, Richard Finkel1, Puneet Bagga
1
1St. Jude Children's Research Hospital, Memphis, United States
Background: Mitochondrial oxidative phosphorylation (OXPHOS) is a driver of muscle activity and physical function. Non-invasive MR-based imaging techniques can demonstrate dynamic OXPHOS activity by capturing the rate of phosphocreatine (PCr) or creatine (Cr) recovery following an exercise activity. We have developed a novel technique, OXPHOS CEST (OXCEST) technique, with improved temporal resolution that provides high Cr rate of recovery measurement accuracy. Our preliminary results demonstrate that OXCEST provides Cr recovery times more closely aligned with 31P-MRS, while also allowing for muscle-specific OXPHOS mapping that 31P-MRS cannot provide. Disrupted mitochondrial function is a common feature to neuromuscular and metabolic diseases. Among neuromuscular diseases, spinal muscular atrophy (SMA) is an excellent choice to study this novel technique. Growing evidence suggests that muscle mitochondrial dysfunction plays a key role in pathophysiology of SMA.
Aim: Development and validation of OXCEST for skeletal muscle OXPHOS imaging in healthy participants and SMA patients.
Methods: We acquired OXCEST and 31P-MRS in seven healthy subjects (male=5, age=34.6±6.1Y) across two sessions to compare OXCEST-derived post-exercise Cr recovery time constant (TCr) with that of PCr (TPCr). OXCEST scanning protocol consists of Pre- and Post-scan reference images, B0/B1 mapping and OXCEST scans separated by a 2-min plantar flexion exercise. OXCEST was measured in lateral (LG) and medial gastrocnemius (MG), and soleus (Sol) muscle groups. During each acquisition, participants performed mild plantar flexion exercise inside the scanner. Additionally, we acquired OXCEST in one SMA Type 3 patient (22Y, male).
Results: OXCEST-derived Cr post-exercise recovery in both LG and MG muscle groups followed a mono-exponential decay (R2>0.97). In healthy cohort, mean TCr was 57±23s and mean TPCr was 51±15s, showing strong correlation (R2=0.83, p=0.005) between OXCEST and gold standard 31P-MRS. In one SMA subject, LG and MG TCr was found to be longer (89sec) compared to an age-matched healthy subject (26sec), suggesting dysfunctional skeletal muscle mitochondrial function.
Conclusion: OXCEST enables high-temporal resolution and reliable assessment of post-exercise Cr recovery. Its strong agreement with 31P-MRS supports its potential as a non-invasive alternative for evaluating skeletal muscle oxidative metabolism. OXCEST has a translatable potential for monitoring mitochondrial OXPHOS in SMA and other neuromuscular disorders.
O1.4 Magnetization transfer Imaging in late-onset Pompe disease
Michele Giovanni Croce1, Farah Naz2, Leonardo Barzaghi2,3, Matteo Paoletti3, Tiziana Enrica Mongini4,
Serena Gasperini5, Massimiliano Filosto6,7, Lorenzo Maggi8, Annalisa Sechi9, Marina Grandis10,11, Michele Sacchini12, Monica Sciacco13, Chiara Bonizzoni3, Niels Bergsland14,15, Francesco Santini16,17, Xeni Deligianni16,17,
Claudia Angela Michela Gandini Wheeler-Kingshott1,18,19, Sabrina Ravaglia20, Anna Pichiecchio1,3
1Department Of Brain And Behavioural Sciences, University Of Pavia, Pavia, Italy, 2Department Of Mathematics, University Of Pavia, Pavia, Italy, 3Advanced Imaging And Artificial Intelligence Center, Neuroradiology Department, IRCCS C. Mondino Foundation, Pavia, Italy, 4Neuromuscular Unit, Department Of Neurosciences, University Of Turin, Turin, Italy, 5Department Of Pediatrics, Università Degli Studi Milano-Bicocca, Fondazione MBBM, San Gerardo Hospital, Monza, Italy, 6Department Of Clinical And Experimental Sciences, University Of Brescia, Brescia, Italy, 7NeMO-Brescia Clinical Center for Neuromuscular Diseases, Brescia, Italy, 8Neuroimmunology And Neuromuscular Diseases Unit, Fondazione IRCCS Istituto Neurologico "Carlo Besta", Milan, Italy, 9Regional Coordinator Center For Rare Diseases, Academic Hospital Of Udine, Udine, Italy, 10Università Di Genova, Genoa, Italy, 11IRCCS Ospedale Policlinico San Martino, Genoa, Italy, 12Unit Of Hereditary Metabolic And Muscular Disorders, Meyer Children's University Hospital, Firenze, Italy, 13IRCCS Fondazione Ca' Granda Ospedale Maggiore Policlinico, Neuromuscular And Rare Disease Unit, Milan, Italy, 14Department Of Neurology, Jacobs School Of Medicine And Biomedical Sciences, Buffalo Neuroimaging Analysis Center, University Of Buffalo, The State University Of New York, Buffalo, NY, United States, 15IRCCS, Fondazione Don Carlo Gnocchi Onlus, Milan, Italy, 16Department Of Radiology, University Hospital Basel, Basel, Switzerland, 17Basel Muscle MRI, Department Of Biomedical Engineering, University Of Basel, Basel, Switzerland, 18Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, Faculty of Brain Sciences, University College London, London, United Kingdom, 19Brain Connectivity Center, C. Mondino National Neurological Institute, Pavia, Italy, 20IRCCS C. Mondino Foundation, Pavia, Italy
Background: Identifying early biomarkers of muscle damage is essential for disease monitoring in late-onset Pompe disease (LOPD). Conventional MRI techniques often detect changes only after fat replacement has occurred.
Aims: To evaluate the sensitivity of Magnetization Transfer Ratio (MTR) as a non-invasive biomarker for early muscle changes in LOPD.
Methods/Patients/Materials: Quantitative muscle MRI of the thigh was performed on 32 LOPD patients (15 asymptomatic, 17 symptomatic) and 34 healthy controls (HCs). Imaging protocol included 6-point Dixon for fat fraction (FF) quantification and magnetization transfer (MT) imaging. Eleven thigh muscles were individually evaluated for this analysis: rectus femoris (RF), vastus lateralis (VL), vastus intermedius (VI), and vastus medialis (VM), sartorius (S), gracilis (G), adductor magnus (AM), semimembranosus (SM), semitendinosus (ST) and both heads of the biceps femoris (BFL, BFS). ROIs were drawn using open source DAFNE software and subsequently validated by an expert operator. FF was computed using the Fatty-Riot Algorithm. MTR was calculated on a pixel-wise basis as: MTR (%) = (Mo − MSat) / Mo × 100 where Mo and MSat refer to images without and with saturation pulse, respectively. Subsequently, MTR values were extracted from each slice position and averaged over all 10 slice positions for each participant. Clinical assessments included manual muscle testing, 6-minute walking test, functional scales, spirometry, and patient-reported outcomes. Correlations between FF, MTR, and clinical measures were analysed. We also assessed whether MTR could detect early fiber damage in muscles with FF < 10%.
Results: MTR was lower, both in the whole thigh and in most individual muscles, in patients compared to controls, and in symptomatic compared to asymptomatic and HCs. Asymptomatic showed lower MTR values compared to controls only in AM and G. In muscles with FF < 10%, only AM, G and BFS showed lower MTR values in patients compared to controls, while asymptomatic patients had lower MTR values than HCs only in AM and G. MTR values negatively correlated with FF and showed significant correlations with clinical tests.
Conclusion: MTR could represent a promising biomarker for monitoring early signs of progression in LOPD, being able to detect mild muscle fiber damage.
O2.1 Towards an automated approach to muscle MRI segmentation, quantification and analysis for the characterisation and diagnosis of neuromuscular diseases
Jose Verdu Diaz1, Carla Bolaño-Díaz1,
Alejandro Gonzalez Chamorro1, Longdan Hao1,
Sam Fitzsimmons1, Volker Straub1, Giorgio Tasca1, Jaume Bacardit2, Jordi Díaz-Manera1, on behalf of the Myo-Guide Consortium
1John Walton Muscular Dystrophy Research Centre, Newcastle University, Newcastle upon Tyne, United Kingdom, 2Interdisciplinary Computing and Complex BioSystems (ICOS) research group, Newcastle University, Newcastle upon Tyne, United Kingdom
Background: Neuromuscular diseases (NMDs) comprise a diverse group of disorders affecting skeletal muscle structure and/or function, often leading to progressive degeneration and replacement of muscle tissue with fat and fibrosis. Diagnosing NMDs is challenging due to clinical heterogeneity, genetic complexity, and overlapping phenotypes, typically requiring time-consuming, multidisciplinary assessments. Muscle MRI has emerged as a powerful diagnostic tool by revealing disease-characteristic patterns of intramuscular fat replacement. However, its full clinical potential is hindered by the manual effort required to interpret these complex imaging patterns. To address this, we introduce Myo-Guide, a unified and automated framework for segmentation, quantification, and analysis of muscle MRI, focused on diagnostic and phenotypic characterisation of NMDs.
Methods: To support the development of Myo-Guide, we established the Myo-Guide Consortium, a collaborative network of 62 clinical and academic sites across 20 countries. We assembled a large, curated database of over 3500 muscle MRI scans (T1-weighted and Dixon axial sequences) spanning more than 100 neuromuscular disorders. A subset of annotated scans was used to train a deep-learning segmentation model based on the U-Net++ architecture. This model was then used to segment the remaining dataset in a fully automated fashion. A semi-automated pipeline was developed to quantify intramuscular fat content from both T1-weighted and Dixon images. Quantitative measures (including global fat fraction, spatial distribution, and radiomic texture features) are being analysed using supervised classification and unsupervised clustering to uncover disease-specific and differential patterns of muscle involvement.
Results: The segmentation model achieved a DICE score above 0.96, performing well even in highly fat-replaced muscles. The quantification pipeline has been successfully implemented for both T1w and Dixon scans. Preliminary analyses suggest consistent, disease-related patterns across several NMD subtypes. Full-scale validation, comparative analysis across diseases, and integration of clinical metadata are ongoing.
Conclusion: Myo-Guide represents a major step toward automated, standardised use of muscle MRI in diagnostic and research workflows. By enabling large-scale, reproducible quantification of muscle involvement, it holds promise for improving diagnostic accuracy, refining disease classification, and supporting patient stratification in clinical trials.
O2.2 Comparison of manual vs. Artificial Intelligence-based muscle segmentation for evaluating disease progression in patients with CMT1A
Etienne Fortanier1,2, Marc-Adrien Hostin1,3,
Constance. P Michel1, Emilien Delmont2, Maxime Guye1, Marc-Emmanuel Bellemare3, Shahram Attarian2,
David Bendahan
1
1CRMBM, Marseille, France, 2Referral Center for neuromuscular diseases and ALS, Marseille, France, 3CNRS, LIS, Aix Marseille University, Marseille, France
Background: Intramuscular fat fraction (FF), assessed with quantitative MRI (qMRI), has emerged as one of the few responsive outcome measures in CMT1A patients. The main limitation for its use in future trials is the time required for the manual segmentation of individual muscles.
Aims: This study aimed at assessing the accuracy and responsiveness of a fully automatic Artificial Intelligence-based (AI)segmentation pipeline to evaluate disease progression in a cohort of CMT1A patients.
Methods: Twenty CMT1A patients were included in this observational prospective longitudinal study. FF was measured twice a year using qMRI in the lower limbs. Muscle segmentation was performed fully automatically using a trained convolutional neural network with or without a human quality check (QC). The corresponding results were compared with the Dice similarity coefficient (DSC) with those obtained from a fully manual (FM) segmentation. In each case, FF progression and its standardized Response Mean (SRM) was computed in individual muscles over the single central slice and a 3D muscle volume at thigh and leg levels to define the most sensitive region of interest.
Results: AI-based segmentation showed excellent DSC values (>0.90). A significant global FF progression was observed at thigh (+0.71 ± 1.28%; p=0.016) and leg levels (+1.73 ± 2.88%, p=0.007), similarly to what was computed from the FM technique (p=0.363 and p=0.634). FF progression of each individual muscle was comparable when computed either from the central slice or the 3D volume. The time necessary for the fully automatic segmentation process using AI with a QC was10 hours for the entire dataset as compared to 90 hours for the FM.
Conclusion: qMRI combined to AI-based segmentation can be considered as a process ready for assessing the longitudinal FF changes in CMT1A patients. Given the slow FF progression and the large heterogeneity among muscles and individuals, FF should be quantified from a 3D volume at the leg level.
O2.3 Quantitative muscle MRI for imaging denervation-induced muscle changes in patients with chronic inflammatory demyelinating polyneuropathy (CIDP)
Petros Chatziandreou1, Johannes Forsting1,
Johanna Thomä1, Martijn Froeling5, Rafael Klimas4, Jeremias Motte4, Anna Lena Fisse4, Kalliopi Pitarokoili4, Lara Schlaffke1,2,3, Elena Enax-Krumova1
1Department of Neurology, BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Heimer Institute for Muscle Research, Bochum, Germany, 4Department of Neurology, St. Josef-Hospital, Bochum, Germany, 5Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Chronic inflammatory demyelinating neuropathy (CIDP) is the most common autoimmune-mediated neuropathy. Despite considerable advances in understanding cellular and molecular pathophysiology and the emergence of advanced therapies, major challenges persist in diagnosing and monitoring. One recent research topic focuses on non-invasive biomarkers for the detection of disease-specific changes in early stages and during disease progression. Quantitative muscle MRI (qMRI) is a promising technique for evaluating muscle injury, inflammation, and degeneration.
Aim: Utilizing qMRI for depicting denervation-caused muscle changes in patients with CIDP, and to correlate these with clinical findings.
Methods: 11 leg muscles from 8 CIDP patients (3 women, age 59.0±10.28 years, BMI 25.92±3.93 kg/m2) and 8 aged and sex-matched healthy controls (HC) were assessed. Clinical assessment included electrophysiology, muscle ultrasound, the MRC Sum Score and R-ODS for functional impairments. MRI scans were performed on a 3T-MRI, including a Dixon-based sequence to measure the fat fraction (FF), a T2-mapping sequence to quantify the water relaxation time, and diffusion sequences to calculate the diffusion parameters FA, MD and RD. A MANOVA was performed to compare patients to HC followed by Tukey post-hoc tests. Kendall's tau was used to correlate the MRI parameters with clinical findings.
Results: The FF (Δ3.617%; p<0.001) and wT2 relaxation time (Δ2.504ms; p<0.001) were significantly higher in CIDP patients. When comparing low-fat-infiltrated muscles (<10%), the difference for the wT2 remained significant (Δ1.989ms; p<0.001). The muscle-wise post-hoc analysis showed significant changes in the FF in 5 muscles and wT2 time in 10 muscles. Correlations with clinical parameters were significant between the MRC Sum Score and FA (τ -0.617; p<0.05) and between the echogenity of the medial gastrocnemius muscle in ultrasound and FA as well as wT2 (τ 0.823, p<0.05).
Conclusion: Quantitative MRI parameters differ between CIDP patients and HC. Notably, wT2 relaxation time emerges as a promising biomarker of disease activity, showing its sensitivity to pathological changes, even in low-fat infiltrated muscle. To validate these findings and support further therapeutic optimization, studies with larger cohorts and extended follow-up periods are essential.
O2.4 The utility of quantitative MRI parameters in monitoring disease progression in patients with Myotonic dystrophy type 2.
Viktória Kokošová1,2, Peter Krkoška1,2,
Daniela Vlažná1,2,3, Marek Dostál2,4, Petra Ovesná5, Kateřina Matulová2, Blanka Adamová1,2
1Department of Neurology, Centre for Neuromuscular Diseases (Associated National Centre in the ERN EURO-NMD), University Hospital Brno, Brno, Czech Republic, 2Faculty of Medicine, Masaryk University, Brno, Czech Republic, 3Department of Rehabilitation, University Hospital Brno, Brno, Czech Republic, 4Department of Radiology and Nuclear Medicine, University Hospital Brno, Brno, Czech Republic, 5Institute of Biostatistics and Analyses Ltd., Brno, Czech Republic
Background: Myotonic dystrophy type 2 (DM2), an adult-onset hereditary muscle dystrophy, involves impairment of limb as well as axial lumbar paraspinal muscles (LPM). The advent of new therapeutic possibilities for neuromuscular disorders led to dire need for biomarkers of disease diagnosis and progression. Quantitative muscle MRI (qMRI), a rapidly evolving field, bears potential to provide such clinically significant biomarkers.
Aims: To evaluate fat fraction (FF) and functional muscle volume (FMV) of LPM and psoas muscle (PS) as potential biomarkers of disease progression with ageing in DM2 patients.
Methods: Data from 35 propensity score matched pairs of DM2 patients and healthy volunteers (HV) were analyzed. All participants underwent MRI of lumbar spine and LPM utilizing a 6-point Dixon gradient echo sequence and FF and FMV of LPM and PS were calculated. Multivariable generalized regression models with FF and FMV of LPM and PS, separately, as dependent variables and age and group as explanatory variables were built to assess age-dependent evolution of qMRI parameters and its difference in LPM and PS between HV and DM2 patients.
Results: The FF of LPM and PS was significantly higher in DM2 patients compared to HV (P < 0.001) and increased significantly with age in both groups (P < 0.001). Nevertheless, there was no significant difference in the rate of its increase in LPM with ageing between DM2 patients and HV. On the contrary, FF of PS in DM2 patients evinced 16% greater increase per decade (P = 0.014) compared to HV. FMV of PS, but not LPM, was lower (P < 0.001) in the DM2 patients. FMV of PS and LPM decreased non-significantly with ageing in both groups.
Conclusion: In DM2 patients the FF of PS, but not LPM, can be utilized as a biomarker of disease progression. FMV of LPM or PS is not a reliable biomarker of disease progression.
The study was supported by the Ministry of Health of the Czech Republic project for conceptual development in research organizations No. 65269705 (University Hospital Brno, Brno, Czech Republic), specific research project MUNI/A/1522/2024, and the European Reference Network for Neuromuscular Diseases, Project ID no. 870177.
O2.5 Quantitative muscle MRI in LGMDR1: Insights from a prospective longitudinal cohort study
Robert Rehmann1,3, Marian Wächter1, Martijn Froeling4, Marlena Rohm1,3, Anne-Katrin Güttsches1,3,
Alice De Lorenzo1, Nicolina Südkamp1,3,
Abdulhadi Kocabas1, Elena Enax-Krumova1,
Matthias Vorgerd1,3, Lara Schlaffke1,2,3, Johannes Forsting1
1Department of Neurology, BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Heimer Institute for Muscle Research, Bochum, Germany, 4Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Limb-girdle muscular dystrophy type R1 (LGMDR1) is the most prevalent LGMD subtype in Europe, characterized by progressive muscle degeneration. Quantitative MRI (qMRI) offers non-invasive insights into disease pathology and progression.
Aim: This prospective longitudinal study evaluated qMRI and clinical parameters over one year to assess disease progression and identify sensitive imaging biomarkers in LGMDR1 patients.
Methods: Thirteen genetically confirmed LGMDR1 patients and 13 matched healthy controls underwent baseline and 12-month follow-up assessments. MRI protocols included Dixon-based fat fraction (FF) mapping, water T2 mapping, and diffusion tensor imaging (DTI). Clinical outcomes included QMFM, ACTIVLIM, and standard functional tests.
Results: Significant deterioration was observed in ACTIVLIM (p = 0.029) and QMFM (p = 0.016). qMRI revealed increased FF and T2 values over time (p < 0.001 and p = 0.016, respectively), particularly in muscles with moderate baseline fat infiltration (10–50%). T2 changes in thigh muscles correlated significantly with clinical deterioration (ρ = −0.621, p < 0.05). No significant longitudinal changes were seen in DTI metrics.
Conclusion: qMRI provides valuable, sensitive markers of disease progression in LGMDR1, especially FF and T2 mapping. FF showed highest responsiveness in moderately affected muscles. Water T2 appears promising for early detection of active muscle damage. These metrics may serve as future outcome measures for clinical trials.
O3.1 Motor Unit Magnetic Resonance Imaging (MUMRI) as a novel biomarker in Spinal Muscular Atrophy (SMA)
Matthew Birkbeck1,2, Ian Schofield1, Ian Wilson1,
Julie Hall1,3, Chiara Bettolo1,4, Volker Straub1,4,
Roger Whittaker1, Andrew Blamire1
1Translational and Clinical Research Institute (Newcastle University), Newcastle upon Tyne, United Kingdom, 2Northern Medical Physics and Clinical Engineering (Newcastle upon Tyne NHS Foundation Trust), Newcastle upon Tyne, United Kingdom, 3Department of Neuroradiology (Newcastle upon Tyne NHS Foundation Trust), Newcastle Upon Tyne, United Kingdom, 4John Walton Muscular Dystrophy Research Centre (Newcastle University), Newcastle Upon Tyne, United Kingdom
Background: Spinal muscular atrophy (SMA) an autosomal recessive disease characterised by the loss of the secondary motor neurons leads to profound muscle weakness. Genetic therapies are rapidly advancing SMA trials, however, treatments need to be delivered early in the disease. A sensitive and non-invasive biomarker to monitor treatment response is required. We have developed a technique called motor unit MRI (MUMRI), which non-invasively detects fasciculation, a common symptom of spinal muscular atrophy (SMA).
Aims: Apply MUMRI in SMA patients and controls to compare fasciculation rates.
Methods: We included 10 patients (all SMA III;6 male;8 walkers,2 sitters) and 10 age comparable and sex matched controls. Three patients had scoliosis. Images of the tongue, upper right arm, paraspinal and bilateral lower limbs (thighs & calves) were acquired using 3-point Dixon and MUMRI (pulsed gradient spin echo) MRI sequences. Muscles were manually delineated, and fat fraction and fasciculation rates were calculated using custom algorithms.
Results: At group level, fat fraction %, was higher in SMA compared to controls for: upper arm (35.0±25.4 vs. 4.2±1.1, p<0.0001), paraspinal (41.4±31.0 vs. 7.4±4.5, p=0.002), thighs (54.8±23.8 vs. 5.7±1.0, p<0.0001) and calves (29.6±23.5 vs. 4.4±0.9, p=0.0003), but not for the tongue (13.9±3.2 vs. 13.0±3.3, p=0.393). Paraspinal fat fraction was significantly higher patients with scoliosis than without (73.3±10.3 vs. 27.8±26.0, p=0.033). Fasciculation rate (number of signal voids per cm3 muscle tissue per minute, cm-3min-1) was higher in SMA compared to controls for: upper arm (0.28±0.61 vs. 0.002±0.001, p=0.014), paraspinal (0.06±0.06 vs. 0.003±0.005, p=0.001), thighs (0.46±0.57 vs. 0.008±0.005, p=0.002) and calves (0.37±0.58 vs. 0.02±0.02, p=0.001), but not for the tongue (0.17±0.17 vs. 0.05±0.08, p=0.082).
Discussion: MUMRI has detected significantly higher fasciculation rates in cervical, thoracic and lumbar innervated muscles, but not bulbar. Although further work in larger cohorts is required, MUMRI is an attractive non-invasive novel biomarker which could be used to monitor progression & response in clinical trials of SMA.
We showed significantly higher fat fraction in the paraspinal muscles, a novel finding in SMA III. We found a higher fat fraction in patients with a scoliosis than without, suggesting Dixon MRI could be used to monitor scoliosis progression.
O3.2 Fasciculations detection in the legs of healthy volunteers using DTI
Karleen Oonk1, Linda Heskamp2, Boudewijn Sleutjes1, Stephan Goedee1, Leonard van den Berg1,
Martijn Froeling2
1Brain center, Department of neurology and neurosurgery, University Medical Center Utrecht, Utrecht, Netherlands, 2Center for Image Sciences, Department of Imaging and Oncology, University Medical Center Utrecht, Utrecht,
Background: Amyotrophic lateral sclerosis (ALS) is a neuromuscular disease characterized by muscle weakness, atrophy, and fasciculations. Fasciculations are an important diagnostic criterium for early diagnosis of ALS. A novel method to detect fasciculations is diffusion weighted imaging (DWI). However, fasciculations can also be observed in healthy individuals, particularly in the lower limbs, without clinical significance. Therefore, reference values for fasciculation rate and size across the entire lower extremity muscles (origin to insertion) of a large healthy cohort are needed. Furthermore, fasciculation rate and size may be influenced by age, BMI and physical activity, but this has not been systematically investigated in a large cohort.
Aim: To quantify fasciculation rate and size in lower extremity muscles in a cross-sectional cohort of healthy volunteers, and to assess the effects of age, BMI, and physical activity.
Methods: Healthy subjects were scanned supine using multi-element posterior-anterior coils. The diffusion tensor imaging (DTI) protocol included b-values of 20(3x), 50(3x), 200(6x), and 500(15x) s/mm2, with unique gradient directions, and TR/TE of 5886ms/55ms. DTI parameters: FOV 450×276mm2, in-plane resolution 3×3mm2, 31 slices (6mm thick, no gap). Full legs were scanned in 5–6 stacks with triple fat suppression, SENSE 2.4, and 3:02min scan time per stack. Data-processing involved image denoising, registration and normalisation. The fasciculations were detected using a custom-built detection algorithm (qmritools.com).
Results: We have included 99 out of the 162 subjects so far, with an average age of 36 years (range: 16-60 years) and BMI of 23.7kg/m2 (range: 18-30kg/m2). The medial and lateral gastrocnemius, and the soleus exhibited the highest fasciculation rates (range: 0.25-1.5%/dm3). Fasciculation size was largest in the soleus, vastus lateralis, and biceps femoris longus (range: 0.5-1.5cm3). A weak positive correlation was observed between age and fasciculation rate (r=0.3, p=0.003), and volunteers who exercised within 48h showed 1.5x higher rates than those who did not.
Conclusion: The results are in line with previous research, showing higher fasciculation rates in the posterior calf. The fasciculation rates in our study were lower compared to literature. This may be due to the inclusion of the entire muscle, rather than the muscle belly with the largest muscle volume.
O3.3 MRI quantification of upper extremity muscle fat fraction in Dystrophinopathies: implications for mobility status stratification
Prathyusha Bellam1, Kelly Rock1, Rebecca J. Willcocks1, Alison M. Barnard1, Donovan J. Lott1, Sean Forbes1, Claudia Senesac1, William Rooney2, Kirsten Zilke2,
Eric Baetscher2, Sub Subramony1, Nizar Chahin2,
Glenn Walter1, Krista Vandenborne1
1University of Florida, Gainesville, FL, United States, 2Oregon Health & Science University, Portland, OR, United States
Background: Duchenne and Becker muscular dystrophies (DMD/BMD) cause progressive muscle degeneration and fat replacement. While lower limbs are widely studied, upper extremity (UE) involvement in relation to mobility status remains underexplored. The deltoid (DEL), biceps brachii (BB), and triceps brachii (TB) are essential for UE function, quantifying their fat infiltration may provide alternative endpoints, especially in non-ambulatory individuals who are often excluded from clinical trials.
Aims: This cross-sectional study aimed to assess and compare muscle fat fraction (FF) in key upper arm muscles (DEL, BB, TB) across functional subgroups: early ambulatory, late ambulatory, and non-ambulatory, in individuals with DMD/BMD.
Methods/Patients/Materials: A total of 111 individuals (DMD: n=54, 4-21 years; BMD: n=57, 18-62 years) underwent localized 8-point and whole-body 3-point Dixon MRI (WBI) to measure FF in the DEL, BB, and TB muscles. Mobility status was categorized based on the ability to rise from the floor and complete the 10-meter walk/run test. Group differences were analyzed using Kruskal Wallis and Dunn's tests, while Spearman correlation and Bland-Altman analysis evaluated agreement between MRI methods.
Results: The two MRI methods showed strong correlation (Spearman's rho = 0.889) and minimal bias (Bland-Altman bias = 0.014). In DMD, FF in the UE remained relatively stable between early (DEL=0.15; BB=0.17; TB=0.15) and late ambulatory stages (DEL=0.22; BB=0.21; TB=0.20), with significantly higher levels of FF (DEL=0.51; BB=0.40; TB=0.51; p<0.05) observed only after loss of ambulation. In contrast, in BMD, FF increased significantly between early (DEL=0.13; BB=0.10; TB=0.12) and late ambulatory stages (DEL=0.20; BB=0.42; TB=0.33; p<0.05) before loss of ambulation. In the late ambulatory stage, both the BB and TB (p<0.05) muscle FF were higher in BMD compared to DMD.
Conclusion: In DMD/BMD UE muscle FF increases as mobility status worsens, yet the increase in muscle FF is observed earlier in the ambulatory functional decline in BMD compared to DMD. This highlights the role of using UE MRI biomarkers in selection and stratification of participants, potentially improving trial design and inclusivity in DMD/BMD.
O3.4 Rethinking diaphragm ultrasound: diaphragm thickening reflects lung volume not contractile effort
Jeroen van Doorn1, Joris van Doremalen1,
Chris de Korte2,3, Nicol Voermans4, Coen Ottenheijm5, Jonne Doorduin6
1Department of Neurology, Clinical Neuromuscular Imaging Group, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, Netherlands, 2Medical Ultrasound Imaging Centre, Department of Medical Imaging, Radboud University Medical Center, Nijmegen, Netherlands, 3Physics of Fluids Group, TechMed Centre, University of Twente, Enschede, Netherlands, 4Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, Netherlands, 5Department of Physiology, Amsterdam UMC, Location VUmc, Amsterdam, Netherlands, 6Department of Intensive Care Medicine, Radboud University Medical Center, Nijmegen, Netherlands
Background: Diaphragm ultrasound is increasingly popular to assess respiratory function in neuromuscular disorders. Specifically, the diaphragm thickening ratio (DTR) is often considered to reflect diaphragm contractility, although recent studies challenge this approach. The precise relation between DTR, lung volume, and diaphragm contractility is important to appreciate the value of diaphragm ultrasound in research and clinical practice. Especially in young children with a neuromuscular disorder, or patients with cognitive impairments, diaphragm ultrasound may provide extra information that is not obtainable with current respiratory function testing.
Aims: To elucidate the complex relationship between DTR, lung volume and diaphragm contractility.
Methods: Two experimental protocol series were performed in healthy participants. During series 1, participants performed slow deep inspirations and isometric maximal voluntary contractions, allowing measurement of DTR during a change in either lung volume or loading. During series 2, participants breathed through a tapered flow resistance device set at 10% to 50% of maximal inspiratory pressure, enabling simultaneous measurement of DTR during changes in lung volume and loading. DTR was measured at the zone of apposition at multiple timepoints during inspiration for all manoeuvres and analysed using mixed models.
Results: Series 1 included 20 participants; series 2 included 21, with comparable demographics and pulmonary function. In series 1, DTR significantly increased during deep inspiration (from 0.85 (95%CI: 0.74-0.97) to 1.98 (95%CI: 1.87-2.10), p<0.001) while not during maximal voluntary isometric contraction (from 1.07 (95%CI: 0.96-1.17) to 1.02 (95%CI: 0.90-1.14), p=0.944). In series 2, DTR was significantly increased with lung volume (from 1.09 (95%CI: 0.98-1.21) to 2.96 (95%CI: 2.72-3.20), p<0.001) and decreased during 50% loading compared to 10% loading (from 1.60 (95%CI: 1.50-1.70) at 50% of inspiratory volume to 1.52 (95%CI: 1.42-1.61), p<0.001).
Conclusion: DTR was not associated with changes in contractility but was associated with changes in lung volume in series 1. This was confirmed by series 2: no relevant association between DTR and contractile force was found. DTR should not be used as a surrogate measure for diaphragm contractility in clinical practice or research.
O3.5 Bilateral analysis of upper limb endpoints in ambulant and non-ambulant Duchenne muscular dystrophy patients
Michel Michaëls1,2, Susanna Rauh3, Evelien Fleerakkers1,2, Candace Moore1,2, Menno van der Holst2,4,
Erik van Zwet5, Erik Niks1,2, Hermien Kan2,3
1Department of Neurology, Leiden University Medical Center, Leiden, Netherlands, 2Duchenne Center Netherlands, Netherlands, 3C J Gorter MRI Center, Leiden University Medical Center, Leiden, Netherlands, 4Department of Rehabilitation and Physiotherapy, Leiden University Medical Center, Leiden, Netherlands, 5Department of Statistics, Leiden University Medical Center, Leiden, Netherlands
Background: In Duchenne muscular dystrophy (DMD), systemic therapeutical approaches face challenges in tissue exposure. Intramuscular dosing may increase tissue concentrations, limit systemic adverse effects, and be relevant to advanced stages of the disease, especially when targeting the upper limb. However, it is unclear if DMD progression is symmetrical and if an untreated arm can serve as an internal control.
Aim: We aim to compare disease progression between the dominant and non-dominant arm in DMD to explore the possibilities of clinical trials using a self-controlled design in intramuscular interventions.
Methods: Upper arm flexor and extensor muscle fat fractions (mFF) and contractile volumes (CV) were determined bilaterally using quantitative MRI (3T, gradient-echo Dixon, 6 echos, TR=15ms, TE=1.60ms, ΔTE=1.40ms, FA=5o, 5mm slice thickness (no gap), center slice defined at 40% humerus length, 13 slices around center slice), offline fat-water reconstruction and manual segmentation. Elbow flexion and extension strength were assessed bilaterally using MicroFET2 (isometric) and Biodex Pro 4 (isokinetic), along with upper limb function using the PUL2.0. Wilcoxon and paired t-tests assessed differences between both arms at baseline and over time.
Results: Twenty-two patients (ten non-ambulant) were assessed at baseline, median age 9.2y, range 7.0–17.0, and ten at 12 month follow-up. Eleven patients underwent MRI at baseline, and four at 12 months, with seven expected to do so before November 2025. At baseline, elbow flexors mFF (16.4±13.4 vs 12.9±13.1%; p=0.04), flexor CV (29.3±10.4 vs 26.7±9.6cm3, p=0.03) and PUL2.0 (36.5±11 vs 35.0±13.5, p=0.003) were different between the dominant and non-dominant arm. Differences between the rate of change over time in the dominant and non-dominant arm were so far only tested for functional measures, and did not show significant differences: PUL2.0 score (–2.0±4.3 vs. –1.0±3.3; p=0.09), isometric flexion (–1.9±6.1 vs. –3.9±10.9N; p=0.87), extension (–8.3±12.7 vs. –5.9±13.6N; p=0.67), isokinetic flexion (–0.5±2.5 vs. –0.5±2.5Nm; p=0.63) and extension (–1.6±0.7 vs. –1.6±0.7Nm; p=1.0).
Conclusion: Our results strengthen clinical observations that differences in DMD progression between dominant and non-dominant upper limbs are limited. This paves the way for clinical trials using unilateral and/or local intramuscular interventions with a patient as his own control design.
O4.1 Non-invasive MRI monitoring of glycogen accumulation in a mouse model of Pompe disease
Nirbhay Yadav1,2, Derek Timm3, Tyler Johnson3, Nickita Mehta3, Brian Fox3, Yuguo Li1,2, Qing Zeng1,2
1Russell H Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, MD, United States, 2F M Kirby Research Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, MD, United States, 3Amicus Therapeutics, Inc., Princeton, NJ, United States
Background: Pompe disease is a rare, inherited, multisystemic disorder caused by functional deficiency of the lysosomal enzyme acid α-glucosidase (GAA), responsible for breaking down glycogen into glucose. Glycogen accumulation, particularly in muscle cells, leads to progressive cardiac, motor and respiratory dysfunction. Enzyme replacement therapy (ERT) with recombinant human GAA aims to slow disease progression and is the only approved treatment modality for Pompe disease. Current clinical methods for tracking the pharmacological effects of ERT in clearing skeletal muscle glycogen, such as biochemical glycogen analysis in muscle biopsies, are invasive. Glycogen is magnetically coupled with water and can be imaged with high sensitivity via the relayed nuclear Overhauser effect (NOE).
Aims: We investigated whether glycoNOE, a non-invasive, MRI-based glycogen quantification method, can detect differences in glycogen levels in a preclinical mouse model of Pompe disease (Gaa knockout mice) compared with wild-type mice and in Gaa knockout mice treated with ERT compared with placebo.
Methods/Patients/Materials: Mice received four biweekly treatments with placebo (Gaa knockout and wild-type mice) or ERT (Gaa knockout mice only). Results from glycoNOE MRI glycogen analysis were compared with biochemical glycogen analysis.
Results: GlycoNOE MRI showed differences between placebo-treated Gaa knockout mice and wild-type controls and between Gaa knockout mice treated with placebo and those treated with ERT. The correlation between glycoNOE MRI and biochemical glycogen analysis was statistically significant.
Conclusion: Results highlight the opportunity to test the utility of glycoNOE MRI in a clinical setting to fill an unmet need in identifying non-invasive methods to quantify glycogen, a direct measure of the efficacy of treatments for patients with Pompe disease. Supported by Amicus Therapeutics, Inc.
O4.3 Comprehensive muscle tissue evaluation via whole-body MR Fingerprinting
Constantin Slioussarenko1, Pierre-Yves Baudin2,
Marc Lapert2, Yves Fromes1, Benjamin Marty1
1Institute of Myology, Paris, France, 2Siemens Healthcare SAS, Paris, France
Background: Neuromuscular diseases (NMD) affect skeletal muscles heterogeneously and can lead to important respiratory or cardiac impairments, making disease monitoring crucial yet challenging. Whole-body MRI has shown promise for NMD1, cancer, and metabolic diseases, but typically relies on qualitative scans. Quantitative multiparametric MRI offers richer information but faces barriers such as lengthy acquisition, motion artifacts, and manual segmentation needs.
Radial MR Fingerprinting (MRF) with fat/water separation (MRF T1-FF) enables simultaneous quantification of FF and water T1 (T1H2O), imaging biomarkers of muscle alterations2, accounting for B0/B1 inhomogeneities.
Aims: This work extends MRF T1-FF to a free-breathing, whole-body 3D protocol for NMD patients, enabling sensitive assessment of muscle alterations, supported by automated post-processing.
Methods: The protocol included four acquisition blocks. For upper body, a motion-corrected MRF T1-FF framework was used: a FLASH scan estimated motion, followed by an axial MRF scan for parametric mapping3. Axial acquisitions for the pelvis, thighs, and legs employed a modified MRF T1-FF, optimized via digital phantoms. Total scan time was around 25 minutes.
Post-processing involved: 1) Motion estimation via VoxelMorph4; 2) Motion-corrected image-series reconstruction; 3) Fast parametric mapping reconstruction via bicomponent dictionary matching; 4) Distortion correction using gradunwarp (https://github.com/Washington-University/gradunwarp); 5) Blocks combination; 6) Automatic muscle segmentation for lower body muscles using a custom nn-UNet5; 7) ROI-based FF and T1H2O extraction.
This was applied on 19 healthy volunteers on a 3T Siemens PrismaFit scanner (Erlangen, Germany).
Results: Whole-body parametric FF and T1H2O maps were generated for all subjects. ROI-level FF and T1H2O values in the lower body extracted from the automated segmentation were in line with the literature, with highest dispersion across subjects for FF in the smallest muscles such as Gracilis. Diaphragm consistent visibility suggested potential for respiratory muscle evaluation in patients.
Conclusion: We present a clinically feasible, free-breathing whole-body MRF protocol enabling automated FF and T1H2O mapping and muscle segmentation. Future work will extend segmentation to upper body muscles and apply the method to larger cohorts, supporting multi-systemic NMD evaluation.
References:
Shelly et al. 2010
Marty et al. 2021
Slioussarenko et al. 2024
Balakrishnan et al. 2019
Baudin et al. 2025
O4.4 Assessing isometric muscle strength using shape and fiber orientation models
Salim Bin Ghouth1, Valentina Mazzoli1
1Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, United States
Background: Muscle strength plays a critical role in physical health, and is negatively affected by age. This is not only due to loss of muscle mass, but also to geometrical remodeling of muscle fibers. These changes in muscle architecture collectively lead to a global shape change, and can be detected non-invasively using Diffusion Tensor Imaging (DTI). Due to its ability to quantify anatomical structures into a set of independent Principal Components (PCs), Statistical Shape Modeling (SSM) holds potential for studying muscle architectural features related to force production.
Aims: The goal of the study was to extend the SSM technique and develop a combined shape and fiber orientation model (SFM) of quadriceps muscle (Vastus Lateralis, VL, Rectus Femoris, RF, Vastus Medialis, VM, Vastus Intermedius, VI), to (1) analyze the 3-dimensional shape and architecture features and describe the geometrical remodeling of skeletal muscles due to aging, and (2) investigate association between PCs obtained from SFMs and isometric muscle strength.
Methods: Using water-only Dixon and DTI images, we developed SFMs of the quadriceps and assessed the association between the PCs obtained from SFMs against isometric muscle strength using a multilinear regression.
Results: The first ten PCs of the SFMs described more than 94.8% of the total variance within the population. The first and second PCs of the SFMs represented variations in muscle size and muscle length-width-thickness. Other PCs described varying features including expansion and compression of the proximal region of the muscles. These shape variations were accompanied by variations of fiber orientations at different regions of the muscles. The R2 of the regression were higher for SFMs (0.78, .67, 0.77, and 0.68) than when using only muscle volume (0.75, 0.64, 0.68 and 0.52 for VL, RF, VM and VI respectively).
Conclusion: SFMs captured shape and architecture features of quadriceps with a few PCs demonstrating robustness of the models. Regression models using SFMs led to better predictions of isometric strength compared to using muscle volume only. This research highlights the importance of muscle shape and architecture in muscle force production and offers new tools to study the decline of muscle strength in older adults.
O4.5 Open MRI Pipeline for muscle strain calculation
Marta Brigid Maggioni1,2, Sabine Melanie Räuber1,2, Francesco Santini1,2
1Basel Muscle MRI (BAMM), Department of Biomedical Engineering, University of Basel, Basel, Switzerland, 2Radiology, Division of Radiological Physics, University Hospital of Basel, Basel, Switzerland
Background: In this work, we present a vendor-agnostic workflow for measuring forearm muscle strain induced by NMES, covering all steps from accelerated CINE 4D-flow data acquisition to reconstruction and analysis with the goal of addressing challenges in quantitatively evaluating rare neuromuscular diseases and standardizing flow data processing. This process is often non-standardized and not openly available, limiting reproducibility and multi-center studies.
Methods: A 4D-flow sequence was implemented in PyPulseq (v1.4.3) using a 3D cartesian GRE acquisition, with CINE acquisition triggered by NMES-induced forearm contractions. Elliptical k-space trimming and Poisson disc undersampling (factor 9, with a fully sampled center) were applied. Complex data were reconstructed using coil sensitivity maps from BART's ESPIRiT algorithm. The protocol was tested on eight healthy subjects (age 24–31, 5 females, 3 males) on a 3T scanner (Prisma, Siemens Healthineers) during NMES-evoked contractions, with an MRI-compatible force sensor recording force.
Our analysis pipeline, which calculates strain tensors from velocity and displacement to assess muscle dynamics by computing the deformation gradient tensor, deriving the Eulerian strain tensor, and extracting its eigenvalues to represent principal strains was used. Strain rates are quantified by fitting time-varying strain curves with sigmoid functions.
Flow sequence parameters included: 27 phases, 2 lines/segment, TR=6.7 ms, TE=4.5 ms, FA=10°, 1.5×1.5×1.5mm3 resolution, venc=0.2 s, and a 5-minute acquisition time. Sequence and analysis tools are available at https://github.com/BAMMri/Pulseq-4DFlow and https://github.com/BAMMri/3D-Dynamic-Velocity.
Results: Undersampling and elliptical scanning reduced the acquisition time from 28 to 5 minutes. NMES-activated muscles are clearly distinguishable on the calculated displacement and strain eigenvalue maps. The build-up rate, determined via fitting, was 2.035 s−1 in the Flexor Digitalis Superficialis. Combining force sensor data with the strain curves showed that strain peaks before force, due to elastic properties in the tendon-muscle complex.
Discussion and Conclusion: This work presents a fully open-source 4D-flow sequence and analysis pipeline for extracting velocity, displacement, and strain from NMES-stimulated skeletal muscles. Future improvements will focus on including automated segmentation from anatomical scans and adopting standardized data formats like musclebids.
Poster Presentations
P1.01 Volume of fasciculation measured on diffusion-weighted MRI correlates with muscle weakness in older adults
Gabrielle Baxter1, Smita Rao2, Els Fieremans1,
Valentina Mazzoli1
1Department of Radiology, NYU Grossman School of Medicine, New York, United States, 2Department of Physical Therapy, NYU, New York, United States
Background: Increased motor unit size is believed to be a hallmark of sarcopenia and associated with lower force production in older adults. Fasciculation, the involuntary contraction due to the spontaneous discharge of an individual motor unit, can be detected in diffusion-weighted (DWI) MRI as areas of signal attenuation. Spatial and temporal characteristics of fasciculation in older adults using DWI could identify neuromuscular factor associated with force production.
Aims: 1) to investigate the relationship between fasciculation event size and detection chance with muscle strength in older adults and, 2) to identify the optimal Diffusion Time (Tm) for the detection and analysis of fasciculation events.
Methods: DWI was performed for the left thigh in 27 subjects (14 males, mean age 73.7) on a 3T scanner (MAGNETON Prisma, Siemens Healthineers). A diffusion-weighted stimulated-echo sequence was used with diffusion times (Tm=30, 60, 100, 200, 400, 700, 1000, 1200ms), TE=38ms, TR=4000, 4000, 4000, 4000, 4000, 4700, 6200, 7200ms, matrix size=24×72×10, voxel size=3.6×3.6×10mm3, NSA=1, δ=6ms, 12 diffusion-encoding directions, b=500s/mm2.
A fasciculation event was defined as a group >5 voxels where the Apparent Diffusion Coefficient (ADC) was 1.5x the average muscle ADC for each Tm. Event volume (% of affected muscle volume, averaged over all volumes) and detection chance (volumes with at least one event/number of volumes) were defined per muscle per Tm.
Isometric strength was measured during knee extension/flexion using an isokinetic dynamometer (Biodex).
Correlations between volume and detection chance and strength were tested using Spearman's correlation coefficient, with p<0.05 indicating statistical significance.
Results: Volume of fasciculation events increased with Tm (r=0.56,p=0.16). Detection chances followed a non-monotonic function of Tm with a peak at 200ms (85.4%). Significant negative correlations (r=-0.45 to -0.67, p<0.05) were found between fasciculation volume and muscle strength for all muscles when Tm>=700ms, suggesting sensitivity to motor unit size. Detection chance and muscle strength were not significantly correlated.
Conclusion: The highest detection chance was observed when Tm was similar to fasciculation length (200ms) and decreased with longer Tm. However, our results suggests that DWI at longer Tm > 700ms could be a promising tool to investigate neuromuscular determinants of strength in older adults.
P1.02 Intraepineurial fat fraction: A novel MR Neurography-based biomarker in Transthyretin amyloidosis polyneuropathy
Eva Sole-Cruz2, Etienne Fortanier1,2,
Constance. P Michel1, Emilien Delmont2,
Annie Verschueren2, Marc-Adrien Hostin1,
Shahram Attarian2, David Bendahan
1
1CRMBM, Marseille, France, 2Referral Center for neuromuscular diseases and ALS, Marseille, France
Background: Hereditary Transthyretin amyloid polyneuropathy (ATTRv-PN) is a rare and progressive neurodegenerative disorder characterized by axonal neuropathy and amyloid deposits. Early detection of disease onset and progression is crucial for timely therapeutic intervention. Quantitative MRI (qMRI) can be used to measure potential biomarkers. While FF is widely used in muscle imaging, its application to peripheral nerves is rare. Intraepineurial fat fraction (ieFF) may reflect lipid droplets in amyloid deposits as described in histological studies or the replacement of nerve fiber loss with fatty rich interfascicular epineurium.
Aims: This study investigates the potential utility of ieFF as a novel imaging-related biomarker in differentiating ATTRv-PN, asymptomatic carriers (ATTRv-C) and healthy controls (HC).
Methods: 53 patients with TTR mutations were imaged (31 ATTRv-PN patients, 22 ATTRv-C, and 24 HC) and both clinical and electrophysiological parameters were quantified. Clinical scores included ONLS, RODS, NIS-LL, MRC; electrophysiology included CMAP, MUNIX, SNAP. 3D Volume, ieFF and Magnetization Transfer Ratio (MTR) were quantified in sciatic and tibial nerves using qMRI. Univariate and multivariate models (adjusted for age and BMI) were used to assess the diagnostic and prognostic value of ieFF
Results: Symptomatic ATTRv-PN patients exhibited significantly higher ieFF in both sciatic (32.4% IQR [24.4-38.1]) and tibial nerves (13.7%, IQR [9.97-20.7]) compared to controls (sciatic 22.3%, IQR [16.6-28.5]; tibial 9.74%, IQR [6.36-12.5]) (p<0.05). ieFF values were positively correlated in both uni and multivariate analysis with the main clinical scores and electrophysiological measures. Asymptomatic carriers also showed increased ieFF values compared to controls (p<0.05). Comparatively, MTR and nerve volumes exhibited less pronounced differences across groups.
Conclusion: This study demonstrates that ieFF effectively differentiates symptomatic and asymptomatic ATTRv patients from HC and correlates strongly with electrophysiological and clinical severity parameters. Furthermore, we compare ieFF with conventional qMRI biomarkers highlighting its superior potential for monitoring nerve structural impairment.
P1.03 Deciphering of skeletal muscle involvement in cystinosis with whole-body muscle MRI
Edouard Berling1,2, Johnny Ye1, Aude Servais3,4,
Hélène Prigent1,2, Alice Rouyer1,2, Clément Guémy1, Guillaume Nicolas1,2, Nadia Venturelli1, Pascal Laforêt1,2, Robert Carlier1,2
1Raymond Poincaré University Hospital, AP-HP, Garches, France, 2Paris Saclay University, Garches, France, 3Necker Enfants Malades University Hospital, AP-HP, Paris, France, 4Imagine Institute, Paris, France
Background: Cystinosis, a rare autosomal recessive metabolic disorder caused by mutations in the CTNS gene, is characterized by abnormal cystine accumulation leading to the formation of cystine crystals in various organs, including skeletal muscles.
Aims: This study aims to depict muscle involvement patterns in cystinosis patients using whole-body muscle MRI (WB-MRI).
Methods: Adult patients with genetically confirmed cystinosis and WB-MRI scans were included in a retrospective cohort. The examinations consisted of axial T1 and T2 fat saturated -weighted images (T2 DIXON technique) and were scored for 47 muscles using Mercuri's score (MS). Patients were categorized into two groups based on the presence or absence of significant MRI involvement, defined as a mean MS >1.1. Clinical data, including age, treatment initiation, plasma cysteamine levels, and muscle strength, were collected.
Results: Nineteen patients were included, with a mean age of 33.6 years. Motor deficits were observed in 15 (79%) of them. All patients were treated with cysteamine and had undergone kidney transplantation. Eleven patients (58%) exhibited respiratory failure, and five (26%) had swallowing disorders.
Eight patients (42%) exhibited significant muscle abnormalities. The most frequently affected muscles were the tongue, trapezius, deltoids, forearm extensors, soleus, and gastrocnemius muscles. No bright signal on T2 fat saturated images was detected.
The groups with and without abnormalities did not significantly differ in terms of age (31.1 vs. 37.1 years, p=0.057) or age at cysteamine initiation (2 vs. 7.6 years, p=0.118). However, the group with abnormalities had a lower vital capacity (53.9% vs. 86.7% of the predicted value, p=0.001) and a higher prevalence of swallowing disorders (62.5% vs. 0%, p=0.0048).
Conclusion: This study highlights a specific pattern of muscle involvement in cystinosis, with a predilection for the tongue, shoulder girdle, and distal muscles of the upper and lower limbs. Severe muscle impairments, including respiratory involvement and swallowing disorders, were correlated with the severity observed on MRI. This suggests that WB-MRI could be considered as a tool for early detection of these complications.
P1.04 Validation of Tractography-Derived Muscle Architecture in the rotator cuff using a fresh cadaveric pig model
David Berry1, Robert Fisch2, Joseph A. Gordon III1, Adam Lane1, Vitaly Galinsky1, Lawrence Frank1, Samuel R Ward1
1University of California, San Diego, La Jolla, United States, 2University of Nevada, Reno, Reno, United States
Background: Skeletal muscle architecture - including fiber length and pennation angle - is predictive of muscle function. Diffusion tensor imaging (DTI) with tractography offers a noninvasive means of estimating these features in vivo. Prior validation has focused on small, simple muscles in the mouse or rabbit leg, or frozen human leg tissue. These models lack the size and complexity of large, multi-pennate muscles. The pig shoulder, with its complex rotator cuff musculature, is a valuable surrogate for studying human shoulder muscle architecture.
An advantage of tractography is its ability to provide fiber architecture distributions across the entire muscle (thousands of fibers). Manual dissection (the gold standard) is limited by sparse sampling, typically only a few measurements per muscle. Tractography enables high-resolution, volumetric mapping and must be validated in freshly harvested, architecturally complex muscle to support its clinical translation.
Aim: To validate tractography-derived measurements of fiber length and pennation angle against manual dissection in pig rotator cuff muscles and assess the distribution of architectural features across methods.
Methods: Eight freshly sacrificed Yorkshire pig shoulders (6 months) were imaged using a Siemens 3T Prisma MRI scanner with a shoulder coil. Each shoulder was scanned using Dixon fat-water separation (1×1×1mm3) and spin-echo DTI (2×2×4mm3, 30directions, b=500 s/mm2). The supraspinatus (SS), infraspinatus (IS), and subscapularis (SubS) muscles were segmented using Horos. Tractography was performed with MRtrix3. Fiber length and pennation angle histograms were generated. Each muscle was then dissected, and five manual measurements of architecture were recorded.
Results: Tractography-derived histograms revealed distinct muscle-specific profiles. Tractography occasionally underestimated long fibers and showed sampling variability in extreme pennation angles. The mean±standard deviation and range of fiber lengths and pennation angles is reported below:
Dissection
SS: 83.2±36.3mm (41.5–172.3), 23.6°±13.1° (0–57)
IS: 62.5±57.8mm (23.5–199.2), 25.7°±9.6° (8–40)
SubS: 46.0±36.5mm (17.4–133.4), 16.1°±9.6° (0–30)
Tractography
SS: 69.0±40.9mm (12.8–250.6), 20.0°±13.9° (0.3–89.9)
IS: 99.7±49.3mm (12.8–250.8), 18.0°±15.3° (0.3–89.9)
SubS: 41.6±20.3mm (12.8–166.7), 17.9°±15.4° (0.3–9.9)
Conclusion: Tractography and dissection show strong agreement in quantifying muscle architecture in pig shoulders. Tractography enables comprehensive fiber distribution assessment. Future work will focus on spatially matching tractography and dissection regions for improved regional validation.
P1.05 From histology to simulation: open-source muscle phantoms for diffusion MRI modeling
David Berry1, Vitaly Galinsky1, Jascha Gaardner1,
Samuel Ward1, Lawrence Frank1
1University Of California, San Diego, La Jolla, United States
Background: The relationships between tissue microstructure, diffusion magnetic resonance imaging (dMRI) sequences and parameters, and the resulting signal are complex and must be systematically characterized for successful clinical translation. In silico modeling provides a powerful framework for investigating these relationships, enabling precise control over fiber geometry, diffusivity, permeability, noise, and pulse sequence parameters. However, most simulations rely on idealized geometries rather than anatomically accurate models derived from histology.
Aims: To create an open-source repository of digital phantoms derived from real muscle histology for use with the open-source dMRI simulation platform DifSim.
Methods: Laminin-stained histology of tibialis anterior muscles from rats subjected to control, cardiotoxin, botulinum toxin (botox), surgical denervation, or tenotomy injuries at 1, 3, 7, 14, and 30 days post-injury were used. Each injury induces distinct patterns of degeneration, regeneration, hypertrophy, or atrophy. Average muscle fiber diameters and surface-to-volume ratios (S/V) were recorded.
A custom MATLAB pipeline binarized and denoised each histological image before extrusion into 3D voxel grids. These were converted to STL files and optimized in Blender, with decimation (90%) and mesh integrity checks (manifold, watertight, outward facing normals). Non-manifold elements were corrected using vertex merging tools. Each mask was saved with a unique identifier for downstream analysis.
Results: Out of 10,944 images, 1,626 high-quality digital phantoms were generated. Mean fiber diameters ranged from 37.7–131.2 μm, and areas from 809.3–8,174 μm2. Across the dataset, the mean ± SD fiber diameter was 78.2 ± 17.3 μm, and area was 3,174 ± 1,323 μm2. The phantoms are saved as both .stl and .mdl files. This dataset captures a wide range of microstructural phenotypes, enabling simulation studies across physiologic and pathologic states.
Conclusion: This study introduces a reproducible pipeline for generating anatomically accurate digital muscle phantoms from histology, facilitating in silico studies of dMRI signal behavior. The phantoms, underlying images, and associated software are freely available via the UCSD Libraries, supporting broader adoption of simulation-based diffusion modeling.
P1.06 pre-and post-skeletal muscle biopsy quantitative magnetic resonance imaging reveals correlations with histopathological findings
Alice De Lorenzo1, Anne-Katrin Güttsches1,2,
Johannes Forsting1, Moritz Kneifel1,2,
Robert Rehmann1,2, Elena Enax-Krumova1,
Martijn Froeling4, Matthias Vorgerd1,2,
Lara Schlaffke1,2,3
1BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Heimer Institute for Muscle Research, Bochum, Germany, 3Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 4Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Skeletal muscle biopsy is the diagnostic gold standard in neuromuscular disorders (NMDs), but its invasive nature limits its use in follow-up. Quantitative muscle MRI (qMRI) offers a non-invasive alternative, though evidence for correlation with histopathology remains sparse.
Aim: To validate and extend prior findings by investigating the correlation between qMRI parameters and histopathological features in skeletal muscle biopsies, using pre- and post-biopsy MRI to precisely co-localize imaging data with the site of tissue sampling.
Methods: This prospective study included 26 patients (mean age 46.4 ± 15.1 years) undergoing skeletal muscle biopsy for clinical diagnosis. qMRI scans were conducted within 72 hours prior to, and within 24 hours following, biopsy. Parameters including fat fraction (FF), water T2 relaxation time, and diffusion metrics (fractional anisotropy [FA], mean diffusivity [MD], axial and radial diffusivity), were extracted from the precise biopsy location, a bigger sphere and the whole muscle and compared with histopathological data.
Results: FF and water T2 relaxation time correlated significantly with histological fat content and inflammatory markers (MHCI expression, CD3-positive T lymphocytes, and CD68-positive macrophages), respectively. FA correlated with the proportion of type 2 fibres, while MD negatively correlated with the autophagic marker p62. No significant variability was observed across segmentation methods, indicating the robustness of qMRI measurements.
Conclusion: qMRI metrics, particularly FF and T2, reliably correlate with key histopathological findings, supporting their utility as non-invasive biomarkers in NMD follow-up. The robustness of qMRI metrics across various segmentation strategies supports its clinical applicability, including in settings where post-biopsy MRI cannot be obtained. Whole muscle analysis revealed similar outcomes as compared to exact biopsy localization, making the post biopsy scan redundant.
P1.08 MYO-RESO: Quantitative muscle MRI as biomarker of muscle involvement in myotonic dystrophy type I
Sebastian Ariel Figueroa Bonaparte1,2,
Jacob Sanchez Dalmau6, Eduard Juanola Mayos1,2,
Alicia Martinez Piñeiro1,2, Anna Martinez Vigueras1, María Mercedes Molleda5, Giuseppe Lucente1,2,
Miriam Almendrote1,2, Maria de la Iglesia Vaya4,
Gemma Montè Rubio3
1Neuromuscular Disorders Unit - Neurology Service -Germans Trias I Pujol University Hospital, Badalona, Spain, 2Germans Trias i Pujol Research Institut (IGTP), Badalona, Spain, 3Center for Comparative Medicine and Bioimaging of Catalonia(CMCiB, IGTP), Badalona, Spain, 4 Joint Unit of Biomedical Imaging and Artificial Intelligence (FISABIO-CIPF), Valencia, Spain, 5Physical Medicine and Rehabilitation Service - Germans Trias i Pujol university Hospital, Badalona, Spain, 6Radiology Service - Germans Trias I Pujol university Hospital, Badalona, Spain
Background: Muscle weakness affects people with myotonic dystrophy type I (MD), leading to wheelchair use in 50% of patients. Quality of life is severely compromised by this condition, and the ability to maintain autonomy is severely affected. Currently, there are no reliable biomarkers for monitoring muscle weakness in MD. In fact, the two scales regularly used to measure muscle strength, the 6MWT and the MRC Scale, are highly inaccurate.
Aims: Objective 1: To assess the progression of muscle involvement over a three-year period using quantitative muscle MRI sequences in 35 patients with MD. Objective 2: To compare radiological progression with muscle tests used as standard of care. Objective 3: To develop an automated muscle segmentation system for quantifying fat replacement in muscle MRI images.
Methods: A cohort of 35 patients will be followed. Whole-body muscle MRI will be performed annually over a 3-year period, using both qualitative and quantitative sequences. Demographic data, muscle strength scores, quality of life, cardiac and respiratory function, and blood tests will also be prospectively collected. Dixon and water T2 sequences will be implemented. In turn, as muscle is a three-dimensional structure, the ability to analyze the same section over time is limited. Therefore, a significant focus of the project is to develop innovative automatic muscle segmentation techniques using AI to facilitate image analysis and reduce inter-observer variability in the quantification of muscle involvement.
Results: This study will start on September 2025
Conclusion: Quantitative muscle MRI could be a useful biomarker of disease progression. Automatic analysis software for quantitative muscle MRI images could improve the measurement of disease progression.
P1.09 Quantitative muscle MRI in inclusion body myositis (IBM): A prospective cohort study
Johannes Forsting1, Robert Rehmann1,2, Matijn Froeling3, Anne-Katrin Güttsches1,2, Elena Enax-Krimova1,
Lara Schlaffke1,2
1BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Heimer-Institute for Muscle Research, Bochum, Germany, 3Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Inclusion body myositis (IBM) is the most common idiopathic inflammatory myopathy in individuals over 50. Given the failure of numerous drug trials, limitations of traditional clinical assessments have been discussed, emphasizing the need for more sensitive outcome parameters. Quantitative MRI (qMRI) has emerged as a promising tool, detecting early muscle changes even before fat infiltration, which could aid treatment decisions. However, the underlying pathophysiological processes remain unclear, particularly in diffusion imaging. Diffusion parameters reflect water molecule movement, providing insights into muscle micro- and macrostructure.
Aims: This study aims to investigate the potential of qMRI as a biomarker for disease assessment in IBM, addressing differences from healthy controls, correlations between qMRI parameters and clinical assessments in IBM and the longitudinal evaluation of qMRI parameters to monitor disease progression.
Methods: Twelve IBM patients (5 females, mean age 69.6, BMI 27.8) and 12 age- and sex-matched healthy controls were examined. Seven patients and corresponding controls were re-evaluated after one year. Clinical assessments included muscle strength (Quick Motor Function Measure, QMFM), functional limitations (IBM Functional Rating Scale, IBM-FRS), and gait analysis (Six-Minute Walk Test). MRI scans of the lower limbs were performed using a 3T scanner, including a Dixon-based sequence for fat fraction (FF), T2 mapping for water T2 relaxation time, and diffusion tensor imaging (DTI) for fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (λ1), and radial diffusivity (RD).
Results: MANOVA revealed significant differences between IBM and controls in all qMRI parameters (p<0.001). In non-fat replaced muscles (FF<10%), IBM patients showed increased water T2 and FA with decreased MD, λ1, and RD (p≤0.020). The strongest correlation with clinical assessments was found for water T2 in thigh muscles (r≤-0.634). Longitudinally, FF (+3.0%), water T2 (+0.6 ms), MD (-0.04×10−3 mm2/s), λ1 (-0.05×10−3 mm2/s), and RD (-0.03×10−3 mm2/s) significantly changed in IBM (p≤0.001), while FA remained unchanged (p=0.242).
Conclusion: qMRI parameters correlate with clinical findings, are sensitive to long-term changes, and may reflect distinct pathophysiological mechanisms. Water T2 appears to be a promising disease activity marker in IBM. Observed diffusion changes over time may indicate intracellular aggregate accumulation, requiring further investigation.
P1.10 Evaluation of cell diameter distribution in a cross-sectional cohort and its correlation with muscle force.
Martijn Froeling1, Roosmarijn Brenninkmeijer2,
Danny R. van der Woude2, Bart Bartels2,
Linda Heskamp1
1Center for Image Sciences, Department of Imaging and Oncology, University Medical Center Utrecht, Utrecht, Netherlands, 2Child Development and Exercise Center, Wilhelmina Children's Hospital, University Medical Center Utrecht, Utrecht, Netherlands
Background: The Random Permeable Barrier Model (RBPM) measures how semi-permeable barriers slow diffusion and then quantifies average cell size from that slowdown. By fitting the frequency-dependent diffusion rate, one can estimate typical cell dimensions. Although the RBPM has been applied to muscle injury and disease, its behaviour in healthy populations remains poorly characterized.
Aims: The aim of this study is to quantify the distribution of muscle cell diameters across different muscle groups in a cross-sectional cohort and to investigate their relationships with(i) biometric and DTI-derived parameters and (ii) measured torque output.
Methods: Data was acquired in 89 subjects (43 Male; age: 35.5 ± 12.9 (16.0 to 60.0)) in two slices in the thigh and two slices in the calf. Stimulated echo EPI DTI was collected with 15 diffusion directions (b = 200 (3x); 500(12x) s/mm2) and four mixing times (20, 100, 300 and 700 ms). Low b-values were omitted to minimized IVIM effects in the radial diffusivity (RD). Other parameters: parameters: voxel size: 12×5×5 mm3; slice gap: 48mm; TE/TR: 42/2000ms; SENSE: 2.4. Data was processed and fitted using QMRITools for Mathematica (qmritools.com). We quantified characteristic cell diameter in six muscle regions: Quadriceps, Adductors, and Hamstrings in the thigh; and the anterior, deep, and superficial compartments of the calf.
Results: Characteristic cell diameter was significantly larger in males than in females and differed across most muscle regions. In our healthy cohort, it correlated weakly with age, weight, height, and BMI and correlated strongly with fractional anisotropy (FA) in all muscles at every mixing time. It also showed moderate correlations with isometric and isokinetic flexion and extension torques; however, these associations became non-significant after adjusting for sex.
Conclusion: In healthy muscle, RBPM-derived cell diameters align closely with DTI-derived fractional anisotropy, suggesting RBPM offers little additional insight into microstructure beyond what FA already provides in healthy subjects. Correlations with demographic and torque measures remain weak or sex-dependent, further limiting its added value in a healthy cohort. Future studies should explore RBPM's other parameters and potential advantages in pathological or highly heterogeneous tissues.
P1.11 Effect of gradient non-linearity correction on whole leg Diffusion Tensor Imaging
Martijn Froeling1, Linda Heskamp1
1Center for Image Sciences, Department of Imaging and Oncology, University Medical Center Utrecht, Utrecht, Netherlands
Background: MRI gradient fields are only linear near the scanner's isocenter. Further away, gradient non-linearities cause image distortions and spatially varying diffusion weighting. While image warping is typically corrected on the scanner, diffusion-weighted imaging (DWI) remains affected, potentially leading to errors in in the parameter estimation and tractography, as first shown by Bammer et al. (2003). In whole-leg imaging, large fields of view place much of the anatomy far from the isocenter, amplifying these effects.
Aims: To quantify the impact of gradient non-linearity correction on DTI parameters and fiber tractography.
Methods: We acquired bilateral lower limb DTI in 101 healthy participants, covering hip to ankle. The protocol included 30 diffusion weightings (b-values: 1 [3x], 20 [3x], 50 [3x], 200 [6x], 500 [15x] s/mm2), triple fat suppression, SENSE factor 2.4, and voxel size 3×3×6 mm3. The total field of view was 480×267×186 mm3, placing the outermost parts of the legs up to 25cm from the isocenter.
Processing included denoising, registration, and DTI fitting with REKINDLE outlier rejection and iWLLS. Muscle-specific MD and FA were extracted using automatic whole-leg segmentation. Gradient non-linearity was corrected by modelling the gradient fields with Legendre polynomials to compute voxel-wise b-matrices based on scan geometry. Diffusion tensors were estimated using both nominal and corrected b-matrices for comparison.
Results: Within typical whole-leg datasets, the effective b-values varied across the legs, with an average of 510±24 s/mm2 (range: 450–654 s/mm2). Gradient non-linearity correction had a variable effect across muscles. The strongest effect was observed for MD which decreased from 1.49±0.04 (1.39–1.54) to 1.43±0.05 (1.27–1.47) 10−3 mm2/s, a reduction of −0.07±0.03 (−0.12 to −0.01), or approximately 5%. FA remained largely unchanged (before: 0.21±0.02, after: 0.21±0.02), though individual muscles showed changes up to 0.015, also representing ∼5% variation. For tractography average tract angles changed by 2.33±2.58° (range: 0.00–17.05°), and whole-muscle tract length increased by nearly 10%, from 94.34±25.46mm (66.44–154.11) to 105.59±30.74mm (77.41–175.73).
Conclusion: Gradient non-linearity correction is essential for accurate MD and FA quantification and reliable muscle fiber tractography in whole-leg DTI. Without correction, MD is systematically overestimated with increasing distance from the isocenter, potentially explaining previously reported along-muscle variations.
P1.13 Muscle diffusion tensor imaging in Late-Onset Pompe Disease
Giulia Guicciardi1,2, Leonardo Barzaghi2,
Matteo Paoletti2, Tiziana Enrica Mongini4,
Serena Gasperini5, Massimiliano Filosto6,7,
Lorenzo Maggi8, Annalisa Sechi9, Marina Grandis10,11, Michele Sacchini12, Monica Sciacco13,
Chiara Bonizzoni2, Niels Bergsland14,15,
Francesco Santini16,17, Xeni Deligianni16,17,
Silvia Nicolosi2, Michele Giovanni Croce3,
Claudia Angela Michela Gandini Wheeler-Kingshott19,3,20, Sabrina Ravaglia18, Anna Pichiecchio3,2
1Department of Mathematics, University of Pavia, Pavia, Italy, 2Advanced Imaging and Artificial Intelligence Center, Neuroradiology Department, IRCCS Mondino Foundation, Pavia, Italy, 3Department of Brain and Behavioural Sciences, University of Pavia, Pavia, Italy, 4Neuromuscular Unit, Department of Neurosciences, University of Turin, Turin, Italy, 5Department of Pediatrics, Università degli Studi Milano-Bicocca, Fondazione MBBM, San Gerardo Hospital, Monza, Italy, 6Department of Clinical and Experimental Sciences, University of Brescia, Brescia, Italy, 7NeMO-Brescia Clinical Center for Neuromuscular Diseases, Brescia, Italy, 8Neuroimmunology and Neuromuscular Diseases Unit, Fondazione IRCCS Istituto Neurologico ‘Carlo Besta’, Milan, Italy, 9Regional Coordinator Center for Rare Diseases, Academic Hospital of Udine, Udine, Italy, 10University of Genoa, Genoa, Italy, 11IRCCS Ospedale Policlinico San Martino, Genoa, Italy, 12Unit of Hereditary Metabolic and Muscular Disorders, Meyer Children's University Hospital, Firenze, Italy, 13IRCCS Fondazione Ca’ Granda Ospedale Maggiore Policlinico, Neuromuscular and Rare Disease Unit, Milan, Italy, 14Department of Neurology, Jacobs School of Medicine and Biomedical Sciences, Buffalo Neuroimaging Analysis Center, University at Buffalo, The State University of New York, Buffalo, NY, United States, 15IRCCS, Fondazione Don Carlo Gnocchi Onlus, Milan, Italy, 16Basel Muscle MRI, Department of Biomedical Engineering, University of Basel, Basel, Switzerland, 17Department of Radiology, University Hospital Basel, Basel, Switzerland, 18IRCCS C. Mondino Foundation, Pavia, Italy, 19Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, Faculty of Brain Sciences, University College London, London, United Kingdom, 20Brain Connectivity Center, C. Mondino National Neurological Institute, Pavia, Italy
Background: Muscle diffusion tensor imaging (mDTI) is a valuable tool to provide indirect information about muscular microstructure, being able to reflect directional changes in water molecule diffusion. mDTI metrics have been proposed as promising quantitative parameters for detecting disease involvement in several myopathies, including Duchenne muscular dystrophy, FSHD, and inflammatory myopathies.
Aims: To evaluate mDTI as a cross-sectional outcome marker of disease involvement in Late-Onset Pompe Disease (LOPD).
Methods: A cohort of 29 individuals diagnosed with LOPD and 29 age-matched healthy controls (HCs) underwent 3T MRI quantitative muscle MRI of the thighs including 6-point Dixon, multi-echo spin-echo (MESE), and diffusion-weighted imaging (DWI) sequences. Eleven regions of interest (ROIs) were automatically drawn on thigh muscles using the open-source DAFNE software and then validated by an expert operator. ROIs were subsequently co-registered to the DWI dataset. For each ROI (left and right sides averaged), we extracted the fat fraction (FF) from the 6-point Dixon sequence, water T2 (wT2) from the 17-echo MESE, and fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), and axial diffusivity (AD) from the DWI sequence. Quantitative parameters analysis was performed both with a muscle-wise and compartment-wise approach.
Results: Correlation analysis showed a moderate positive correlation between FA and FF across all muscles in patients. FA and RD did not show significant differences between LOPD subjects and HCs. In contrast, AD and MD were significantly reduced in several muscles of the medial and posterior compartments compared to HCs. When analyzing muscles with FF < 10%, AD showed the strongest significant reduction compared to HCs in the majority of muscles.
Conclusion: mDTI metrics, in particular axial diffusivity, may serve as promising parameters for detecting early muscle abnormality and potentially for monitoring disease progression in LOPD, especially in muscles with minimal fat replacement.
P1.14 Creatine-CEST-based pH mapping in healthy volunteer leg muscles
Valentin Henriet1, Pierre-Yves Baudin1,
Benjamin Marty1, Marc Lapert2, Harmen Reyngoudt1
1Institute of Myology, Paris, France, 2Siemens Healthcare SAS, Courbevoie, France
Background: Intramuscular pH is a promising biomarker for neuromuscular disease research. Chemical Exchange Saturation Transfer (CEST) enables high-resolution pH mapping by measuring exchange rates (kex) between metabolites and free water.
Aim: Here, we propose a creatine (Cr)-CEST approach to estimate intramuscular pH in healthy leg muscle, based on Bloch-McConnell (BMC) fitting. To extend the pH range in muscle, an exercise protocol was performed.
Methods: Fifteen healthy volunteers (9 males, 41.6±8.2 years, range 23.9-64.8 years) were scanned on a 3-T clinical system at rest and after 10 minutes of flexion/extension of the right foot with a stretching band (strength=18kg).
A pulsed CEST-GRE sequence was acquired between -4.0ppm and 4.0ppm (0.2ppm steps), centered on the water frequency. A 600ms saturation was applied at 1.0µT. The CEST-WASAB1 method was used for B0-B1 mapping. Water-T2 and water-T1 values were estimated based on multi-spin-echo and MR fingerprinting sequences, respectively, and were included in the fitting model alongside B1 maps. After BMC-fitting, resulting in the kex between Cr and water, pH was estimated with: pH=log10(kex/(2.9×10^9)) +13.62. A basic CEST asymmetry (2.0ppm) analysis was performed resulting in magnetization transfer ratio asymmetry (MTRasym).
Five muscles were manually segmented: gastrocnemius medialis (GM), gastrocnemius lateralis (GL), soleus, tibialis anterior, peroneus longus. Paired t-tests were performed to compare pH and Cr concentration before and after exercise and Pearson-R2 was computed to estimate the correlation between imaging biomarkers
Results: Although the level of pH decrease varies from one volunteer to another, a significant pH decrease was observed after exercise in GM and GL (-0.05 on average, p<0.001). A moderate correlation was observed between MTRasym and pH (R2 = 0.48), indicating the relationship between the increase of Cr concentration and the exercise-induced ‘acidification’. No correlations were found between pH and water-T2 and water-T1 (R2=0.18 and R2=0.09, respectively).
Conclusion: CEST-based pH is sensitive to subtle pH variations in the leg after exercise. The absence of correlation between pH and either water-T2 or water-T1 suggests that pH could provide independent and complementary information. This highlights the potential of pH as a relevant biomarker for investigating muscle physiology, particularly in the context of neuromuscular disorders.
P1.15 Determinants of qMRI variation in skeletal muscle: Effects of sex, age and muscle volume
Linda Heskamp1, Martijn Froeling1
1Center for Image Sciences, Department of Imaging and Oncology, University Medical Center Utrecht, Utrecht, Netherlands
Background: Quantitative MRI (qMRI) is widely used to evaluate neuromuscular disorders. It is sensitive to pathology but often lacks specificity and a link to function. Our ongoing MOTION study collects whole leg qMRI, muscle force, and lifestyle data from 162 healthy volunteers to identify key factors influencing qMRI outcomes and their associations with muscle function and lifestyle.
Aim: Here, we investigate which physiological factors explain between-subject and between-muscle variation in qMRI measures, accounting for known confounders.
Methods: We scanned lower extremity muscles bilaterally, from hip to ankle, in 101 healthy participants. The MRI protocol included 1) 10-echo Dixon-based sequence for fat fraction (FF) quantification, 2) diffusion tensor imaging (DTI) with 30 weightings and 3) 13-echo multi-echo spin-echo sequence for measuring muscle water T2 (T2water).
Images were automatically processed using qMRITools. Diffusion data were corrected for gradient non-linearities. T2water was fitted using a two-compartment EPG model.
We used a full factorial two-level linear mixed-effect model with qMRI measures as dependent variables. Independent variables included physiological factors (muscle volume, sex, age) and confounding factors (FF, SNR). Effect sizes (ES) are reported per one standard deviation of the independent variable.
Results: Each qMRI parameter showed several significant associations; the most relevant findings are presented here.
Muscle volume had the strongest effect on DTI measures: FA decreased (ES: -0.03cm3) and MD increased (0.26cm3) with increasing volume, with effects varying by muscle. Sex affected MD but not FA; age had no effect. SNR had minimal impact on FA and MD, while FF only affected MD.
T2water was higher in females (0.4ms) and increased with age (0.3ms), but was independent of volume. T2water also increased slightly with FF (0.05ms).
FF increased slightly with age (0.2%), and decreased with volume (-1.4%), especially in smaller muscles. SNR had a small negative effect (-0.3%).
Conclusion: Interpretation of qMRI measures should consider age, sex, and muscle volume. For diffusion metrics, muscle volume explains most between-subject and between-muscle variation, especially in smaller muscles. For T2water, sex and volume are the main determinants. Fat fraction is most affected by age, and in small muscles, by volume, likely due to partial volume effects.
P1.16 Whole-body DTI for assessing fasciculation and muscle microstructure in ALS in 10 minutes
Linda Heskamp1, Karleen Oonk2, Boudewijn Sleutjes2, Stephan Goedee2, Leonard van den Berg2,
Martijn Froeling1
1Center for Image Sciences, Department of Imaging and Oncology, University Medical Center Utrecht, Utrecht, Netherlands, 2Brain center, Department of neurology and neurosurgery, University Medical Center Utrecht, Utrecht, Netherlands
Background: Three Dutch neuromuscular expertise centers have jointly developed a harmonized whole-body MRI protocol for diagnosis and longitudinal monitoring of muscular dystrophies. This protocol includes a turbo spin-echo (TSE) Dixon sequence to assess STIR positivity, fat replacement, and muscle atrophy, and a gradient echo (GRE) Dixon sequence for quantitative fat fraction mapping. However, in amyotrophic lateral sclerosis (ALS) additional pathological features, particularly fasciculations, microstructural muscle changes, and denervation, are important for early and accurate diagnosis. These features can be captured by combining TSE Dixon with diffusion tensor imaging (DTI).
Aim: To develop a rapid whole-body DTI protocol to assess fasciculations and muscle microstructure, with the intent to prospectively apply this in 250 people suspected of ALS
Methods: Participants were scanned supine using a head coil and multi-element posterior-anterior coils.
The 3T protocol includes:
DTI: Assesses muscle microstructure and fasciculation, using diffusion weightings of 5 (3x), 100 (10x), 200 (10x), and 400 (10x) s/mm2 (resolution=1.5x.1.5×8mm3, TR=2000-2600ms, TE=57ms, triple fat suppression). Acceleration is achieved with multi-band factor 2 and SENSE 2.6 (anterior-posterior).
TSE Dixon: Multi-acquisition 3-echoes TSE sequence for qualitative assessment of STIR positivity (e.g. edema/denervation) and atrophy quantification (TSE factor=23, effective TE1/ΔTE=60/0.77ms, TR=3000-3300ms, resolution=1.5×1.5×8mm3, SENSE 2.1 [anterior-posterior])
Feet-head coverage was 208 mm (lower back, hips, and upper legs), 176 mm (shoulder and lower legs), or 108 mm (head). In-plane FOV was 480×300 mm.
Results: To date, three healthy volunteers have been successfully scanned using our protocol within a 30-minutes clinical scan slot. Total acquisition time was 15 minutes (DTI: 10min, TSE Dixon: 5min), leaving ample time for participant positioning and calibration scans. Image quality was sufficient in all body regions for DTI and TSE Dixon.
Conclusion: We developed a fast, whole-body DTI protocol and combined this with TSE Dixon for simultaneous assessment of fasciculations, muscle microstructure, edema/denervation, and atrophy. With a scan time within 30 minutes, this protocol delivers the next translational step towards clinical implementation of fasciculation imaging using MRI. Future work will assess the added diagnostic value compared to clinical examination and needle EMG in 250 people suspected of ALS and ALS carriers, and extension to 1.5T.
P1.17 Dynamics not Magnitude of exercise hyperemia correlate with aerobic muscle metabolic performance
Melissa Hooijmans1, Sandra van den Berg2,
Jeroen Jeneson3
1Department of Human Movement Sciences, Faculty of Behavioral and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, Netherlands, 2Department of Radiology and Nuclear Medicine, Amsterdam UMC, University of Amsterdam, Amsterdam, Netherlands, 3Child Development and Exercise Center, Wilhelmina Children's Hospital, University Medical Center Utrecht, Utrecht, Netherlands
Background: Working muscles recruit blood flow to support their elevated demand for aerobic ATP synthesis, a phenomenon known as exercise hyperemia. In vivo optical recordings of blood flow dynamics in arm muscle of healthy individuals revealed substantial inter-individual differences in peak blood flow index in response to the surrogate exercise intervention of cuff inflation and deflation. We hypothesize that the magnitude of the hyperemic response is closely related to and influences muscle metabolic performance. To explore this, we investigated in a cohort of healthy individuals performing bouts of dynamic exercise if such variations in peak hyperemic blood flow affected homeostasis of energy and pH balance in working muscle and post-exercise metabolic recovery kinetics.
Methods: Here, we performed interleaved quantitative in vivo 1H magnetic resonance blood oxygenation level-dependent (BOLD) imaging (MRI) and 31P MR spectroscopy (31P MRS) recordings from the upper arm musculature prior to, during and after a five-minute bout of arm-cycling against a 15 W workload, respectively. From these datasets we derived three MRI parameters (BOLD MAX (%), time-to-peak (TTP; s) and initial slope of BOLD signal change (BOLD IS; %)) informing on magnitude and timing of the hyperemic response to exercise, respectively, and three MRS parameters (change in phosphocreatine level (ΔPCr; %) and pH drop during exercise, and time to 95% recovery of PCr (95RT) post-exercise) informing on aerobic metabolic performance of the working muscle.
Results: Each subject recruited all available motor units of the upper-arm musculature to complete the arm-cycling task evidenced by >80% depletion of intramuscular PCr stores. We found a threefold range of variation in both BOLD MAX and TTP, and a fivefold range in BOLD IS among individuals within the tested cohort. Correlation analysis next revealed that the absolute value of BOLD IS, not BOLD MAX, best predicted aerobic muscle metabolic performance.
Conclusion: These results suggest that individual vascular responsiveness to dilatory versus constrictive agents rather than magnitude of exercise hyperemia per se govern aerobic metabolic performance of healthy working upper-arm muscles.
P1.18 Advancing clinical trials in myotonic dystrophy type 1: refining radiological, clinical and patient-reported outcome measures
Louise Iterbeke1, Lotte Huysmans2,3, Kobe Bamps4,5, Ronald Peeters6, Veerle Goosens6, Frederik Maes2,3, Patrick Dupont7, Kristl Claeys1,8
1Laboratory for Muscle Diseases and Neuropathies, Department of Neurosciences, KU Leuven, and Leuven Brain Institute (LBI), Leuven, Belgium, 2Department ESAT/PSI, KU Leuven, Leuven, Belgium, 3Medical Imaging Research Center, University Hospitals Leuven, Leuven, Belgium, 4Department of Cardiovascular Sciences, KU Leuven, Leuven, Belgium, 5Department of Cardiology, University Hospitals Leuven, Leuven, Belgium, 6Department of Radiology, University Hospitals Leuven, Leuven, Belgium, 7Laboratory for Cognitive Neurology, Department of Neurosciences, KU Leuven, and Leuven Brain Institute (LBI), Leuven, Belgium, 8Department of Neurology, University Hospitals Leuven, Leuven, Belgium
Introduction: Myotonic dystrophy type 1 (DM1) is a progressive, multi-systemic disorder. In classic DM1, the predominant symptoms are slowly progressive, predominantly distal muscle weakness and myotonia. As novel therapeutics emerge for DM1 patients, there is a need to define reliable, sensitive and non-invasive outcome measures to assess future treatment efficacy.
Methods: Thirty-three adult classic DM1 patients and 33 age- and sex-matched healthy controls were assessed at baseline, 12, and 18 months. Muscle MRI included 6-point Dixon and T2H2O sequences. Proton density fat fraction (PDFF) was quantified in 18 proximal and 10 distal leg muscles through semi-automated 3D segmentation of each muscle. Clinical outcome measures consisted of the Motor Function Measure 32 (MFM32), 6-Minute Walk Distance (6MWD), 10-Meter Walk Test (10MWT), 30-Second Sit-to-Stand (30SSS), Medical Research Council Sum Score (MRCSS) and 9-Hole Peg Test. Strength assessments included isometric dynamometry, handgrip strength, key and tip pinch strength, and peak cough flow. Patient-reported outcomes comprised the DM1-ActivC, Brief Pain Inventory, Individualized Neuromuscular Quality of Life, and the Fatigue and Daytime Sleepiness Scale.
Results: At baseline, PDFF was significantly higher in all leg muscles of DM1 patients compared to controls. After 12 months, PDFF increased significantly in 8/18 proximal leg muscles (weighted mean of all proximal muscles: +1.1%, 95%CI: 0.09 to 2.19) and 8/10 distal leg muscles (weighted mean of all distal muscles: +1.8%, 95%CI: 1.21 to 2.45) in patients. At baseline, patients also performed significantly worse than controls on all clinical and strength measures. After 12 months, significant declines were observed in MFM32 ‘transfer and standing’ subscore (-5.0%, 95%CI: -10.04 to -0.04), MRCSS (-2.2, 95%CI: -3.20 to -1.10), and DM1-ActivC (-2.6, 95%CI: -5.19 to -0.12) in patients. In contrast, no significant changes in MRI or clinical measures were observed among healthy controls. At baseline, total PDFF of the 18 proximal and 10 distal leg muscles significantly correlated with MFM32, 6MWD, 10MWT, and 30SSS.
Conclusion: Based on results at 12 months, quantitative muscle MRI, along with the MFM32, MRC SS, and DM1-ActivC, are promising outcome measures for future clinical trials in adult classic DM1 patients. Detailed 18-month results will be presented at MYO-MRI 2025.
P1.19 Quadriceps fatiguability and recovery after exercise through the menstrual cycle
Carly Jones1, Emily Kraus1, Garry Gold1
1Stanford University, Stanford, United States
Background: Menstrual cycle impact on athletic performance is a top concern for female athletes. Some studies found increased strength, slower muscle recovery and higher fatiguability mid-cycle; others found no difference. Previous studies were limited by variable menstrual cycle phase definitions, contraceptives use, and muscle function quantification. T2 and T2* relaxation may be more sensitive to muscle fatiguability and acute recovery than traditional methods.
Aims: Determine if muscle fatiguability and acute recovery vary with the menstrual cycle.
Methods: Three naturally menstruating volunteers were scanned in a 3T MR scanner using a T2 (1:49 min, TE: 6.456-103.296ms, 16 echoes) and T2* (38s, TE: 1.128-25.608ms, 16 echoes) mapping sequence before (1 series) and after (16 series) a wall-sit exercise. Participants held wall sits for the same length of time in both sessions.
Quadriceps muscles were segmented automatically (MuscleMap) and T2 and T2* relaxation was calculated using a mono-exponential decay model (MATLAB).
T2 and T2* signal decay post-exercise was fit to a mono-exponential decay model. Difference from peak to baseline signal, Δ, is related to muscle fatigue, and signal decay constant, τ, is related to acute muscle recovery.
Results: Subject 1 (irregular) had similar T2 (visit 1: Δ:8.2ms, τ:14.2min; visit 2: Δ:6.1ms, τ:13.2min) and T2* (visit 1: Δ:6.2ms, τ:12.2min; visit 2: Δ:4.4ms, τ:13.4min) signal behaviour after exercise for two visits spaced two months apart.
Subject 2 (regular) had longer T2 (visit 1: Δ:5ms, τ:17.6min; visit 2: Δ:4.9ms, τ:14.6min) and T2* (visit 1: Δ:5ms, τ:25.6min; visit 2: Δ:3.9ms, τ:14.6min) signal decay in visit 1 (luteal phase) compared to visit 2 (follicular phase).
Subject 3 (regular) had similar T2 (visit 1: Δ:7ms, τ:12.1min; visit 2: Δ:4.5ms, τ:12.5min) and T2* (visit 1: Δ:6.5ms, τ:8.2min; visit 2: Δ:3.6ms, τ:8.5min) signal behaviour after exercise for two follicular phase visits.
Conclusion: A pronounced T2 and T2* response to exercise and subsequent signal decay was observed in all volunteers and was similar at follow-up for all participants. T2 and T2* signal decay were longer in the luteal phase compared to the follicular phase. A larger cohort will be recruited to further explore how the menstrual cycle may influence muscle function.
P1.20 Progression of Miyoshi muscular dystrophy in thigh muscles monitored by quantitative MRI
Ivica Just1,2, Petra Hnilicova3, Radka Klepochova1,4,5,
Monika Koprušáková Turčanová2,
Martin Kolísek6, Martin Krššák1,2
1Division of Endocrinology and Metabolism, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria, 2High Field MR Centre, Department of Biomedical Imaging and Image guided Therapy, Medical University of Vienna, Austria, Vienna, Austria, 3Department of Neurology, Jessenius Faculty of Medicine, Comenius University, Martin, Slovakia, 4Institute of Experimental Endocrinology, Biomedical Research Center, Slovak Academy of Sciences, Bratislava, Slovakia, 5Institute of Measurement Science, Slovak Academy of Sciences, Bratislava, Slovakia, 6BioMed Martin, Jessenius Faculty of Medicine, Comenius University, Martin, Slovakia
Background: Miyoshi myopathy (MM) is a rare autosomal recessive distal myopathy characterized by progressive atrophy of leg muscles, typically presenting in late adolescence or early adulthood. It is caused by mutations in the DYSF gene encoding dysferlin.
Aim: This study aimed at monitoring progression of the atrophy over 1 year in MM patients in later stages and comparing to their healthy relatives using quantitative fat fraction imaging.
Methods: Study was performed at 3T Prisma Fit (Siemens) using 1H body flex coil (Siemens) in two time points year apart. Seven siblings were participating in the study, four patients with MM (2f/2m, age 37±5y, BMI 22±3kg/m2, disease duration 13y, 20y, 21y and 25y) and three healthy siblings (f/2m, age 38±6y, BMI 23±2kg/m2). Measurement protocol consisted of multi-echo Dixon, T1-w and T2-w images. Segmentation was done in 3D slicer, analyzing 9 individual muscles and whole muscle segment as well. Volumes and average fat fraction (FF) were calculated for each single muscle and global muscle segment and comparison in two time points was done. As a measure of effect change, standardized response mean SRM was reported additionally to t-test.
Results: Whole muscle segmentation showed increase in MM patients (71.3%-81.2%) with the SRM of 0.9 in comparison to healthy siblings (12.3%-12.1%) with SRM of 0.34. Comparing individual muscles, biggest increase in FF, supported by t-test and also SRM, was in gracilis, sartorius and biceps femoris longus muscle. Differences in the pattern of dystrophy progression was shown - person with the longest disease duration (25y) manifested lowest FF in biceps femoris longus (41.1%) and adductor magnus muscle (65.2%) and person with average disease duration (20y) showed highest FF in gracilis (70.3%) and sartorius muscle (95%).
Conclusion: Although in the later stages of the disease, fat inflitration in thigh muscles can vary individually between patients, even siblings. Muscles that can differ in patterns are sartorius, gracilis, biceps femoris longus and adductor magnus.
P1.21 Tracking muscle degeneration and disease activity in FSHD using Qualitative Longitudinal MR Imaging
Anna-Lena Mayer1, Anika Starke1, Matthias Türk1,
Michael Uder1, Arnd Dörfler1, Armin Nagel1,
Elisabetta Gazzerro3, Teresa Gerhalter1,2
1Universitätsklinikum Erlangen, Erlangen, Germany, 2Medical University of Graz, Graz, Austria, 3Charité Universitätsmedizin und Max-Delbrück-Centrum für Molekulare Medizin, Berlin, Germany
Background: Facioscapulohumeral Muscular dystrophy (FSHD) is a hereditary neuromuscular disease caused by genetic mutations that result in inappropriate DUX4 gene expression, resulting in progressive, often asymmetric muscle weakness, fatty degeneration, and atrophy. Understanding disease progression is vital for evaluating new therapies in clinical trials.
Aim: This study assessed one-year progression of lower leg muscle involvement in FSHD using 3T magnetic resonance imaging (MRI), focusing on hyperintense lesions in short-tau inversion recovery (STIR) sequences as potential markers of active muscle damage. Fatty degeneration was evaluated using T1-weighted images.
Methods: Thirty-one genetically confirmed FSHD patients (21 men; mean age 45.2 ± 17.2 years) and 30 healthy controls underwent MRI scans. Fat infiltration was graded using the Goutallier visual score on T1-weighted images; edema was assessed via STIR. Clinical muscle function was evaluated using the Medical Research Council (MRC) scale. Follow-up imaging was conducted after approximately one year.
Results: At baseline, MRC scores averaged 26.6 ± 4.2 (right leg) and 26.5 ± 4.7 (left leg), out of 30. The gastrocnemius, soleus, and tibialis anterior muscles were most commonly affected. Some participants showed pronounced fatty degeneration (Goutallier score 4) in the fibularis longus and tibialis posterior muscles. Flexor digitorum/hallucis longus muscles were less involved (maximum score 2).
After 53.9 ± 4.5 weeks, follow-up of 27 patients showed mild progression of fatty degeneration, particularly in less affected muscles. Muscles previously showing STIR hyperintensity often progressed to fatty replacement. MRC scores changed only slightly (right: -0.2 ± 0.6; left: 0.1 ± 0.6). Notably, advanced fatty degeneration did not always correspond with loss of function; some patients with extensive changes remained ambulatory.
Conclusion: STIR hyperintensities may serve as early indicators of subsequent fatty degeneration in FSHD. MRI provides a sensitive, objective tool for tracking disease progression, supplementing clinical assessments and aiding evaluation of therapeutic efficacy.
P1.22 Multi-parametric 1H MRI of lower leg muscle in patients with Becker muscular dystrophy
Yvonne Mileder1,2, Valentina Schunk2,
Pierre-Yves Baudin3, Susi Rauh4, Katharina Tkotz2, Moritz Zaiss2, Frank Roemer2, Arnd Dörfler2,
Michael Uder2, Elisabetta Gazzerro5, Armin Nagel2, Teresa Gerhalter2,6
1Graz University of Technology, Graz, Austria, 2University Hospital Erlangen, Erlangen, Germany, 3Institute of Myology, Paris, France, 4Leiden University Medical Center, Leiden, Netherlands, 5Charité University of Medicine and Max Delbrück Center of Molecular Medicine, Berlin, Germany, 6Medical University of Graz, Graz, Austria
Background: Becker muscular dystrophy (BMD) is a hereditary neuromuscular disorder caused by dystrophin gene mutations, leading to impaired energy metabolism, inflammation, altered muscle fiber size, and progressive replacement of muscle by fat and fibrotic tissue. Quantitative MRI has been used to track disease progression, with fat fraction (FF) correlating well with functional measures. However, even non-fat-replaced muscles show reduced function, highlighting the need for non-invasive markers reflecting early tissue changes.
Aim: This study investigates diffusion tensor imaging (DTI), water T2 relaxation times, and chemical-exchange saturation transfer (CEST) as potential early markers in ambulatory BMD patients.
Methods: Thirty BMD patients (mean age 35.6 ± 18.6 years) and 22 healthy controls (mean age 39.6 ± 17.8 years) underwent 3T MRI of the lower legs. FF was assessed using a 6-point DIXON method. Water T2 values were calculated using tri-exponential fitting. DTI parameters were calculated from diffusion-weighted spin-echo imaging and CEST effects were investigated at 3.5 ppm (amide-weighted), 2.5 ppm (phosphocreatine-weighted) and 2.0 ppm (total creatine-weighted) away from the water signal using a low-power saturation pulse approach. Functional performance was measured using the 10-meter walk test (10mwt) and supine-to-stand (STS) test.
Results: Patients showed slower functional performance (10mwt: BMD 1.2±0.2 m/s, controls 1.4±0.2 m/s, p=0.013; STS: BMD 4.5±3.5 s, controls 2.5±0.1 s, p=0.045) and higher FF (BMD 0.15±0.19, controls 0.06±0.04, p=0.025), especially in gastrocnemius medialis and soleus. Dystrophic muscles had reduced mean diffusivity (p=0.012) and increased water T2 heterogeneity (p=0.004), but no significant change in mean T2 or fractional anisotropy. CEST effects were elevated in patients across all markers: amide (BMD 0.010±0.003, controls 0.008±0.004, p<0.001), phosphocreatine (BMD 0.011±0.003, controls 0.009±0.004, p<0.001), and total creatine (BMD 0.010±0.003, controls 0.008±0.004, p=0.001).
Conclusion: While DTI and T2 changes were prominent in fat-infiltrated muscles (FF>0.1), CEST abnormalities were also found in preserved muscles (FF < 0.1), suggesting early metabolic disturbances detectable via CEST MRI. Ongoing 1-year follow-up (to be completed September 2025) will assess the relationship between these imaging biomarkers and clinical outcomes.
P1.23 Baseline quantitative whole-body muscle MRI and functional outcomes from a prospective natural history study in adults with FSHD1.
Matthias Opsomer1,2, Lotte Huysmans3,4,
Kobe Bamps5,6, Ronald Peeters7, Veerle Goosens7, Frederik Maes3,4, Patrick Dupont8, Kristl Claeys1,2
1Laboratory for Muscle Diseases and Neuropathies, Department of Neurosciences, KU Leuven and Leuven Brain Institute (LBI), Leuven, Belgium, 2Department of Neurology, University Hospitals Leuven, Leuven, Belgium, 3Department ESAT/PSI, KU Leuven, Leuven, Belgium, 4Medical Imaging Research Center, University Hospitals Leuven, Leuven, Belgium, 5Department of Cardiovascular Sciences, KU Leuven, Leuven,, 6Department of Cardiology, University Hospitals Leuven, Leuven, Belgium, 7Department of Radiology, University Hospitals Leuven, Leuven,, 8Laboratory for Cognitive Neurology, Department of Neurosciences, KU Leuven, and Leuven Brain Institute (LBI), Leuven, Belgium
Background: Facioscapulohumeral muscular dystrophy type 1 (FSHD1) is a slowly progressive muscle disorder, caused by contraction of the D4Z4 repeat on chromosome 4q35.
Aim: We report the baseline findings of quantitative whole-body muscle MRI, clinical outcome measures (COMs) and patient-reported outcome measures (PROMs) from a prospective natural history study in adults with FSHD1.
Methods: Each patient underwent a whole-body muscle MRI including T1-weighted, 6-point Dixon and STIR sequences. Semi-automated 3D whole-muscle segmentation was applied to quantify proton density fat fraction (PDFF) in proximal and distal leg, back, arm and shoulder muscles. COMs included the Vignos and Brooke scales, 10-grade FSHD clinical severity score (FSHD CSS; Ricci Score), 15-grade clinical score and FSHD-COM. Functional measures comprised Motor Function Measure 32 (MFM32), 6-Minute Walk Test (6MWT), 30-Second Sit-To-Stand (30STS). PROMs included FSHD-RODS, ACTIVLIM, Individualized Neuromuscular Quality of Life Questionnaire (INQoL) and Fatigue Severity Scale (FSS).
Results: A total of 60 symptomatic FSHD1 patients were included (mean age 52.1 ± 15.7 years; 55.0% female). The mean age of onset of neuromuscular symptoms was 34.4 ± 17.8 years and mean age of diagnosis 43.3 ± 18.2 years. The mean D4Z4 repeat number was 7.2 ± 1.8. Clinical assessment showed a median Vignos scale of 2 (IQR 1-3), median FSHD CSS of 1.5 (IQR 1.0-2.5), mean FSHD clinical score of 4.8 ± 3.2 and a median FSHD-COM score of 12 (IQR 6-19). The median MFM32 was 90% (IQR 56-97%) for D1, 94% (IQR 89-97%) for D2, and 100% (IQR 95-100%) for D3. Mean 6MWT distance was 490 ± 148 meters. Median 30STS was 12 (IQR 5-16). PROMs revealed a median FSHD-RODS score of 55 (IQR 43-62) and a median ACTIVLIM of 31 (IQR 26-35). INQoL indicated that weakness was the main complaint in most patients (median 52.6%), followed by pain (36.8%) and fatigue (31.6%). The median FSS was 41 (IQR 27-46). Muscle MRI PDFF quantification results are being analysed and will be presented at the MYO-MRI 2025 meeting.
Conclusion: The ongoing quantitative muscle MRI analysis will further elucidate the relationship between muscle fat infiltration and functional impairment in this cohort of adult patients with FSHD1.
P2.02 Associations between MRI muscle structure and tensiomyography contractile parameters in older adults
Katarina Puš1,2,3, Diana A. Madrid Fuentes4,
Ashley A. Weaver4, Jeannette R. Mahoney5,
Boštjan Šimunič1
1Institute for Kinesiology Research, Science and Research Centre Koper, Koper, Slovenia, 2Department of Health Sciences, Alma Mater Europaea University, Maribor, Slovenia, 3Faculty of Sport, University of Ljubljana, Ljubljana, Slovenia, 4Department of Biomedical Engineering, Center for Injury Biomechanics, Wake Forest University School of Medicine, Winston-Salem, United States, 5Renaissance School of Medicine, Stony Brook University, Department of Neurology, Stony Brook, United States
Background: Muscle composition changes with age, including increased adipose tissue in skeletal muscle impairing muscle strength, mobility and function. Tensiomyography (TMG), a non-invasive tool, has been shown to detect early muscle atrophy and muscle composition changes before imaging techniques, yet its sensitivity to detect age-related fat infiltration remains unexplored.
Aims: This study aims to explore the association between TMG-derived parameters, physical performance and MRI-derived muscle fat fraction (MFF) in older adults.
Methods: Fifty-one Slovenian older adults (mean age 82±8 years, 53% female) were recruited from nursing homes, community and retirement groups. Muscle contractile properties were evaluated with TMG-derived parameters (delay time, radial displacement, contraction velocity) on vastus lateralis (VL) and biceps femoris (BF) of the right leg. Physical performance was evaluated with 5 sit-to-stand (STS), timed up-and-go (TUG) and gait speed (over the 4 meters) tests. MFF was extracted from MRI Dixon sequence of the thigh and the analysis incorporated five slices, using Dafne and 3D Slicer. MFF was calculated as the percentage of the fat signal intensity to the sum of the fat and water signal intensities. Correlation analysis was used to investigate associations.
Results: In the mechanical component, we found strong associations between knee extensors MFF and radial displacement of VL (r=-.522, p<.001) and contraction velocity of VL (r=-.546, p< 001). Physical performance component showed positive moderate correlation between TUG and MFF of knee flexors (r=.366, p =.012) and extensors (r =.388, p =.008). Additionally, we found a moderate positive correlation between STS and knee extensors MFF (r =.345, p=.022). Gait speed presented a negative trend with the MFF of both muscle groups (flexors: r=-.280, p=.062, extensors: r=-.279, p=.067).
Conclusion: TMG effectively captured muscle quality decline linked to functional deficits in older adults with radial displacement and contraction velocity exhibiting the strongest ties to MFF. Physical performance outcomes were moderately associated with MFF, supporting the relevance of muscle composition for mobility and strength. The strong associations suggest that TMG captures intramuscular fat infiltration usually captured with MRI. While MRI remains essential for precise MFF mapping, TMG's portability could complement imaging in resource-limited settings.
P2.03 Investigating the effect of pulse-width in NMES with dynamic MRI in the forearm muscles
Sabine Melanie Räuber1,2, Francesco Santini1,2,
Marta Brigid Maggioni1,2
1University of Basel, Allschwil, Switzerland, 2University Hospital of Basel, Basel, Switzerland
Background and Aims: Neuromuscular electrical stimulation (NMES) combined with dynamic MRI and force measurement offers a reproducible method to assess muscle activity. This presents a unique opportunity to investigate, in real-time, the influence of NMES parameters during acquisition, rather than before or after.
Dynamic MRI ideally requires a minimally fatiguing, rhythmic isometric tetanic contraction for each on/off cycle throughout the acquisition while also improving patient comfort by minimising pain associated with NMES. Since pulse width is reported in literature to influence muscle fatigue, we aim to investigate how pulse width affects measured muscle strain while maintaining constant output grip force.
Methods: An MRI-compatible grip force sensor measured NMES-evoked forearm muscle grip strength during acquisition. This work compared two NMES protocols (60 Hz frequency, 0.78 s on/0.78 s off time) producing the same output force with varying pulse duration (80 µs vs 450 µs). In between the two stimulation protocols, the volunteers rested for 3 minutes. A 4-point accelerated 4D flow MRI sequence with CINE acquisition triggered by NMES was acquired on three healthy female volunteers (27-31 years old) on a 3 T system.
We used the top 5% of maximum intensity of each strain map to identify high-strain regions. Coordinates of these peak-strain voxels were extracted for distance-based clustering analysis (smaller distances between high strain voxels indicate higher clustering).
Results: The short-pulse protocol exhibits considerably more high-strain points (approximately one order of magnitude) than the wide-pulse protocol. In the short-pulse protocol, the stimulated muscle exhibited larger areas of high strain, suggesting a broader region of muscle activation, while the wide-pulse protocol resulted in smaller high-strain areas, indicating more defined regions of muscle activation.
Conclusion: We explored the influence of short vs long NMES pulse widths on muscle contractility in the forearms of three volunteers by utilising dynamic MRI with synchronised NMES and force recording. While both pulse types produced similar force outputs, the results suggest a difference in the inherent muscle recruitment. These differences need further physiological interpretation and should potentially be taken into account when extending this technique in studies with patients affected by neuromuscular diseases.
P2.04 The "muscle toolbox": A multi-center multi-parametric natural history study in children with neuromuscular diseases using a harmonized protocol
Susi Rauh1, Donnie Cameron2, Linda Heskamp3,
Jeroen L.M. van Doorn2, Evelien Fleerakkers1,
Roosmarijn Brenninkmeijer3, Marinus J. Becks2,
Mark C. Kruit1, Rutger A.J. Nievelstein3, Maaike Pelsma2, Selma C. Tromp1, Corrie E. Erasmus2, Erik H. Niks1, W. Ludo van der Pol3, Martijn Froeling3, Bart Bartels3, Nens van Alfen2, Hermien E. Kan1
1Leiden University Medical Center, Leiden, Netherlands, 2Radboud University Medical Center, Nijmegen, Netherlands, 3University Medical Center Utrecht, Utrecht, Netherlands
Background: Neuromuscular disorders (NMDs) are rare; therefore, pooling data across multiple centers is crucial to accelerate trial readiness. Differences in current procedures to obtain muscle strength, motor function, MRI, or ultrasound can result in differences in outcome measures and might require additional examinations.
Aim: This study aims to 1) harmonize imaging and muscle function protocols for NMDs to enable data pooling across three Dutch NMD expert centers, 2) characterize the natural history over 4 years in children with NMDs using our harmonized quantitative function and quantitative imaging protocols, and 3) improve current diagnostics and clinical care by including whole-body fat fraction measures as standard of care in diagnosis and follow-up visits. We aim to determine disease progression and relationships among clinical, functional, and imaging outcomes, and investigate the ability of imaging metrics to predict clinical milestones.
Methods: This is a longitudinal study for 4 years with yearly visits. We aim to include 150 children (<18 years at inclusion) with a confirmed diagnosis of spinal muscular atrophy (SMA), myotonic dystrophy type I (DM1), Becker, Duchenne, facioscapulohumeral, and limb-girdle muscular dystrophies (BMD, DMD, FSHD, LGMD). The participants will undergo whole-body imaging, including either MRI, ultrasound, or both. The MRI protocol includes a whole-body Dixon protocol for fat fraction determination, 1H spectroscopy for pH determination in the soleus muscle, and additional disease-dependent sequences to characterize inflammation or diffusion. Ultrasound protocol includes quantitative assessment of eight muscles to determine echogenicity. Each participant will be assessed with motor function scales and quantitative strength tests.
Results: Currently, 14 patients are included (BMD=5, DMD=3, LGMD=3, SMA=1, DM1=2), aged 6-16, of whom all underwent muscle function and strength tests, 7 MRI, 6 ultrasound, and 4 both. The whole-body Dixon scan can be acquired within 5 minutes, and the disease-specific sequences within 15 minutes, making the MRI protocol feasible and tolerable for most children.
Conclusion: This work is an important step towards harmonisation of muscle measurements in NMDs. Furthermore, it will give insight into the additional value of MRI biomarkers beyond fat fraction. In the future, the harmonized protocol will be extended to adults and implemented in the diagnosis.
P2.05 Deep automatic and interactive segmentation in MRI of pathological skeletal muscles
Louis Rigler1, Benjamin Marty1, Pierre-Yves Baudin1
1Institute of Myology, Paris, France
Background: Quantitative MRI provides increasingly relevant biomarkers for the characterization and monitoring of neuromuscular disorders (NMD). Such analyses generally require delineating regions of interest (ROI) to separate the different muscles or muscle groups. While recent deep learning methods provide high quality automatic segmentations, mislabeling errors remain, especially in highly fatty-replaced muscles.
Methods: We propose a deep learning-based method for semi-automatic 3d muscle segmentation where users can correct errors by providing guidance to the network in the form of clicks or scribbles.
Results: Data used for training and testing were 3D (64 slices) MRI Dixon out-of-phase legs of healthy and NMD patients acquired on 3T in Paris, with a large range of fatty-replacements. Annotations consisted in 9 manually segmented muscles in 1 out of every 4 slices.
Using the popular nnU-Net framework, we modified the training pipeline to introduce user-guidance in the form of clicks. In each training batch, simulated clicks are progressively and randomly added into dedicated input channels with probabilities depending on the size of the mislabeled regions. We also used a modified loss that forces the network to focus at first on the click-less segmentation and to prioritize click influence in later epochs.
For testing, we simulated 20 successive clicks based on the ground truth, computing relevant error metrics (Dice coefficient, largest error radius) for each prediction for all labels. We compared the resulting guided segmentations to an unmodified non-interactive nnU-Net (‘no-click model’).
Without click, the proposed model had Dice (0.895 ± 0.033) and largest error radius (4 ± 1.52 pixels) equivalent to the no click model (0.895 ± 0.032 and 4.2 ± 1.51 pixels respectively). After 10 clicks, our model improved the resulting segmentation (mean Dice 0.906 ± 0.029, p-value = 0.001, largest error radius 2.8 ± 1.01 pixels, p-value = 1e-4) and furthermore after 20 clicks (mean Dice 0.914 ± 0.026, p-value = 2e-5, largest error radius 2.35 ± 0.88 pixels, p-value = 3e-6).
Conclusion: After further validation, this approach could be implemented into a user-interface to automatically generate muscle ROIs and, correct remaining mislabeling errors with a few simple gestures.
P2.06 MRI quantification of lower body muscle fat fraction is associated with NSAD functional decline in men with Becker muscular dystrophy
Kelly Rock1, Donovan J. Lott1, Rebecca J. Willcocks1, Alison M. Barnard1, Sean Forbes1, Claudia Senesac1, Prathyusha Bellam1, Shakeel Ahmed1, Eric Baetscher2, William Rooney2, Glenn Walter1, Krista Vandenborne1
1University Of Florida, Gainesville, FL, United States, 2Oregon Health & Science University, Portland, OR, United States
Background: Becker muscular dystrophy (BMD) leads to skeletal muscle fat replacement and impaired function. Quantitative magnetic resonance (qMR) whole-body imaging (WBI) allows for rapid data acquisition of multiple muscles and the development of muscle fat fraction (FF) composites. Functionally-relevant biomarkers, like qMRI, are needed to enhance selection criteria and efficacy assessment in BMD clinical trials. The North Star Assessment for Limb-girdle type muscular dystrophy (NSAD) has the potential to assess functional performance in men with BMD across a wide functional spectrum.
Aims: To assess how baseline qMRI FF of lower body muscles relates to NSAD performance and longitudinal changes in NSAD in men with BMD.
Methods: 38 men with BMD (18-62 years) completed a 2-year natural history study including 3-point Dixon qMR WBI and NSAD assessments. Baseline lower body muscle fat fraction (FF) was calculated as the average FF of 9 key muscle groups of the right trunk and lower extremity. Participants were stratified into four groups: FF<20, FF20-40, FF40-60, and FF60-80. Spearman's rho assessed correlations, one-way ANOVA tested between-group differences, and longitudinal NSAD changes over 1-2 years were reported.
Results: Lower-body FF ranged from 6 to 79% (median: 44%) and baseline NSAD scores ranged from 2 to 54 (median: 25). FF was strongly negatively correlated with baseline NSAD score (rs=-0.92; p<0.001), indicating that higher FF is associated with poorer motor function. When stratified by baseline FF, NSAD showed significant between-group differences in NSAD scores (p<0.001) with highest scores in the lowest FF groups (FF<20, FF20-40, FF40-60). NSAD scores were not significantly different between FF40-60 and FF60-80 groups. NSAD scores declined most in individuals with FF40-60 with median changes of -2.0 at 1 year and -3.5 at 2 years, followed by FF20-40 (-1.5, -2.0) and FF60-80 (-1.0 -1.5). Those with FF<20 showed negligible change.
Conclusion: Lower-body muscle FF, measured by qMR WBI, is associated with functional performance in men with BMD. Muscle FF corresponds with baseline function and longitudinal decline, particularly in individuals with lower body muscle FF greater than 20%. Stratifying participants by FF may improve clinical trial design by identifying participants who are likely to show functional decline.
P2.07 Determining the repeatability of the rapid qDESS sequence in the quantification of muscle-water T2 of the upper leg.
Gabriel Rossetto1, David M. Higgins2,
Kieren G. Hollingsworth1
1Newcastle University, Newcastle upon Tyne, United Kingdom, 2Clinical Science, Philips, Farnborough, United Kingdom
Background: The T2 relaxation time of skeletal muscle is a key biomarker of acute disease activity in muscular dystrophies. Quantification is possible using the accepted MESE sequence, however the limitations imposed by the specific absorption rate and T1 relaxation on the TR make this a time-consuming method over larger volumes of tissue. qDESS is a partially balanced 3D gradient echo sequence with a short TR, acquiring two echoes per repetition, from which a T2 map can be modelled from the ratio of the signal magnitudes of both echoes. Therefore, the rapid qDESS sequence can make early detection and monitoring of potential curative treatments significantly easier over whole limb/body scans.
Aims: To determine the test-retest repeatability of the measured T2 values from the upper leg of healthy volunteers using qDESS and compare it with MESE.
Methods: Test-retest T2 maps of 8 healthy volunteers (7F, 21-57y) with repositioning were acquired centrally in the upper leg with qDESS and MESE at 3.0T. ROI analysis on matched slices covering the full length of the acquisition was performed on the qDESS and MESE scans. qDESS took 1:12mins/17 slices and MESE took 4:44mins/5 slices. A Bland-Altman analysis was performed on all results to determine the test-retest repeatability.
Results: Bland-Altman analysis for the sequences (qDESS/MESE) showed no significant statistical nor clinical bias between the test-retest T2 values at the upper leg (-0.29/-0.11 ms), with tight bounds for the 95% limits of agreement (+LOA = 2.68/1.52 ms, -LOA = -3.26/-1.74 ms). There was also no significant bias with tight bounds for the LOA for three analysed muscles: vastus lateralis (0.52/-0.18 ms, +LOA = 2.53/1.04 ms, -LOA = -3.57/-1.40 ms), vastus intermedius (-0.47/-0.09 ms, +LOA = 2.05/1.76 ms, -LOA = -3.00/-1.94 ms) and semitendinosus (0.21/-0.04 ms, +LOA = 3.41/1.76 ms, -LOA = -2.99/-1.85 ms).
Conclusion: qDESS shows good repeatability in test-retest T2 values across the anterior to posterior muscles in the upper leg with an acquisition time 14x faster per slice than that of MESE. It is a promising T2-mapping technique for longitudinal assessment of muscular dystrophies over large volumes of skeletal muscle in clinically acceptable times.
P2.08 Use of a rapid qDESS sequence to measure skeletal muscle T2 during intense exercise compared with MESE and spectroscopy
Gabriel Rossetto1, Rory Brown1, David M. Higgins2,
Kieren G. Hollingsworth1
1Newcastle University, Newcastle upon Tyne, United Kingdom, 2Clinical Science, Philips, Farnborough, United Kingdom
Background: Quantification of skeletal muscle water T2 relaxation time is a key biomarker of disease activity in muscular dystrophies. However, standard quantification methods do not provide whole muscle/body coverage in a clinically feasible time. We hypothesise that the 3D qDESS sequence (quantitative double echo steady state sequence) can play that role.
Aims: To compare the measurement of T2 by qDESS with MESE and spectroscopy in the lower leg before and immediately after an intense plantar flexion exercise that is expected to elevate T2 temporarily in relevant muscle groups.
Methods: 10 healthy volunteers (7M, 26-53y) pushed a 10kg load on an in-scanner ergometer for 5 minutes at 83 pushes/minute. qDESS and MESE were acquired at 3.0T with three matched slices at mid-calf for ROI analysis: STEAM spectroscopy was acquired from the peroneus muscle. qDESS took 1:03mins/15 slices and MESE took 4:44mins/5 slices. All three sequences were acquired pre-exercise and immediately post-exercise in the order qDESS-STEAM-MESE. Correlation and Bland-Altman analysis was performed on all results.
Results: Peroneus T2 by both qDESS and MESE had a strong and significant linear correlation with STEAM spectroscopy (κ>0.96, p<0.001). Bland-Altman analysis showed no significant statistical nor clinical bias between skeletal muscle T2 measured by qDESS and STEAM (0.34ms) or between MESE and STEAM, (1.77ms), though the Bland Altman analysis between qDESS and MESE showed a bias of (-2.12ms), i.e. MESE measured a slightly higher T2 than qDESS.
Peroneus, lateral and medial gastrocnemii ROIs show significant T2 elevation post exercise on both qDESS and MESE T2, while soleus and tibialis anterior show no change.
Conclusion: qDESS was able to accurately measure T2 relaxation time before and after exercise in skeletal muscle compared to both MESE and STEAM spectroscopy. This opens the opportunity of using a fast 3D sequence to assess T2 in muscular dystrophies where wide anatomical coverage can be of benefit (such as FSHD) or during exercise challenges.
P2.09 MyoQMRI 2.0 - A comprehensive open-source pipeline for quantitative muscle imaging
Jessica Schäper1, Francesco Santini1,2,
Claudia Weidensteiner1
1Basel Muscle MRI, University of Basel, Basel, Switzerland, 2Department of Radiology, University Hospital of Basel, Basel, Switzerland
Background: The main quantitative imaging measures for neuromuscular diseases are water T2 (wT2), which indicates edema and acute disease activity, and the fat fraction, which quantifies muscle degeneration. While the most widely used methods are conceptually simple, changes in implementation details of the sequences and in the data analysis pipelines can decrease reproducibility.
Aims: The goal of this work was to develop a vendor-independent multi-echo spin-echo (MESE) sequence and create a standardized open-source pipeline called MyoQMRI 2.0 for a comprehensive quantitative muscle image evaluation, based on multiple tools for image generation, reconstruction, and postprocessing.
Methods: First, a MESE sequence was programmed in PyPulseq v1.4.2. Details about the sequence and a comparison with a vendor sequence on healthy volunteers can be found in [3]. Together with this sequence, several open-source tools were combined into a comprehensive, cohesive package to form a complete analysis pipeline for muscle data, containing:
the PyPulseq MESE sequence
the ormir-mids converter, which converts dicom data into a common input format an EPG-based wT2 fitting algorithm
a fat/water separation algorithm for two-echo GRE data, using hierarchical multi-resolution graph-cuts
The MESE sequence was tested on six healthy volunteers and compared with the vendor's implementation with the same parameters. The fat/water algorithm was tested on previously acquired 6-echo GRE data and compared with the FattyRiot reconstructions.
Results: The developed PyPulseq sequence yields similar images as the product sequence when processed with the MyoQMRI package. However, PyPulseq values (mean: 31±1ms) were slightly higher than Siemens (mean: 28±1ms) values. The 2-echo fat/water reconstruction yielded similar results to the 6-echo reconstruction. The package can be found at: https://github.com/BAMMri/MyoQMRI
Conclusion: The difference in T2 demonstrates that even for ostensibly equal parameters, different sequence implementations can yield different results. This makes it clear why a vendor-independent open-source approach is beneficial. The 2-echo fat/water sequence is more popular in clinical studies and yields similar results to the more complex 6-echo method.
P2.10 Quantitative muscle MRI to assess muscle tissue preservation during a 10-Day extended fast: A Single-case pilot study
Lara Schlaffke1,2,3, Johannes Forsting1, Johanna Thomä1, Lionel Butry1, Wiebke Fenkse4, Saif Al Basri4
1Department of Neurology, BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Heimer Institute for Muscle Research, Bochum, Germany, 4Department of Internal Medicine, Endocrinology & Diabetes and Gastroenterology & Hepatology, BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany
Background: Fasting induces profound metabolic and molecular adaptations that have been linked to increased lifespan and improved health outcomes in animal models. While the metabolic shifts during human fasting are increasingly well characterized, the preservation or loss of skeletal muscle tissue during prolonged fasting is a critical yet understudied aspect of metabolic adaptation.
Aim: Characterise the metabolic effects of a 10 -day fast on leg musculature using quantitative MRI.
Method: In this single-case pilot study, we investigated skeletal muscle changes during a 10-day (max 250 kcal/day) fast using quantitative muscle MRI (qMRI) as a non-invasive biomarker. MRI scans were conducted at baseline, and after refeeding to assess changes in muscle volume, composition, and signal characteristics related to tissue integrity and metabolism. The participant was instructed to perform 20 push-ups and 20 burpees per day to ensure minimal muscle activation. To complement imaging data, whole-body composition was evaluated using bioimpedance analysis (BIA), and energy expenditure was assessed through indirect calorimetry.
Results: According to the BIA analysis, body weight was reduced by 5.15kg (7%), of which 3.21kg was recognized as fat and 1.1kg as muscle mass. When analysing qMRI parameters, muscle volume in the thigh decreased from 121cm3 to 118cm3, Dixon fat fraction decreased from 2.8% ±1% to 2.6% ± 0.9%. FA slightly decreased (0.23 to 0.21) while RD slightly increased (1.26 to 1.32 ×10-3mm3/s).
Conclusion: This case provides initial insights into the preservation of muscle tissue during extended fasting with minimal muscle activation and demonstrates the feasibility of integrating qMRI with physiological measures to monitor fasting-induced tissue alterations. Our pilot findings provide important evidence that there is no risk of clinically relevant loss of muscle mass during fasting.
P2.11 Agreement between 3D ultrasound and DTI for assessing tibialis anterior muscle architecture - a pilot study
Lara Schlaffke1,2,5, Johannes Forsting1, Paolo Tecchio3, Brant J. Raiteri3,4, Daniel Hahn3,4
1BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Human Movement Science, Faculty of Sport Science, Ruhr University Bochum, Bochum, Germany, 4School of Human Movement and Nutrition Sciences, University of Queensland, Brisbane, Australia, 5Heimer Institute for Muscle Research, Bochum, Germany
Background: Accurate assessment of skeletal muscle architecture in vivo is critical for understanding muscle function, neuromuscular disorders, and rehabilitation outcomes. The tibialis anterior (TA) muscle is the major ankle dorsiflexor, and quantifying its muscle architecture and geometry could thus provide diagnostic and functional insights about walking independence. Conventional imaging techniques such as B-mode ultrasound are limited to two-dimensional views and operator-dependent alignment, while diffusion tensor imaging (DTI), though capable of estimating 3D fiber orientation, is costly and time consuming.
Aim: To validate 3D ultrasound for quantifying TA muscle architecture and aponeurosis geometry, and to compare fascicle length changes with those derived from DTI.
Methods: An automated 3D ultrasound (3DUS) method combined with DTI to quantify the entire architecture of the TA muscle at rest and during isometric contraction was examined in 4 healthy volunteers. Using a spatially tracked ultrasound transducer, volumetric muscle images across various ankle joint angles (−5°, 15°, and 35° plantarflexion) were acquired. DTI scans were performed on the same day using a 3T MRI scanner with tractography protocols for all 3 angles. The 3DUS data were reconstructed into isotropic voxel grids. A PCA-based coordinate system of the central aponeurosis mesh was applied to align and quantify fascicle length, pennation angle, and aponeurosis geometry. Muscle fascicles were identified algorithmically and automatically and reconstructed based on anatomical boundaries.
Results: Preliminary findings indicate strong agreement between 3DUS and DTI in fiber bundle/fiber length changes across joint angles and reveal notable inter- and intra-subject variability in aponeurosis geometry. The estimated lengths of the fiber bundles (3D ultrasound) and fibers (DTI) respectively increased by 2.2mm versus 5.6mm from -5° to 15° plantar flexion and by 7.4mm versus 7.6mm from -5° to 30° plantar flexion.
Conclusion: These data suggest that 3DUS is capable of accurately assessing TA muscle architecture non-invasively, with high spatial resolution over a large capture volume. This multimodal imaging framework (i.e. combining 3D ultrasound with DTI) provides a scalable, objective tool for evaluating muscle architecture in clinical and research settings. It offers a promising alternative for monitoring muscle adaptation, guiding rehabilitation strategies, and improving diagnostic accuracy for neuromuscular pathologies.
P2.12 Spatial heterogeneity of strain from dynamic magnetic resonance imaging using automated anatomically relevant partitioning of individual calf muscles
Saskya Soriano2, Gilliana Loyola2, Vadim Malis1,
Kendall Johnson2, Shantanu Sinha1, Usha Sinha2
1University of California San Diego, San Diego, United States, 2San Diego State University, San Diego, United States
Background: Intramuscular strain heterogeneity has been shown in several studies and is critical for understanding muscle coordination and force transmission. However, most analyses rely on averaging strain values over the entire muscle or limit analyses to one or more volumes of interest (VOI) or regions of interest (ROI), all of which can obscure spatial heterogeneity. With the potential to acquire data from the entire muscle volume using novel accelerated MRI flow sequences, it is important to develop methods to characterize the strain in sub-volumes of a muscle.
Aims: To develop a robust framework for anatomically aligned sub-volume segmentation of individual calf muscles and to extract strain indices from volumetric dynamic data on calf muscles.
Methods: 16 young subjects (11 M) were included in the study. Volumetric dynamic images of the calf at 45 and 60% Maximum Voluntary Contraction were acquired during isometric plantarflexion using a compressed sensing accelerated 4D Flow MRI sequence at 3T. Each calf muscle (medial gastrocnemius (MG), lateral gastrocnemius (LG), soleus (SOL), anterior tibialis (AT)) was manually segmented. We applied principal component analysis (PCA)-based method to the 3D coordinates of all nonzero voxels in the segmentation mask of each muscle to partition into 12 anatomically meaningful sub-volumes (3×2×2 grid aligned to the muscle's principal axes). The grid consisted of three bins along the first principal component (superior-inferior), two bins along the second (medial-lateral), and two along the third (anterior-posterior). This approach provides consistent anatomical labeling across subjects while adapting to individual muscle orientation and shape. Strain metrics for each sub-volume was calculated using displacement fields derived from the velocity data of the 4D Flow data.
Results: The MG and LG showed highest strain heterogeneity, especially in distal/lateral regions. This suggests regionally selective activation in the gastrocnemius during plantarflexion. SOL showed uniform strain and is likely fully recruited even at submaximal MVCs. TA, as antagonist, showed mild variation but no significant subregional differences.
Conclusion: This framework allows one to build a database of strain distributions in calf muscle sub-volumes under different contraction paradigms and supports future studies on strain heterogeneity in clinical populations.
P2.13 Body Mass Index related differences in strains and co-activation in calf muscles using compressed sensing accelerated 4D Flow Magnetic Resonance Imaging
Gilliana Loyola2, Usha Sinha2, Savannah Orrill2,
Adam Levay2, Vadim Malis1, Kendall Johnson2,
Shantanu Sinha
1
1University of California San Diego, San Diego, United States, 2San Diego State University, San Diego, United States
Background: Obesity can alter gait patterns and is a significant risk factor for developing osteoarthritis. Body Mass Index (BMI) is a widely used index that can reflect obesity.
Aim: To explore BMI related altered strain distribution in calf muscles from altered activation patterns as well as altered levels of coactivation between agonist and antagonist muscles.
Methods: 8 young subjects (6M) with a mean BMI of 23.7 were the control group (CG) while 8 young subjects (5M) with a mean BMI of 29.2 were the obese/overweight group (OG). We performed strain mapping of the calf muscles at three %Maximum Voluntary Contraction (30,45 and 60%) during isometric plantarflexion by volumetric dynamic imaging using a compressed sensing accelerated 4D Flow MRI sequence at 3T. Structural imaging included fat quantification as well as diffusion tensor imaging.
Results: No significant differences were found in muscle volumes or in fat fraction between CG and OG though both values were higher in the OG. Significant differences were found in compressive strains between 30 and 60% MVC with larger absolute values at higher %MVC. We also found significant differences in the compressive strain between the Medial Gastrocnemius (MG), Lateral Gastrocnemius (LG), and Soleus (SOL) with LG having the biggest strain values. There were significant (Muscle×Cohort) interactions in the compressive strain. In the CG, LG strain was significantly higher than MG while in the OG, significant differences in strain were found between all three plantarflexors with LG > MG > SOL. These strain differences may indicate subtle differences in gait patterns present before other compositional changes (like fat fraction). Significant differences between the mean of the plantarflexor (PF) and dorsiflexor (DF) muscles were seen in both CG and OG groups with PF having higher strain values. No significant (Muscle×Cohort) interactions were seen when comparing the PFs ad DFs though we hypothesized an increase in coactivation of DFs in OG; this increase may be observed in groups with much higher BMIs.
Conclusion: Dynamic volumetric strain mapping of calf muscles enables detection of altered strain in higher BMI cohorts.
P2.14 A simulation framework for dynamic phase-contrast MRI of the muscle
Maaike Smit1, Melissa Hooijmans2,3, Luuk Vos4,
Gustav Strijkers4, Hermien Kan1, Susanne Rauh1
1C.J. Gorter MRI Center, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, 2Department of Radiology and Nuclear Medicine, Amsterdam University Medical Center, Amsterdam, Netherlands, 3Department of Human Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, Netherlands, 4Department of Biomedical Engineering and Physics, Amsterdam University Medical Center, Amsterdam, Netherlands
Background: Time-resolved phase-contrast MRI (PC-MRI) is an important technique to study muscle contraction dynamically. It is particularly interesting for investigating neuromuscular diseases, where the decrease in muscle strength cannot fully be explained by the loss of muscle mass. Pathophysiological factors, like fat and inflammation, and scan settings affect the PC-measurements. Disentangling these influences is virtually impossible in vivo and in silico, but can be studied using simulations.
Aim: We aimed to develop a simulation framework for PC-MRI to optimize sequence parameters and to study the effect of pathophysiological changes on scan outcomes in muscle.
Methods: A simulation framework using the Bloch equations was developed in Python. The evolution of the MR signal was simulated during a PC-MRI pulse sequence containing two rectangular bipolar gradients (flip angle = 8°, t_RF = 0.45 ms, G_z = 14.82 mT/mm, t_G = 2.20 ms). Simulations were performed for a 1D segment (L = 30 mm, N_pix = 101). To simulate a rigidly moving muscle, the segment was displaced with 25 different velocities (range -6.0 to 6.0 cm/s, step size of 0.5 cm/s). Using the phase change at the end of the sequence (ΔΦ) and the velocity encoding parameter (VENC), the velocity was calculated as: v = ΔΦ*VENC/π. The percentage error between the calculated and the predefined velocity was determined for verification.
Results: When velocity was applied to the 1D segment, a change in phase was observed at the end of the sequence. As expected, a net phase change of zero was found when no velocity was applied. The phase change increased with increasing velocity. The calculated velocities corresponded with the predefined velocities, with an error varying from 2.36*10^-8 to 1.13*10^-6%.
Conclusion: We successfully developed a simulation framework for time-resolved PC-MRI using Bloch simulations. As a next step, the framework will be extended to 3D, allowing the modeling of a velocity vector with components in multiple spatial dimensions. Future work also includes incorporating fat replacement in the simulated muscle and studying the effect of specific sequence parameters on the signal.
P2.15 Evaluation of quantitative muscle MRI and intelligent phenotyping housing system as advanced and objective phenotyping methods in a mouse model of calpainopathy
Nicolina Südkamp1,3, Marlena Rohm1,2, Gabriele Russo4, Xavier Helluy4,5, Abdulhadi Kocabas1,3, Martijn Froeling6, Denise Manahan-Vaughan4, Frank Jacobsen1,3,
Matthias Vorgerd1,3, Johannes Forsting1,
Lara Schlaffke1,2,3
1BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Heimer Institute for Muscle Research, Bochum, Germany, 4Department of Neurophysiology, Ruhr-University Bochum, Bochum, Germany, 5Department of Biopsychology, Institute of Cognitive Neuroscience, Ruhr-University Bochum, Bochum, Germany, 6Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Calpainopathy is a rare genetic muscle disorder without causal treatment available, but recently many different treatment options, including genetic treatment and immunomodulation have been developed and are currently tested pre-clinically in murine models. Traditional behavioral assays frequently fail to detect early motor deficits in corresponding mouse models.
Aim: This study investigates whether quantitative magnetic resonance imaging (qMRI) and automated motor assessments can reveal early pathological changes in a mouse model of calpainopathy, particularly during presymptomatic stages.
Methods: A standardized qMRI protocol, previously validated in a Pompe disease model, was applied to Calpain 3-knockout mice (Capn3 KO) and wild-type controls at 5 and 15 months of age. Fat fraction, water T2 relaxation time, and diffusion tensor imaging parameters were analyzed across selected hindlimb muscles. Additionally, voluntary motor activity was monitored using an intelligent phenotyping housing system.
Results: At 5 months, no significant differences were detected between genotypes. At 15 months, Capn3 KO mice exhibited elevated water T2 in the soleus and gastrocnemius muscles, corresponding with immune cell infiltration, suggesting early muscle inflammation. Diffusion parameters and fat fraction remained unchanged, and motor activity assays revealed minimal and inconsistent genotype effects. Gender-based differences in qMRI metrics were significant, influencing fat distribution and T2 values.
Conclusion: Muscle qMRI, particularly T2 mapping, can sensitively detect subclinical muscle changes in aging Capn3 KO mice, potentially identifying early inflammatory processes. Gender differences and variability underscore the need for evaluation of both genders separately when investigating muscle structure. Extending observations beyond 15 months may improve detection of later-stage pathology in this mouse model of calpainopathy.
P2.16 Cross-manufacturer comparison of quantitative muscle MRI in healthy volunteers
Johanna Thomä1, Lionel Butry1, Johannes Forsting1, Martijn Froeling4, Lara Schlaffke1,2,3
1BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Heimer Institute for Muscle Research, Bochum, Germany, 4Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Quantitative muscle magnetic resonance imaging (qMRI) is a promising method for detecting muscular changes, particularly in neuromuscular diseases (NMDs). Cross-manufacturer reproducibility plays a crucial role in distinguishing pathological from physiological features and in ensuring data comparability in multi-center studies on rare NMDs. However, reproducibility and manufacturer-related measurement variability have so far been insufficiently studied, especially regarding the water T2 relaxation time (wT2).
Aim: The aim of this study is to evaluate the cross-manufacturer reproducibility of qMRI parameters in a healthy population.
Methods: Ten healthy volunteers were examined within one week using two 3T magnetic resonance scanners from the manufacturers Siemens and Philips. Quantitative assessment of 12 leg muscles was performed muscle wise using Dixon technique to determine fat fraction (FF), T2 mapping to measure water T2 relaxation time (wT2), and diffusion-weighted sequences for diffusion parameters: fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), and axial diffusivity (λ1). Intraclass correlation coefficients (ICC(2,k)), coefficients of variation, paired t-tests, and Bland-Altman plots were generated to assess reproducibility and comparability.
Results: The ICCs demonstrated moderate to excellent reproducibility across all muscles for most parameters (FF = 0.923, wT2 = 0.798, FA = 0.771, MD = 0.708, RD = 0.850), except of λ1 (ICC = 0.361). Significant differences were observed between the manufacturers for all parameters (p < 0.05). The Bland-Altman analysis revealed systematic biases ranging from 44.2% for FF to 3.6% for FA. In the muscle-specific analysis, the highest coefficients of variation were found in the rectus femoris (ranging from 43% for FF to 10% for λ1).
Conclusion: The study demonstrates that qMRI parameters are generally reproducible in healthy volunteers. However, the absolute measurement values are not directly comparable between manufacturers and require correction. These differences should be considered particularly in multi-center studies, to minimize manufacturer-related discrepancies and ensure comparability of results.
P2.17 Development of a quantitative muscle ultrasound protocol for murine models of neuromuscular disorders
Jeroen van Doorn1, Donnie Cameron2, Nens van Alfen1
1Department of Neurology, Clinical Neuromuscular Imaging Group, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, Netherlands, 2Department of Medical Imaging, Clinical Neuromuscular Imaging Group, Radboud University Medical Center, Nijmegen, Netherlands
Background: Quantitative muscle ultrasound (QMUS) and MRI enable non-invasive assessment of skeletal muscle tissue for screening and follow-up of neuromuscular disorders (NMDs). Echogenicity as obtained with QMUS is a human imaging biomarker investigated for use in drug-evaluating trials as a muscle biopsy substitute. Assessment of preclinical NMD treatment effects still depends on muscle biopsies and animal sacrifice. Surprisingly, there is currently no protocol for QMUS in murine models of NMDs. Such a protocol could aid in preclinical experiment design, reducing the number of animals, as they would not need to be euthanised during the study. Further, comparing imaging results could help translate preclinical studies to clinical drug-evaluating trials. In this project we have taken the first step to standardised murine muscle imaging.
Aim: To develop a hindlimb muscle QMUS scanning protocol in mice.
Methods: A convenience sample of surplus mice were scanned directly after scheduled euthanisation using an ultrahigh frequency ultrasound device. Three positions were scanned in the left hindleg: the lower leg and thigh compartments, and the hip region with the gluteus maximus and gluteus medius muscles. Mice were positioned on their right side in a preclinical MRI animal cradle to enable future comparisons with MRI. Exploratory statistics were used for associations with strain, age, and weight.
Results: 50 mice from 8 different strains were scanned. All muscle groups in the lower leg and hip regions could be visualised and analysed, whereas image quality in the thigh was insufficient in 30% of the scans. Echogenicity differed between strains. There was a negative correlation between echogenicity and age in the deep and posterior muscle groups of the lower leg, and a negative correlation between echogenicity and weight in all muscle groups of the lower leg and thigh.
Conclusion: QMUS of murine hindlimb muscles is feasible in the lower leg and hip, but more difficult in thigh region. Echogenicity depends on mouse strain, age, and weight; therefore, a matched control group should be included in any experiment. Alternatively, normative values could be collected in a cohort of the same strain that will be used in future experiments.
P2.18 Respiratory muscle shear wave elastography to assess respiratory muscle function in congenital myopathies
Jeroen van Doorn1, Dieke Jans1, Chris de Korte2,3,
Baziel van Engelen4, Nicol Voermans4, Nens van Alfen1, Coen Ottenheijm5, Damien Bachasson6,
Jonne Doorduin7
1Department of Neurology, Clinical Neuromuscular Imaging Group, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, Netherlands, 2Medical Ultrasound Imaging Centre, Department of Medical Imaging, Radboud University Medical Centre, Nijmegen, Netherlands, 3Physics of Fluids Group, TechMed Centre, University of Twente, Enschede, Netherlands, 4Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, Netherlands, 5Department of Physiology, Amsterdam UMC, Location VUmc, Amsterdam, Netherlands, 6Sorbonne Université, INSERM, UMRS1158 Neurophysiologie Respiratoire Expérimentale et Clinique, Paris, France, 7Department of Intensive Care Medicine, Radboud University Medical Center, Nijmegen, Netherlands
Background: Respiratory muscle dysfunction is prevalent in congenital myopathies and often leads to respiratory failure, highlighting the importance of periodic monitoring and early detection. Transdiaphragmatic pressure (Pdi) is the gold standard for measuring diaphragm contractility, but the invasive nature of this test impedes its widespread use, and accessory respiratory muscles are not assessed. Ultrasound shear wave elastography (SWE) measures the elastic modulus of tissues and has also been proposed as a non-invasive alternative for assessing contractility of respiratory muscles, including the diaphragm.
Aim: to evaluate the use of SWE to assess respiratory muscle function in both healthy individuals and patients with a congenital myopathy, while also determining the reliability of SWE as a potential clinical and research tool for monitoring respiratory dysfunction.
Methods: Shear wave velocity of the diaphragm, parasternal intercostal, sternocleidomastoid, external oblique abdominal, internal oblique abdominal, and transverse abdominal muscles was acquired in healthy participants and patients with a congenital myopathy during quiet breathing, deep inspiration or expiration and (sub)maximal isometric efforts. Mouth pressure (Pmo) was acquired simultaneously with SWE acquisition. Furthermore, Pdi was acquired but only in healthy participants. All measurements were repeated by a first observer and a second observer to assess intra-observer, inter-observer and test-retest reliability.
Results: 20 healthy participants (age: 22.5y [20.5-25.5], 8 males) and 21 patients with a congenital myopathy (age: 34.0y [22.0-42.0], 11 males) were included. Changes in Pmo and Pdi were not reflected by changes in shear wave velocity (Pmo: from -23.0±2.81 cmH2O to -60.4±4.01 cmH2O, p<0.001, Pdi: from 27.5±14.1 cmH2O to 65.7±27.2 cmH2O, p=0.001, and shear wave velocity: from 2.29±0.65 m/s to 2.57±1.11 ms/s, p=0.179). Shear wave velocity did not differ between healthy participants and patients. Reliability estimates were, on average, poor.
Conclusion: The absence of differences in measured shear wave velocity between healthy participants and patients, and the low reliability estimates, indicate that SWE of the diaphragm and other respiratory muscles is not suitable for monitoring respiratory dysfunction in routine clinical care or research. The complex mechanical properties of the respiratory muscles, further amplified in congenital myopathies, introduce challenges that may require alternative SWE implementations.
P2.19 Spatiotemporal relationship between hamstring muscle activation and strain rate during dynamic knee flexion: A combined multi-channel electromyography and 3D time-resolved phase contrast study
Luuk Vos1, Dorus Colijn2, Susanne S. Rauh3,
Maaike S. Smit3, Hermien E. Kan3, Gustav J. Strijkers1, Melissa T Hooijmans2,4
1Amsterdam Umc, Amsterdam, Netherlands, 2Faculty of Behavioral and Movement Sciences, Vrije Universiteit, Amsterdam, Netherlands, 3C.J. Gorter Center for High Field MRI, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, 4Department of Radiology & Nuclear Medicine, Amsterdam UMC, Amsterdam, Netherlands
Background: Electromyography (EMG) is considered the gold standard in evaluating muscle function, typically measured on the surface of the skin. The measured signal represents the summation of activated motor units and represents only a small area of the muscle. MRI-based methods such as velocity-encoded phase contrast (VE-PC) imaging provide insights into three-dimensional muscle deformation. The relationship between EMG and VE-PC is relatively unknown.
Aims: The goal of this study was to investigate the spatiotemporal relationship between electric activation and muscle velocity in the hamstring muscles during dynamic knee flexion.
Methods: Six healthy volunteers (25.0 ± 2.9 years, 2f) participated in this study and visited the hospital twice. During the first visit, the activation patterns were determined in the m. biceps femoris long head (BF), m. semimembranosus (SM) and semitendinosus (ST)) in the dominant leg using multichannel EMG. Participants performed repeated knee flexion exercise with a 3-second cycle at 15% of the maximal voluntary contraction force. During the second visit the participants performed the same knee flexion exercise in a 3T MRI scanner. The MR protocol consisted of a static 3D gradient echo mDixon scan for anatomical reference and a 3D time-resolved VE-PC acquisition (VENC 10cm/s; 30 time-points, voxel size=3×3×6mm3, scan duration=502s; pseudo-spiral undersampling factor 10). Analyses were performed between 15-60% which is the relevant part of the contraction cycle. Differences in mean activation and mean velocity between electrode locations within muscles were analysed using a Friedman test. Spearman correlations were used to determine the relation between EMG and VE-PC.
Results: Visually, there seems to be a proximo-distal distribution in VE-PC velocity, but not in muscle activation. However, no differences in mean activation and mean velocity between electrode location of all muscles were found. Muscle activation was negatively correlated with muscle velocity (BF (r=-0.66, p<0.001), ST (r=-0.41, p<0.001) and SM (r=-0.65, p<0.001)).
Conclusion: Current study might be underpowered to show differences between locations. Negative correlations between muscle activation and velocity indicate that muscles shorten when activated and elongate with relaxation which is in line with muscle physiology. Future studies will include more participants and advanced strain rate analysis.
P2.20 Correlations between muscle fat fraction MRI and instrumented gait assessments in Dysferlinopathy patients
Ian Wilson1, Lisa Alcock1, Meredith James2,
Heather Hilsden2, Karen Wong2, Mark Richardson3, Emma-Jayne Robinson2, Emma Grover2, Philip Brown3, Carla Bolano-Diaz2, Kieren Hollingsworth1,
Peter Thelwall1, Jordi Diaz-Manera2, Andrew Blamire1, Volker Straub2
1Translational and Clinical Research Institute (Newcastle University), Newcastle Upon Tyne, United Kingdom, 2John Walton Muscular Dystrophy Research Centre (Newcastle University), Newcastle upon Tyne, United Kingdom, 3The Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, United Kingdom
Background: Fat replacement in muscles measured by Dixon MRI provides an excellent marker for disease progression in dysferlinopathy patients. Disease-specific clinical scales are used to evaluate motor performance. Instrumented gait analysis provides objective, quantifiable outcomes which may be useful for monitoring disease progression and evaluating therapeutic response.
Aim: This study aimed to establish if there were any correlations between the clinical outcome measures of MRI and gait analysis.
Methods: Patients with genetically confirmed dysferlinopathy were assessed as part of the Jain Foundation funded Clinical Outcomes Study. 19 participants were assessed (age range 23-75y, 10 females, number of years since symptom onset range: 2-33 years). All patients completed the following assessments during a single visit:
Muscle MRI: Dixon MRI data were collected from the thigh and leg muscles. Mean fat fraction (FF) values were calculated and contractile cross-sectional area (cCSA) was determined.
Gait: Using an instrumented walkway, gait was evaluated while patients walked at a self-selected preferred pace. Up to six 10-m intermittent walks were collected. 8 temporal-spatial features of gait were averaged across all steps.
NSAD: The North Star Assessment for limb-girdle type muscular dystrophies (NSAD) is a 29-item scale of motor performance, with lower scores indicating reduced motor performance. *
Results: Moderate to strong correlations were found between fat fraction and cCSA measurements in the thighs and calves with gait measurements of stride length, velocity, step time, stance time, swing time and single leg support. (rho values ranged from 0.460 to 0.783, P = < 0.05). Strong correlations were also found between NSAD scores and Thigh MRI (rho >0.7, P < 0.05). Moderate to strong correlations were also found for gait and MRI measurements and time from symptom onset in patients. (rho 0.460-0.783, P < 0.05).
Conclusions: These results provide further evidence that MRI measurements of fat replacement in muscles of the lower limb closely correlate with changes in motor performance and locomotor control in dysferlinopathy patients. Further studies will investigate the role of individual muscles in the loss of gait functions.
P2.21 Blood flow restriction training induced morphology changes in M. quadriceps femoris – a prospective pilot study
Lionel Butry1, Alice De Lorenzo1, Johannes Forsting1, Martijn Froeling4, Arthur Praetorius45,6,
Christian Raeder5,6, Robert Rehmann1,3, Tobias Ruck1,3, Christian Schoepp5,6, Janina Tennler5,6, Lara Schlaffke1,2,3
1BG Universitätsklinikum Bergmannsheil gGmbH, Bochum, Germany, 2Medical Engineering, FH Dortmund - University of applied science and arts, Dortmund, Germany, 3Heimer Institute for Muscle Research, Bochum, Germany, 4Center for Image Sciences, Precision Imaging Group, Division Imaging & Oncology, University Medical Centre Utrecht, Utrecht, Germany, 5Athletikum Rhein Ruhr, Department of Arthroscopic Surgery, Sports Traumatology and Sports Medicine, BG Hospital Duisburg, Duisburg, Germany, 6Department of Arthroscopic Surgery, Sports Traumatology and Sports Medicine, BG Hospital Duisburg, Duisburg, Germany
Background: Low-load blood flow restriction (BFR) resistance training is an innovative training method with similar increases in strength as traditional high-load resistance training. While immediate physiological changes have been described, biomarkers for BFR resistance training induced muscle morphology changes are needed to assess its safety and efficacy in healthy and patient populations.
Aim: Identification of quantitative magnetic resonance imaging (qMRI) biomarkers of BFR resistance training.
Methods: A prospective pilot study with 3 participants was conducted. Each participant received a 5-week knee extension training program including 3 sessions per week with 6 sets until muscle failure at 40% repetition maximum. One leg was trained with BFR (60% limb occlusion pressure), the other leg without BFR. At baseline and at the end of the training program, qMRI of a 15 cm section of the thighs was acquired including T2 mapping to assess for oedematous tissue alterations, DIXON to assess intramuscular fatfracton and diffusion-weighted imaging to assess the muscle architecture, and processed using QMRITools.
Results: The volume-equated mean total mechanical training load was 16676±2289 kg for the BFR and 16913±2775 kg for the non-BFR leg. Change from baseline in water T2 relaxation time (BFR: 0.33 ms [0.5%], non-BFR: 0.34 ms [0.9%]) and fat fraction (BFR: 0.09% [4%], non-BFR: -0.19% [-7%]) were within the measurement error. An increase in muscle volume (BFR: 421 ccm [31%], non-BFR: 430 ccm [30%]) and a decrease in fractional anisotropy (BFR: -0.01 [-5%], non-BFR: -0.008 [4%]) with stable mean diffusivity (BFR: -0.0004×10-3 mm2/s [-0.03%], non-BFR: -0.01×10-3 mm2/s [-0.8%]) and radial diffusivity (BFR: 0.005×10-3 mm2/s [0.4%], non-BFR: -0.003×10-3 mm2/s [0.2%]) were observed.
Conclusion: The BFR and non-BFR leg showed similar signs of hypertrophy in the M. quadriceps femoris without inflammatory processes, indicating a safe and efficient use of low-load BFR resistance training in future studies and rehabilitation programs. To evaluate the efficacy of BFR resistance training, a sham-BFR group will be included in the full study. Adaptions to patients with hereditary or inflammatory neuromuscular disorders need to be evaluated carefully
Author Index
A
Adamová, Blanka S12
Ahmed, Shakeel S26
Al Basri, Saif S28
Alcock, Lisa S32
Almendrote, Miriam S18
Attarian, Shahram S11, S16
B
Bacardit, Jaume S10
Bachasson, Damien S31
Baetscher, Eric S13, S26
Bagga, Puneet S9
Bamps, Kobe S22, S24
Barnard, Alison M. S13, S26
Bartels, Bart S19, S26
Barthelemy, Inès S8
Barzaghi, Leonardo S10, S20
Baudin, Pierre-Yves S15, S21, S24, S26
Baxter, Gabrielle S18
Becks, Marinus J. S26
Bellam, Prathyusha S13, S26
Bellemare, Marc-EmmanuelS S11
Bendahan, David S11, S16
Bergsland, Niels S10, S20
Berling, Edouard S17
Berry, David S17, S18
Bettolo, Chiara S12
Bilfeld, Marie Faruch S5
Birkbeck, Matthew S12
Blamire, Andrew S6, S12, S32
Blot, Stéphane S8
Bolaño-Díaz, Carla S10, S32
Bonizzoni, Chiara S10, S20
Brenninkmeijer, Roosmarijn S19, S26
Brown, Rory S27
Brown, Philip S32
Burman, Ritambhar S9
Butry, Lionel S28, S30, S33
C
Cameron, Donnie S3, S26, S31
Carlier, Robert S17
Chahin, Nizar S13
Chatziandreou, Petros S11
Claeys, Kristl S22, S24
Colijn, Dorus S32
Croce, Michele Giovanni S10, S20
D
Díaz-Manera, Jordi S10
Dörfler, Arnd S24
Dalmau, Jacob Sanchez S18
de Almeida Araujo, Ericky Caldas S8
de Korte, Chris S14, S31
de la Iglesia Vaya, Maria S18
De Lorenzo, Alice S12, S18, S33
De Ridder, Willem S5
Deligianni, Xeni S10, S20
Delmont, Emilien S11, S16
Diaz-Manera, Jordi S32
Doorduin, Jonne S14, S31
Dostál, Marek S12
Dupont, Patrick S22, S24
E
Enax-Krumova, Elena S11, S12, S18, S19
Erasmus, Corrie E. S26
Estaffio, Fadi S8
F
Fenkse, Wiebke S28
Fieremans, Els S16
Figueroa Bonaparte, Sebastian Ariel S18
Filosto, Massimiliano S10, S20
Finkel, Richard S9
Fisch, Robert S17
Fisse, Anna Lena S11
Fitzsimmons, Sam S10
Fleerakkers, Evelien S14, S26
Forbes, Sean S13, S26
Forsting, Johannes S11, S12, S18, S19, S28, S30, S33
Fortanier, Etienne S11, S16
Fox, Brian S14
Frank, Lawrence S17, S18
Froeling, Martijn S6, S11, S12, S13, S18, S19, S20, S21, S26, S30, S33
Fromes, Yves S8, S15
G
Güttsches, Anne-Katrin S12, S18, S19
Gaardner, Jascha S18
Galinsky, Vitaly S17, S18
Gandini Wheeler-Kingshott, Claudia Angela Michela S10, S20
Gaspar, Andreia Sofia S4
Gasperini, Serena S10, S20
Gazzerro, Elisabetta S24
Gerevini, Simonetta S7
Gerhalter, Teresa S7, S24
Ghouth, Salim Bin S15
Goedee, Stephan S13, S21
Gold, Garry S23
Gonzalez Chamorro, Alejandro S10
Goosens, Veerle S22, S24
Gordon III, Joseph A. S17
Grandis, Marina S10
Grandis, Marina S20
Grover, Emma S32
Guémy, Clément S17
Guicciardi, Giulia S20
Guye, Maxime S11
H
Hahn, Daniel S28
Hall, Julie S12
Hao, Longdan S10
Helluy, Xavier S30
Henriet, Valentin S21
Heskamp, Linda S13, S19, S20, S21, S26
Higgins, David M. S27
Hilsden, Heather S32
Hnilicova, Petra S23
Hollingsworth, Kieren G. S27
Hollingsworth, Kieren S32
Hooijmans, Melissa S9, S22, S30
Hooijmans, Melissa T S32
Hostin, Marc-Adrien S11, S16
Hughes, SamuelS S9
Huysmans, Lotte S22, S24
I
Iterbeke, Louise S22
J
Jacobsen, Frank S30
James, Meredith S32
Jans, Dieke S31
Jeneson, Jeroen S9, S22
Johnson, Tyler S14
Johnson, Kendall S29
Jones, Carly S23
Just, Ivica S23
K
Kan, Hermien S14, S30
Kan, Hermien E. S26, S32
Katzberg, Hans S8
Khan, Sono S8
Klepochova, Radka S23
Klimas, Rafael S10
Kneifel, Moritz S18
Kocabas, Abdulhadi S12, S30
Kokošová, Viktória S12
Kolísek, Martin S23
Koopman, Fieke S9
Krššák, Martin S23
Kraus, Emily S23
Krkoška, Peter S12
Kruit, Mark C. S26
L
Laforêt, Pascal S17
Lane, Adam S17
Lapert, Marc S15, S21
Leung, Doris S7
Levay, Adam S29
Li, Yuguo S14
Lott, Donovan J. S13, S26
Loyola, Gilliana S29
Lucente, Giuseppe S18
M
Madani, Afarin S2
Madrid Fuentes, Diana A. S25
Maes, Frederik S22, S24
Maggi, Lorenzo S10, S20
Maggioni, Marta Brigid S16, S25
Mahoney, Jeannette R. S25
Malis, Vadim S29
Manahan-Vaughan, Denise S30
Marty, Benjamin S8, S15, S21, S26
Matulová, Kateřina S12
Mayer, Anna-Lena S24
Mayos, Eduard Juanola S18
Mazzoli, Valentina S15, S16
Mehta, Nickita S14
Meier-Ross, Kaitlynn S8
Melkus, Gerd S8
Mensch, Alexander S5
Michaëls, Michel S14
Michel, Constance. P S11, S16
Mileder, Yvonne S24
Molleda, María Mercedes S18
Mongini, Tiziana Enrica S10, S20
Moore, Candace S14
Motte, Jeremias S11
N
Nagel, Armin S24
Naz, Farah S10
Nederveen, Aart S9
Nicolas, Guillaume S17
Nicolosi, Silvia S20
Nievelstein, Rutger A.J. S26
Niks, Erik S14
Niks, Erik H. S26
O
on behalf of the Myo-Guide Consortium S10
Oonk, Karleen S13, S21
Opsomer, Matthias S24
Orrill, Savannah S29
Ottenheijm, Coen S14, S31
Ovesná, Petra S12
P
Paoletti, Matteo S10, S20
Peeters, Ronald S22, S24
Pelsma, Maaike S26
Piñeiro, Alicia Martinez S18
Pichiecchio, Anna S10, S20
Pitarokoili, Kalliopi S9
Praetorius, Arthur S33
Prigent, Hélène S17
Puš, Katarina S25
R
Räuber, Sabine Melanie 16, S25
Raaphorst, Joost S9
Raeder, Christian S33
Raiteri, Brant J. S28
Rao, Smita S16
Rauh, Susanna S14
Rauh, Susi S24, S26
Rauh, Susanne S28
Rauh, Susanne S. S32
Ravaglia, Sabrina S10, S20
Rehmann, Robert S12, S18, S19, S33
Reyngoudt, Harmen S8, S21
Richardson, Mark S32
Rigler, Louis S26
Robinson, Emma-Jayne S32
Rock, Kelly S13, S26
Roemer, Frank S24
Rohm, Marlena S12, S30
Rooney, William S13, S26
Rossetto, Gabriel S27
Rouyer, Alice S17
Rubio, Gemma Montè S18
Ruck, Tobias S33
Russo, Gabriele S30
S
Südkamp, Nicolina S12, S30
Sacchini, Michele S10, S20
Sampaio, Marcos L S8
Santini, Francesco S4, S10, S16, S20, S25, S28
Sarkar, Tanay S8
Schänzer, Anne S5
Schäper, Jessica S28
Schlaffke, Lara S5, S11, S12, S18, S19, S28, S30, S33
Schoepp, Christian S33
Schofield, Ian S12
Schunk, Valentina S24
Sciacco, Monica S10, S20
Sechi, Annalisa S10, S20
Senesac, Claudia S13, S26
Servais, Aude S17
Šimunič, Boštjan S25
Sinha, Shantanu S29
Sinha, Usha S29
Sinha, Shantanu S29
Sleutjes, Boudewijn S13, S21
Slioussarenko, Constantin S15
Smit, Maaike S30
Smit, Maaike S. S32
Smith, Ian C. S8
Sole-Cruz, Eva S16
Soriano, Saskya S29
Starke, Anika S24
Straub, Volker S10, 12, S32
Strijkers, Gustav S9, S30
Strijkers, Gustav J. S32
Subramony, Sub S13
T
Türk, Matthias S24
Tasca, Giorgio S10
Tecchio, Paolo S28
Tennler, Janina S33
Thelwall, Peter S32
Thomä, Johanna S11, S28, S30
Timm, Derek S14
Tkotz, Katharina S24
Tromp, Selma C. S26
Turčanová, Monika Koprušáková S23
U
Uder, Michael S24
V
van Alfen, Nens S26, S31
van den Berg, Leonard S13, S21
van den Berg, Sandra S22
van der Holst, Menno S14
van der Pol, W. Ludo S26
van der Woude, Danny R. S19
van Doorn, Jeroen S14, S31
van Doorn, Jeroen L.M. S26
van Doremalen, Joris S14
van Engelen, Baziel S31
van Zwet, Erik S14
Vandenborne, Krista S13, S26
Venturelli, Nadia S17
Verdu Diaz, Jose S10
Verschueren, Annie S16
Vigueras, Anna Martinez S18
Vlažná, Daniela S12
Voermans, Nicole S9
Voermans, Nicol S14, S31
Voorn, Eric S9
Vorgerd, Matthias S12, S18, S30
Vos, Luuk S30, S32
W
Wächter, Marian S12
Walter, Glenn S13, S26
Wang, Fengdan S7
Ward, Samuel R S17
Ward, Samuel S18
Warman-Chardon, Jodi S8
Weaver, Ashley A. S25
Weidensteiner, Claudia S28
Whittaker, Roger S12
Willcocks, Rebecca J. S13, S26
Wilson, Ian S12, S32
Wong, Karen S32
Wooten, Lauren S9
Y
Yadav, Nirbhay S14
Ye, Johnny S17
Z
Zaiss, Moritz S24
Zeng, Qing S14
Zilke, Kirsten S13
