Journal of Neurology (2023) 270:960–974 https://doi.org/10.1007/s00415-022-11447-7 ORIGINAL COMMUNICATION Muscle MRI in immune‑mediated necrotizing myopathy (IMNM): implications for clinical management and treatment strategies Laura Fionda1 · Antonio Lauletta1 · Luca Leonardi1 · Jorge Alonso Perez2 · Stefania Morino1 · Gioia Merlonghi1 · Girolamo Alfieri1 · Rocco Costanzo1 · Laura Tufano1 · Fiammetta Vanoli1,3 · Elena Rossini1 · Eduard Gallardo Vigo2 · Tommaso Tartaglione4 · Marco Salvetti1 · Giovanni Antonini1 · Jordi Diaz‑Manera2,5,6 · Matteo Garibaldi1 Received: 13 July 2022 / Revised: 20 October 2022 / Accepted: 21 October 2022 / Published online: 3 November 2022 © The Author(s) 2022 Abstract Objectives Immune-mediated necrotizing myopathy (IMNM) is the most severe idiopathic inflammatory myopathy (IIM) and early aggressive poly-immunotherapy is often required to reduce long-term disability. The aim of this study is to investigate muscle MRI in IMNM as outcome measure for disease activity, severity, progression, response to treatment, and to better characterize the pattern of muscle involvement. Methods This is a retrospective, observational, cross-sectional, and longitudinal study including 22 IMNM patients, divided into three groups based on timing of first MRI and if performed before or under treatment. T1 score and percentage of STIR positive muscles (STIR%) were considered and analyzed also in relation to demographic, clinical and laboratory characteristics. Results STIR% was higher in untreated patients and in those who performed MRI earlier (p = 0.001). Pelvic girdle and thighs were in general more affected than legs. T1 score was higher in patients with MRI performed later in disease course (p = 0.004) with a prevalent involvement of the lumbar paraspinal muscles, gluteus medius and minimus, adductor magnus and hamstrings. 22% of STIR positive muscles showed fat replacement progression at second MRI. Higher STIR% at baseline correlated with higher risk of fat replacement at follow-up (p = 0.003); higher T1 score correlated with clinical disability at follow-up, with late treatment start and delayed treatment with IVIG (p = 0.03). Interpretation Muscle MRI is a sensitive biomarker for monitoring disease activity and therapy response, especially when performed early in disease course and before treatment start, and could represent a supportive outcome measure and early prognostic index in IMNM. Keywords Immune mediated necrotizing myopathy (IMNM) · Whole body muscle MRI · Inflammatory myopathies · Immunomodulating therapy · Follow-up study 3 * Matteo Garibaldi IRCCS Istituto Neurologico Carlo Besta, Milan, Italy matteo.garibaldi@uniroma1.it 4 Department of Radiology, Istituto Dermopatico 1 dell’Immacolata, IRCCS, 00167 Rome, Italy Neuromuscular and Rare Disease Centre, Department 5 of Neuroscience, Mental Health and Sensory Organs John Walton Muscular Dystrophy Research Centre, (NESMOS), SAPIENZA University of Rome, Sant’Andrea Translational and Clinical Research Institute, Newcastle Hospital, 00189 Rome, Italy University, Newcastle Hospitals NHS Foundation Trust, 2 Newcastle Upon Tyne NE1 3BZ, UK Neurology Department, Neuromuscular Disorders Unit, 6 Universitat Autònoma de Barcelona, Hospital de la Santa Centro de Investigación Biomédica en Red en Enfermedades Creu I Sant Pau, 08041 Barcelona, Spain Raras (CIBERER), 08041 Barcelona, Spain 13 Vol:.(1234567890) Journal of Neurology (2023) 270:960–974 961 Introduction disease severity, progression, and treatment response and to better characterize the pattern of the muscle involve- Immune-mediated necrotizing myopathy (IMNM) is a spe- ment in IMNM. cific nosological entity of idiopathic inflammatory myopa- thy (IIM) characterized by rapidly progressive muscular weakness with marked increase of serum creatine kinase Materials and methods (CK) and prominent myofiber necrosis and regeneration with mild or absent inflammatory infiltrates at muscle Patients biopsy [1, 2]. Two different antibodies have been associ- ated to IMNM: the anti-signal recognition particle (SRP) This is a retrospective, observational, cross-sectional, and and the anti-hydroxy-3-methylglutaryl-CoA reductase longitudinal study including IMNM patients followed at (HMGCR), that account for about two-thirds of IMNM Sant’Andrea Hospital of Sapienza University of Rome and patients [3]. IMNM, may be triggered by statin exposure at Hospital de la Santa Creu i Sant Pau of UAB University in about two-thirds of HMGCR positive cases [4]. Con- of Barcelona. Patients’ informed consent to MRI, which versely, anti-SRP patients usually show severe clinical includes a statement about image preservation for research, presentation, with higher frequency of cardiac involve- was obtained from all participants prior to performing MRI, ment, extramuscular manifestations, and malignancy [5]. in compliance with ethical standards of local ethical com- A third group of seronegative IMNM (anti-SRP and anti- mittees, the Helsinki Declaration, and the Good Clinical HMGCR negative) represents about one-third of patients Practice. and shares distinctive features, including female predom- All patients received a diagnosis of IMNM accordingly inance, higher frequency of connective tissue disorders to the 224th European Neuromuscular Centre (ENMC) and malignancy [6]. Treatment strategies include conven- international workshop on IMNM [3], in particular muscle tional first-line treatments (corticosteroids), but frequently biopsy required the presence of necrosis and regeneration require second- (and/or immunosuppressive agents) or at different stages, poor macrophagic-prevalent inflamma- eventually third-line (IVIG or Rituximab) immunother- tory infiltrates, variable and faint expression of MHC-I and apy to achieve the best clinical outcome [7]. IMNM rep- c5b-9 complement deposition on sarcolemma with support- resents the most severe form of IIM and early aggressive ive detection of antibodies against SRP or HMGCR. All poly-immunotherapy is often required to reduce the long- patients performed at least one MRI study including STIR term residual disability [8]. This could be easily detected and T1 sequences of between January 1, 2014 at the date of by muscle magnetic resonance (MRI), which represents January 1, 2020 for diagnostic and clinical purposes. In the the gold standard technique for muscle imaging study in participating centers, the diagnostic workup in suspected muscle diseases [9]. Despite its usefulness, it is largely IIM includes a full clinical evaluation, muscle MRI study, established for inherited myopathies to recognize the pat- muscle biopsy and laboratory analysis with CK and antibody tern and severity of fat replacement by T1-sequences and testing. All patients underwent a full screening for malig- muscle edema/inflammation by T2-Short-tau-inversion- nancy during the diagnostic workup. recovery (STIR)-sequences [10], its application to IIM, All patients underwent clinical examination at baseline and to IMNM in particular, is still not of customary use. and during the clinical follow-up. Neurological examination Muscle MRI could be helpful to assess disease activity included a revised version of the MRC-sum score for proxi- and severity in IMNM, other than to identify the target mal muscles (MRC-60) obtained from the following muscle site for biopsy [11]. Some studies investigated the role groups of each side: arm abductors, elbow flexors, elbow of muscle MRI in IIM showing that it could play a role extensors, hip flexors, knee flexors and knee extensors. Since for monitoring the disease progression and response to MRC sum score provide a 5-point scale of evaluation, the therapy [12, 13]. Concerning IMNM, muscle MRI showed maximum score allowed by summing the strength of each a higher proportion of thigh muscle oedema, atrophy and group of muscles aforementioned is 60, so we called this fat replacement compared to other IIMs with more severe parameter MRC-60. Only evaluations performed in prox- muscle involvement in SRP than HMGCR positive patients imity of the MRI studies (± 1 month) were considered for [14–16]. statistical analysis. A comprehensive clinical evaluation Nevertheless, the pattern of muscle involvement by MRI including muscle strength evaluation of facial, bulbar, axial and its longitudinal use for monitoring disease progression and upper and lower limb impairment scored by MRC scale, and treatment response in IMNM is still warranted. were used to establish the overall disability for clinical out- In this framework, we investigated the use of muscle come purpose: asymptomatic (no muscular weakness), mild MRI to establish its usefulness as prognostic tool for (MRC-60 score: 41–60, mild axial weakness, without facial, bulbar, respiratory or distal muscle weakness, rise from the 13 962 Journal of Neurology (2023) 270:960–974 floor without aid), moderate (MRC-60 score: 21–40, mod- protocols were followed [19] accordingly to international erate axial weakness, with minimal facial, bulbar, respira- standards of muscle biopsy [20] for conventional histologi- tory or distal muscle weakness, rise from the floor with aid) cal, histochemical and immunohistochemical techniques. or severe (MRC-60 score: 0–20, moderate to severe axial, Open muscle biopsies were obtained from deltoid or facial, bulbar, respiratory or distal muscle weakness, unable quadriceps muscles in all patients. Conventional histological to rise from the floor). Disease remission was considered for and histochemical techniques, 8–10 μm thick cryostat sec- patients with normal CK level and stationary best clinical tions were stained with hematoxylin and eosin (HE), modi- outcome for at least 3 months without or tapering immu- fied Gomori trichrome (GT), periodic acid Schiff technique notherapy, regardless of residual clinical impairment. In (PAS), Oil red O, reduced nicotinamide adenine dinucleotide patients with follow-up MRI study, STIR% score improve- dehydrogenase-tetrazolium reductase (NADH-TR), succinic ment > 90% at last MRI or STIR negativity at first MRI per- dehydrogenase (SDH) and cytochrome c oxidase (COX). formed after 6 months from symptom onset (LATE group, The immunohistochemical (IHC) study using antibodies see below) was also considered a supportive criterion. against CD3, CD4, CD8, CD31, CD56, CD68, HLA-ABC, For statistical analysis patients were divided into two HLA-DR and c5-b9 were matched with negative control major groups: (1) patients with MRI performed before slides and visualized using immunoperoxidase techniques. 6 months from symptoms onset regardless of treatment Muscle biopsy was considered diagnostic for IMNM in (EARLY group) and (2) patients with MRI performed after presence of prominent myofiber necrosis and regeneration 6 months from symptoms onset, who were all under treat- in different stages, mild or absent inflammatory cells, faint ment (LATE group). For supplementary analysis the EARLY or absent sarcolemmal MHC-I positivity and c5b-9 com- group was also divided into patients who performed the first plement deposition on sarcolemma in accordance with the MRI before treatment (EARLY-BT subgroup) and patients international standards [21–23]. who performed the first MRI under treatment (EARLY-UT). Muscle MRI studies Laboratory analysis All muscle MRIs were obtained using a 1.5-T MRI follow- ing previously described protocols [24] in accordance with All patients performed a full laboratory analysis, before the international consensus recommendations [25]. A total starting treatment, including blood cell count, hepatic and of 39 muscles of lower body (LB) including pelvic girdle renal function, thyroid hormones, lactate dehydrogenase and lower limb muscles and 18 muscles of upper body (UB) (LDH), and serum creatine kinase (CK). including scapular girdle and arms were studied from each Patients' sera were tested for the presence of all myosi- side, analyzing T1 Turbo spin echo (T1-TSE) and T2-Short tis-specific autoantibodies (MSA) and myositis-associated tau inversion recovery (T2-STIR) sequences. T1 and STIR autoantibodies (MAA) with a line blot test kit (Euroimmun, sequences of scapular girdle or arms were not available for 3 Lubeck, Germany) including 16 different antigens (Mi-2alfa, and 7 patients, respectively. STIR sequences were not avail- Mi-2beta, TIF1gamma, MDA5, NXP2, SAE1, Ku, PM-7, able for two patients at the first MRI acquisition. Scl100, PM-Scl75, Jo-1, SRP, PL-7, PL-12, EJ, OJ, Ro-52) Fat replacement was evaluated on T1 sequences using a [17]. Anti-HMGCR antibodies were studied using ELISA 5-point scale (0–4) according to Fisher classification [26], (Inova Diagnostics, San Diego, CA, USA) following the while edema/inflammation on STIR sequences using a manufacturer instructions [18]. 2-point scale (0: negative, 1: positive). Serum CK activity was analyzed at diagnosis, at first MRI STIR sequences were used to evaluate the presence of study (± 1 month) and last clinical follow-up for all but one muscle inflammation and were considered as “positive” of the patients. Analysis performed in proximity of the MRI when an abnormally increased signal in the intra-muscular studies (± 1 month) were considered for statistical analysis. tissue could be detected, compared with the unaffected sur- rounding muscles. We calculated the percentage of STIR- Morphological study positivity as a fraction of STIR-positive muscles over the total number of muscles evaluated (STIR% = total no STIR- Open muscle biopsy for diagnostic confirmation of IMNM positive muscles/total muscles evaluated*100) separately in was performed in all patients. Proximal upper or lower limb UB and LB (STIR%UB and STIR%LB). muscle was selected accordingly to clinical evaluation of The overall burden of fat replacement was calculated for muscle weakness or evidence of STIR positive signal at LB as a sum of total values of T1 score for each muscle muscle MRI. In 13 patients, muscle MRI was performed per patient (T1-LB score) with a range between 0 (no fat before or within 1 month from muscle biopsy. Internal replacement in any muscle) and 312 (complete replacement 13 Journal of Neurology (2023) 270:960–974 963 in all LB muscles) and it is also expressed as percentage of analysis, using mean fat replacement as the value analyzed, replacement. and graphical representation as a heatmap was performed For longitudinal evaluation of MRI changes, we evaluated using R software, V.4.0.3. the change over time as ΔSTIR and ΔT1 scores, between the first and last MRI available for each patient. Two independent Neurologists with experience in MRI Results analysis (MG, LF) blinded to demographic and clinical fea- tures analyzed all MRIs. In muscles with different T1 and Patients STIR scoring, observers reviewed the muscles together to agree the final score. Twenty-two patients (6 males, 16 females), aged between 23 and 83 years (mean 59.52 ± 17.35) were included in the Statistical analysis study. Seven patients had anti-HMGCR antibodies, 8 anti- SRP antibodies and 7 were seronegative. There were no dif- Demographic, clinical and radiological features, including ferences in median age at onset and gender prevalence in the age at MRI, age at symptoms onset, disease duration, treat- different serotype groups. Muscle biopsies were performed ment start, delay and response, and time elapsed between on deltoid (15 patients), quadriceps (6 patients), biceps fem- symptoms onset, therapy start and MRI study, were collected oris (1 patient). Total body TC scan did not detect presence retrospectively for each patient. For patients with more than of cancer in any patients. Median disease duration at first one MRI study, only the first and last MRI was considered. MRI study was 4 months (IQR 27, range 0–330, SD 82.2). MRC scores and laboratory analysis (serum CK) were Fifteen patients performed the first MRI before considered at the moment of the first MRIs (± 1 month) and 6 months from symptom onset (EARLY group; median of at follow-up end. For subgroup analysis correlation and lon- 3 ± 1.84 months, range 0–6): of these, eleven patients were gitudinal studies, we considered only the scores acquired for untreated at first MRI (EARLY-BT subgroup; median time the LB (STIR%LB and T1-LB and related Δ scores) because of MRI from symptom onset: 2 months, range 1–6) and four of lacking data for UB in some patients. We refer as “base- were under treatment (EARLY-UT subgroup; all patients line” the data at the time of first MRI, and “follow-up” data with corticosteroids from 1 month; median time of MRI at last MRI. from symptom onset 4 months, range 3–5). Seven patients Cohen's kappa coefficient (κ) was calculated to assess the performed MRI later than 6 months from symptom onset inter-rater agreement. Continuous variables were expressed and all these patients were under treatment from a variable as median, range and inter quantile range (IQR). Chi-square period (LATE group; median time of MRI from symptom test was used for comparison of categorical variables. We onset: 66 ± 114 months, range 15–330). identified that none of the variables analyzed was normally Fourteen patients underwent more than one MRI, with distributed using Kolmogorov–Smirnov and therefore we a median follow-up duration of 16.5 months (IQR 15.5, used non-parametric statistical studies. Mann–Whitney range 4–30) between the first and the last MRI. Of these, U test was used to identify whether differences observed eight patients belonging to the EARLY group (six patients between two groups were significant. Kruskal–Wallis with EARLY-BT and two patients EARLY-UT group) while six Dunn’s multiple comparison test was used for comparison patients to the LATE group. of more than two groups (e.g., age of onset comparison All patients received immunotherapy during the dis- between seronegative/HMGCR/SRP patients). Wilcoxon ease course. Median time from symptoms onset to treat- signed rank test was used to determine whether differ- ment start was 3 months (range 0–324). All but one patient ences observed in continuous variables at two time points received prednisone (up to 1 mg/kg/die) as first-line therapy. (baseline and follow-up) were statistically significant (e.g., One patient received IVIG as first-line treatment with dis- T1-LB score at baseline and at the end of follow-up period). ease remission and mild residual disability. Four patients Spearman rank-order test was run to assess if correlations only received steroids obtaining disease remission in two, between variables were statistically significant. Correlation improvement in one and disease stabilization in one. Sev- coefficients are expressed as r and considered strong corre- enteen patients received a second-line therapy, with immu- lation if higher than 0.8 and good if higher than 0.6. When nosuppressant agents (methotrexate n = 11; azathioprine multiple comparisons were performed, we applied a pos-hoc n = 2; mycophenolate mofetil n = 1) or intravenous immu- Bonferroni correction. Two-sided p values were calculated noglobulin (IVIg n = 3). Eleven patients also needed a third- for all analyses; values of < 0.05 were considered significant. line therapy to achieve the best clinical outcome with IVIg All these analyses, as well as the graphics development, in six patients or a different immunosuppressant in five were performed using JASP Statistics 0.16 (IBM, Armonk, (cyclosporine A n = 2; rituximab n = 1; azathioprine n = 1; New York, USA) and GraphPad Prism 8.2.1. Hierarchical mycophenolate mofetil n = 1). 13 Table 1  Clinical and MRI features 964 Treatment Group FU MRI Pt Sex MSA Age (ys) Timing STIR%LB ΔSTIR T1-LB ΔT1 MRC CK onset/first MRI/ No. of IVIG/ Disease Clinical MRI first/second first/ BS/FU FU end (UI/L) treat- delay outcome outcome 13 from second ments (mo) DO/btw MRIs (mo) Untreated Early- BT Yes P1 F HMGCR​ 39 2/7 26/0 − 100% 4/4 0% 52/60 7400/7400/78 3 Y/6 R A P2 F HMGCR​ 34 1/10 46/4 − 92% 0/2 + 0.6% 54/60 8800/8800/79 3 Y/5 R A P3 F HMGCR​ 74 6/9 49/0 − 100% 40/46 + 1.9% 44/53 10,500/10,500/800 1 Y/6 R Mild prox weak- ness P4 F SRP 57 4/16 51/21 − 60% 15/110 + 30.4% 40/56 6000/6000/250 3 Y/15 I Mild prox weak- ness P5 F SRP 78 1/4 37/0 − 100% 18/18 0% 43/57 8000/8000/80 3 Y/5 R Very mild prox weak- ness P6 F – 45 5/17 32/3 − 92% 11/15 + 1.3% 50/60 3250/3250/150 2 N R A No P7 M HMGCR​ 61 3 52 31 42/50 25,000/25,000/2000 1 N I Mild prox weak- ness P8 F SRP 83 2 n.a 30 48/58 3000/3000/95 2 N R Very mild prox weak- ness P9 M – 65 0 27 29 54/60 n.a./n.a./290 1 N R A P10 F – 84 4 58 102 26/34 3000/3000/96 2 N I Mod prox- dist-ax weak- ness P11 F – 22 0 8 0 48/60 30,000/30,000/450 1 N R A Treated Early- WT Yes P12 M HMGCR​ 69 3/23 15/0 − 100% 50/50 0% 54/60 8200/8200/62 2 N R A P13 M SRP 62 4/22 40/27 − 32% 31/54 + 7.4% 49/60 7700/7700/250 2 Y/5–9 R Axial weak- ness Journal of Neurology (2023) 270:960–974 Table 1  (continued) Treatment Group FU MRI Pt Sex MSA Age (ys) Timing STIR%LB ΔSTIR T1-LB ΔT1 MRC CK onset/first MRI/ No. of IVIG/ Disease Clinical MRI first/second first/ BS/FU FU end (UI/L) treat- delay outcome outcome from second ments (mo) DO/btw MRIs (mo) No P14 F SRP 66 5 21 36 43/60 4000/400/150 3 Y/5 R A P15 F – 83 3 14 62 48/52 2100/2100/25 1 N I Mild prox weak- ness Journal of Neurology (2023) 270:960–974 Late Yes P16 M HMGCR​ 76 53/30 4/1 − 66% 54/56 + 0.6% 49/58 5300/410/270 3 Y/7 R Mild prox weak- ness P17 F HMGCR​ 30 330/15 0/0 – 162/162 0% 46/46 n.a./205/80 2 N R Mild prox weak- ness P18 F SRP 62 66/27 2/4 + 100% 62/62 0% 50/57 3600/170/150 3 N R Mild prox weak- ness P19 F SRP 49 15/6 40/42 + 6.5% 45/62 + 5.4% 54/54 7300/5000/300 3 N S Mild prox weak- ness P20 F SRP 53 181/210 0/0 0% 178/180 0% 48/48 3500/n.a./37 3 Y/96 R Mild-mod prox weak- ness P21 F – 39 164/17 23/23 0% 116/116 0% 48/48 n.a./n.a./n.a 3 N S Mild-mod prox-ax weak- ness No P22 M M – 58 9 14 52/60 1200/n.a./700 3 Y/9 R A FU follow-up; DO disease onset; btw between; BS baseline; Y yes; N no; R remission; I improvement; S stabilization/refractory; A asymptomatic 13 965 966 Journal of Neurology (2023) 270:960–974 Fig. 1  Median T1-LB and STIR%LB scores in EARLY/LATE and (B) and with shorter disease duration (A), when muscles are still not treated/untreated patients. Muscle MRI is more informative when replaced by fat (T1-LB score) fat with respect to patients with longer performed early in disease course and before treatment, showing disease course (C) and under treatment (D) higher inflammation (higher STIR%LB score) in patients untreated MRC-60 evaluation at baseline showed a median value (STIR%UB: 26%, range 0–57%) and LB (STIR%LB: 22%; of 48.5 (range 40–54, IQR 7), while at follow-up end it was range 0–55%). 56.5 (range 48–60, IQR 6) (p < 0.0001). The median score of STIR%LB positivity was 38% in CK values dropped from a mean of 6000 U/L (range EARLY-BT group (range 8–57%), 22% in EARLY-UT group 39–10,455, IQR 7792.5) at baseline to a mean of 154 U/L (range 14–40%) and 11% in LATE group (range 0–40%) (range 37–887, IQR 202.5) at last visit (p < 0.0001). (p = 0.001). Considering treatment, the median score of At follow-up end, the best clinical outcome consisted of STIR%LB positivity in treated patients was 15% (range disease remission in 16 patients (8 asymptomatic and 8 with 0–40%) versus the 38% of untreated patients (corresponding residual disability). Four patients improved without disease to the EARLY-BT group) (p < 0.005) (Fig. 1A, B). remission while two patients only achieved a disease stabi- In the LB, pelvic girdle and thighs were in general more lization without improvement. Of symptomatic patients, 11 affected than legs (32% and 34% versus 12% respectively). had variable mild proximal weakness (with associated axial The most frequently STIR positive muscles were adduc- weakness in three, very mild in two, mild-moderate in two), tor magnus (57%), gluteus medius (55%) and obturatorius and one patient had moderate proximal weakness of upper internus (50%) in LB, followed by gluteus minimus, biceps and lower limbs (MRC score 20–40). All patients still need femoris caput longum, semimembranosus, adductor longus therapy to maintain the best clinical outcome. and brevis, gastrocnemius medialis (range of STIR positivity Major clinical and MRI characteristics of each patient are 45–48%) Gastrocnemius medialis was the most frequently summarized in Table 1. affected STIR positive muscle in legs (52%). In the UB, trapezius and supraspinatus were the most affected mus- Cross‑sectional MRI analysis cles (55% and 50% respectively), followed by subscapularis (47%) and infraspinatus (42%) (Fig. 2A). Considering the first MRI studies in all patients, the percent- No differences of STIR%LB score were observed among age of STIR positive muscles (%STIR) was similar in UB the three serological subgroups (SRP: 29%; HMGCR: 26%; seronegative: 26%). However, a slight difference in the 13 Journal of Neurology (2023) 270:960–974 967 Fig. 2  Single muscle STIR% and T1 score in upper and lower body. ularis and infraspinatus (A, right). Similarly, fat replacement occurs In LB STIR positivity is higher in pelvis and thighs than legs, with earlier in lumbar paraspinal muscles, gluteus medius and minimus, adductor magnus, gluteus medius and obturatorius internus and adductor magnus and hamstrings (B, left) in LB, and in subscapula- gastrocnemius medialis as the most frequently STIR positive mus- ris, latissimus dorsi and paraspinous muscles in UB (B, right). Each cles, followed by gluteus minimus, biceps femoris caput longum, column represents the sum of scores obtained for each muscle in the semimembranosus and adductors (A, left). In the UB, trapezius and totality of patients, subdivided in the three groups EARLY-BT (red), supraspinatus show the higher STIR positivity, followed by subscap- EARLY-UT (yellow) and LATE (blue) pattern of STIR positive muscles was observed. Lumbar, adductor magnus and hamstrings, especially the biceps fem- gluteus minimus and hamstring muscles were more com- ori caput longum and the semimembranosus (Fig. 2B). In monly involved in SRP patients; gluteus medius and pelvic UB subscapularis, latissimus dorsi and paraspinous were muscles in HMGCR-positive patients, while quadriceps and the most affected muscles. No statistical differences were hamstring muscles were more frequently detected in seron- observed between the three serological subgroups in terms egative patients. of T1-LB score and distribution of fat replacement. T1 sequence analysis revealed a median T1-LB score of 33.5 (range 0–178). At first MRI, patients of EARLY-BT Longitudinal MRI analysis group had a median T1-LB score of 18 (average 28.4; 8% of replacement), while EARLY-UT group had a median T1-LB A follow-up MRI study was performed in 14 patients with score of 43 (14% of replacement) and patients of LATE a median delay of 16.5 months from the first MRI (range group had a median T1-LB score of 62 (29% of replacement) 4–30, IQR 15.5). The average follow-up distance was (p = 0.004). Considering treatment, the median T1-LB score 11 months in the EARLY-BT subgroup (n = 6), 23 months in treated patients was 58 (range 31–178; average 77.8; 24% in the EARLY-UT subgroup (n = 2) and 21 months in the of replacement) versus the 18 of untreated patients (corre- LATE group (n = 6). sponding to the EARLY-BT group) (p = 0.01) (Fig. 1C, D). In this group of patients, the median STIR%LB score Muscles with higher degree of fat replacement in LB were was 29% (range 0–51%, IQR 29.25) at first MRI and 2% the lumbar paraspinal muscles, gluteus medius and minimus, (range 0–42%, IQR 21.5) at second MRI (p = 0.0024). This 13 968 Journal of Neurology (2023) 270:960–974 Fig. 3  Longitudinal STIR%LB and T1-LB score variation. Varia- second MRI) is observed in patients treated early and aggressively tion of percentage of STIR positivity (red columns) and amount of (EARLY-BT group). Note the higher T1 score progression in cases fat replacement in T1 sequences (blue columns) between first and of p8 (EARLY-BT) and p6 (EARLY-UT) who received IVIG later last MRI studies in LB in different patients divided in EARLY-BT, in disease course. T1 sequences do not show consistent progression EARLY-UT, and LATE groups. STIR sequences represent a use- in patients with longer disease duration (LATE group), sometimes ful outcome measure for treatment response and disease remission resembling LGMD (p17, p21), while STIR sequences could reveal in the EARLY groups. The best outcome (lower STIR and T1 at patients refractory to treatment or with disease relapses (p12, p15) result was more evident in the 8 patients (8/14) of the whole muscles at first MRI that showed a progression of T1 score EARLY group: STIR%LB was 37% (range 15–51%, IQR over the second MRI was similar among groups: 37/183 24) at first MRI and 1% (range 0–40%, IQR 24.5) at sec- muscles in EARLY-BT subgroup (20.2%), 11/43 in EARLY- ond MRI (p = 0.0039), with a median time to the second UT subgroup (25.6%) and 13/54 in LATE group (24.1%) MRI of 22 months (range 6–30, IQR 15.25). The average (p = 0.6) (Fig. 4A). Conversely, the number of muscles that STIR variation between first and last MRI in LB (ΔSTIR-LB showed a T1 progression over the second MRI regardless of score) was −44% considering all patients. Nevertheless, con- STIR positivity at first MRI was higher. In the EARLY-BT sidering subgroups, the average ΔSTIR-LB was −90.6% subgroup 39/76 muscles were not STIR-positive at first MRI in EARLY-BT subgroup and −66.1% in EARLY-UT sub- (51.3%), 4/15 muscles in the EARLY-UT subgroup (26.6%) group, while in the LATE group ΔSTIR-LB was + 21.3% and 7/20 in the LATE group (35%) (p = 0.13) (Fig. 4B). (p < 0.0001) (Fig. 3). The heatmap analysis shows a good overlap between A milder difference was also observed in T1 score change STIR positive muscles and fat replaced muscles in T1 between first and last MRI study from a median score of 56.1 (Fig. 5). at first MRI (range 0–178) to 66.8 at second MRI (range 2–178). The ΔT1-LB was + 5.7% in EARLY-BT subgroup, Correlation analysis compared to + 3.7% in EARLY-UT subgroup and + 6.1% in LATE group (Fig. 3). Inter-rater agreement (κ) for MRI scoring between observers Overall, 22% of STIR positive muscles showed some was 0.87. In general, short disease duration correlated with degree of T1-score progression (at least 1 point) in the sec- higher STIR%LB score (p = 0.02, R = 0.5) while long disease ond MRI study (range 0–87%). The number of STIR positive 13 Journal of Neurology (2023) 270:960–974 969 end also correlated with late start of treatment (p = 0.03, R = 0.47) and in particular with delayed treatment with IVIG (p = 0.03, R = 0.7) (Fig. 6). At baseline, CK values correlate with STIR%LB hyper- intensity (p = 0.03, R = 0.266) and with its change during follow-up (ΔSTIR-LB) (p = 0.04, R =  − 0.6); CK also corre- lates with disease duration (p = 0.015, R =  − 0.223), and with T1-LB scores at baseline and at follow-up end (p = 0.009, R =  − 0.272) (p = 0.02, R =  − 0.16). Variation of CK during follow-up correlated with ΔSTIR%LB hyperintensity (p = 0.04, R = 0.58). Sex, age, and autoantibodies positivity were not associ- ated with MRI alterations. Discussion Muscle MRI in IMNM is a sensitive biomarker for disease activity and treatment response when performed early in dis- ease course, possibly before treatment start. As expected, our data show that STIR positive signal is higher in early stages of disease and in untreated patients and progressively decreases in treated patients and with long disease dura- tion. In fact, ΔSTIR could represent a supportive outcome measure and early prognostic index for treatment response in IMNM, as it correlates with clinical disability and disease Fig. 4  STIR positivity and T1 progression matching. Similar percent- remission other than MRC-60 score and CK variation during age among groups of T1 progression at FU in STIR positive muscles follow-up. Even if MRC significantly varied during follow- at baseline (A). Not all T1 progressed muscles at follow-up were STIR positive at baseline (B) up, we introduced an overall clinical evaluation to better understand retrospectively the clinical course of our cohort of patients (Table 1). Indeed, lower variation of ΔSTIR cor- duration correlated with higher T1-LB score (p = 0.02, relates to higher degree of fat replacement (T1-score) at fol- R = 0.48). low-up—corresponding to higher disability—as it occurs in Shorter timing of muscle MRI study from disease onset undertreated patients or refractory conditions. T1 sequences in untreated patients (EARLY-BT group) correlated with provides information about the overall burden of disease, higher STIR%LB score at baseline and STIR change over representing the most important outcome measure which time (ΔSTIR-LB) (p = 0.001, R = 0.67 and p < 0.001, positively correlates with muscular weakness and clinical R = 0.83 respectively), whereas MRI performed later in impairment. Higher T1-score at follow-up end is also asso- already treated patients (LATE group) correlated with higher ciated with delayed start of treatment from disease onset, T1-LB score (p = 0.004, R = 0.6). in particular with delayed start of IVIG therapy, confirm- Higher STIR%LB score correlated with higher change in ing its key role in inducing disease remission in IMNM. T1 score at follow-up end (ΔT1 score) (p = 0.003, R = 0.76). The importance of IVIG in the management of IMNM Furthermore, lower changes in STIR signal at follow-up is represented by clinical history of three representative (ΔSTIR-LB) correlated with higher T1-LB score at follow- patients belonging to the EARLY-BT group: P1, P4 (Fig. 7) up end (p < 0.001, R = 0.85). and P3. P1 represents the classical management of IMNM Both higher T1-LB and lower ΔSTIR-LB scores corre- characterized by a progressive adding-on therapy, started lated with worse clinical outcome at follow-up end, spe- with prednisone as first-line therapy with partial improve- cifically: with lower MRC-60 score (p < 0.001, R = 0.7 and ment, subsequent add-on of immunosuppressive agents for p < 0.002, R = 0.6 respectively), lower disease remission rate second-line therapy and lastly, the use of IVIG at 6 months (p = 0.001, R = 0.65 and p < 0.004, R = 0.56 respectively) from disease onset to reach a disease remission and complete and worst clinical outcome with higher diffuse/axial muscle clinical recovery. Conversely P4, despite having a similar weakness (p < 0.001, R = 0.73 and p = 0.001, R = 0.8 respec- history with early start of first and second-line treatments, tively). Noteworthily, the higher T1-LB score at follow-up started IVIG therapy only 15 months later disease onset with 13 970 Journal of Neurology (2023) 270:960–974 Fig. 5  STIR and T1 matching heatmap at baseline and follow-up. the general overlap between STIR positive muscles (active phase of Patients listed basing on timing at first MRI from disease onset in disease, EARLY group, up) and fat replaced muscles in T1 (most months (bold, left) and time elapsed between MRIs for each patient. advanced phases of disease, LATE group, down). Each column repre- Note the inverse gradient of STIR positivity decrease and T1 sever- sents one muscle of LB ity increase from up (EARLY groups) to down (LATE group), and 13 Journal of Neurology (2023) 270:960–974 971 importance of MRI to assess disease remission regardless of residual clinical disability. In general, we observed a good overlap between inflam- mation and fat replacement (Fig. 5), in accordance with the assumption that fat replacement in inflammatory myopathies is the consequence of muscle inflammation. In our cohort, during the period of longitudinal evaluation, about 20% of STIR positive muscles progressed in T1 score at follow- up, regardless of belonging group, while more than 50% of muscles that showed T1-score progression were not STIR positive in the first MRI. This unexpected finding could have different explication. First, STIR sequences have not enough sensitivity to detect all inflammatory processes in muscles, in fact muscle biopsies in patients affected by IIMs can Fig. 6  IVIG starting treatment and long-term fat replacement correla- show inflammatory infiltrates also in STIR negative mus- tion. Correlation analysis shows that delayed treatment with IVIG (in cles, suggesting that cellular mechanisms of inflammation months) is associated with higher fat replacement (LB-T1 score) at are not constantly captured by muscle MRI, probably due to last FU (p = 0.03) lack of muscle oedema above the threshold needed for MRI detection. Nevertheless, inflammatory infiltrates still ensure marked improvement, but considerable residual disability the pathophysiological mechanism underlying the muscle due to the fat replacement occurred in muscles, confirmed by damage and fat replacement progression. On the other hand, marked increase of T1 score (ΔT1) at follow-up MRI study. MRI can detect muscle inflammation in patients with no Finally, P3 obtained disease control only by IVIG as first- clinical sign of muscular involvement and in biopsy-negative line therapy because of contraindication of corticosteroids patients [27, 28] because of sampling errors, or conversely and immunosuppressive agents, supporting the beneficial because of STIR hyperintensity could occur for a transitional use of IVIG also as first-line therapy alone. period. Probably, other mechanisms than muscle inflamma- On the other hand, T1-score is higher in LATE group. tion could participate to the pathophysiology underlying Most of these patients are in a remission phase of the disease fat replacement and disease progression in inflammatory and muscle MRI provides information on the consequences myopathies. of past periods of muscle inflammation (at onset and dur- Even if a preferential pattern of involvement could be ing relapses) and correlates with clinical disability. Longer recognized in our cohort of patients according to those disease course is generally associated to higher degree of fat already described [15, 29], no significant differences among replacement, probably because of possible periods of low different serotypes have been observed. The only minor disease control (due to treatment refractory), or even because variability concerned the pattern of muscle STIR positiv- of an initial period of misdiagnosis (e.g. seronegative IMNM ity—which was more prominent in lumbar, gluteus minimus for toxic myopathies). In these cases, especially if there is and hamstring muscles in SRP, gluteus medius and pelvic a permanent disability at neurological examination, persis- muscles in HMGCR and quadriceps and hamstring muscles tence of STIR positive signal can help to assess treatment in seronegative patients suggesting a possible implication unresponsiveness or even detect disease relapses. Conversely of the antibody-related disease mechanism in the muscle- STIR negativity is associated to chronic or remission phases. specific susceptibility, a well-known phenomenon frequently In most chronic cases, muscle MRI could resemble that of observed in inherited myopathies [30]. Certain muscles are inherited myopathies or muscular dystrophies. Two cases early involved in disease course, while others remain spared of LATE group are representative: P21 and P17 (Fig. 7) until the late-end stage of disease, even if all muscles in the that performed first MRI very late in disease course. In the body share the same genetic background. This variability former case, MRIs were performed during a lasting period represents the cornerstone of the disease-specific patterns of disease activity and refractory to different immunosup- of muscle involvement described in several myopathies and pressive treatments, while in the latter MRIs were both per- muscular dystrophies. Mechanisms underlying the variable formed during a period of disease remission. In these cases, muscle-specific susceptibility/resistance to injury is not yet STIR negativity helps to identify disease remission despite fully understood but it is probably the result of complex residual muscular weakness from those with STIR positivity interactions between different gene expression and modu- due to still active disease, which could improve with alter- lation between muscles, and muscle tissue environment. native or more aggressive immunotherapy, underling the Different inflammatory pathways among IIMs could lead to the preferential involvement of certain muscles as inclusion 13 972 Journal of Neurology (2023) 270:960–974 Fig. 7  Muscle MRI findings. Representative cases of the importance of MRI study in the management of IMNM patients. Complete disease remission and clinical recovery with early aggressive immunotherapy (p1, EARLY-BT group). Incomplete disease remission (still positive STIR signal) and progression of fat replacement (T1 worsen- ing) due to delayed start of IVIG (p4, EARLY-BT group). Persistence of STIR positiv- ity and mild progression of fat replacement despite poly-immu- notherapy in treatment refrac- tory patient (p21, LATE group). STIR negativity (not shown) and T1 stationarity in a chronic- remission phase resembling LGMD (p17, LATE group) body myositis (IBM) (flexor digitorum profundus, quadri- [40], but also in SRP or seronegative patients, as in the cases ceps and gastrocnemius medialis), DM and Antisynthetase P17, P20 and P21 of the LATE group. The risk to develop syndrome (ASS) (symmetrical pelvic and relative adductor- severe clinical phenotypes resembling LGMD should be sparing thigh muscles) or myasthenia-myositis association linked to history of long disease duration in misdiagnosed, (upper limb extensors and more prominent in myotendinous undertreated or treatment-refractory patients, rather than to junctions) [19, 31–33], helping to differentiate the pattern a specific serological group. of muscle involvement from IMNM. The predominant axial This study has some limitations. The major is due to the and pelvi-femoral muscle involvement with prominent STIR retrospective analysis and the heterogeneity of the sample, positivity could also help to differentiate IMNM from other with MRIs performed at different time from diagnosis. The myopathies [34–39]. IMNM may resemble limb-girdle mus- small size of our cohort did not allow us to have a rele- cular dystrophies (LGMD) not only in HMGCR serogroup vant number of patients in each serological group possibly 13 Journal of Neurology (2023) 270:960–974 973 reducing the statistical differences between them. MRI References study did not include upper body for all patients, and during follow-up it was performed for clinical purposes only (not 1. Dalakas MC (2015) Inflammatory muscle diseases. N Engl J Med following a standardized protocol). Finally, we could not 372:1734–1747. https://​doi.​org/​10.​1056/​NEJMr​a1402​225 2. De Bleecker JL, De Paepe B, Aronica E et al (2015) 205th ENMC correlate the MRI results with muscle biopsy, because they international workshop: pathology diagnosis of idiopathic inflam- were not performed concurrently in all cases. matory myopathies part II 28–30 March 2014, Naarden, The Netherlands. Neuromuscul Disord 25:268–272. https://​doi.​org/​ 10.​1016/j.​nmd.​2014.​12.​001 3. Allenbach Y, Mammen AL, Benveniste O et al (2018) 224th ENMC international workshop. Neuromusc Disord 28:87–99. Conclusions https://​doi.​org/​10.​1016/j.​nmd.​2017.​09.​016 4. Mammen AL, Chung T, Christopher-Stine L et al (2011) Autoanti- Muscle MRI is a sensitive biomarker for monitoring dis- bodies against 3-hydroxy-3-methylglutaryl-coenzyme A reductase in patients with statin-associated autoimmune myopathy. Arthritis ease activity and therapy response in IMNM, notably if per- Rheum 63:713–721. https://​doi.​org/​10.​1002/​art.​30156 formed early in disease course and before treatment start, 5. Watanabe Y, Uruha A, Suzuki S et al (2016) Clinical features and and could represent a supportive outcome measure and prognosis in anti-SRP and anti-HMGCR necrotising myopathy. J early prognostic index. In particular higher STIR positivity Neurol Neurosurg Psychiatry 87:1038–1044. https://​doi.​org/​10.​ 1136/​jnnp-​2016-​313166 at baseline and delayed treatment start (in particular with 6. Lim J, Rietveld A, De Bleecker JL et al (2019) Seronega- IVIG) are associated to higher degree of fat replacement tive patients form a distinctive subgroup of immune-mediated during disease course and worst clinical outcome. necrotizing myopathy. Neurol Neuroimmunol Neuroinflamm 6:e513. https://​doi.​org/​10.​1212/​NXI.​00000​00000​000513 Author contributions LF, MG: study concept, design and supervision, 7. Allenbach Y, Benveniste O, Stenzel W et al (2020) Immune- analysis and interpretation of clinical and neuroimaging data, manu- mediated necrotizing myopathy: clinical features and pathogen- script writing, responsibility for the integrity of the study; AL: analysis esis. Nat Rev Rheumatol 16:689–701. https://​doi.​org/​10.​1038/​ and interpretation of clinical and neuroimaging data, manuscript writ- s41584-​020-​00515-9 ing, responsibility for the integrity of the study; JAP, SM, GA, RC, 8. Kassardjian CD, Lennon VA, Alfugham NB et al (2015) Clini- LT, FV, ER, MS, GA: patients’ clinical evaluation, critical revision for cal features and treatment outcomes of necrotizing autoimmune intellectual content; GM, EGV: database creation and data reporter, myopathy. JAMA Neurol 72:996–1003. https://​doi.​org/​10.​1001/​ critical revision of data collection; TT: analysis and interpretation of jaman​eurol.​2015.​1207 neuroimaging studies; JDM, LL: statistical analysis; drafting and revis- 9. Ten Dam L, van der Kooi AJ, Verhamme C et al (2016) Muscle ing manuscript; clinical evaluation of patients, study concept, local imaging in inherited and acquired muscle diseases. Eur J Neurol study supervisor; all authors have read and approved the submission. 23:688–703. https://​doi.​org/​10.​1111/​ene.​12984 10. Warman-Chardon J, Diaz-Manera J, Tasca G et al (2020) 247th Funding Open access funding provided by Università degli Studi di ENMC International Workshop: Muscle magnetic resonance Roma La Sapienza within the CRUI-CARE Agreement. This study did imaging—Implementing muscle MRI as a diagnostic tool for not receive any funding. The authors have nothing to disclose. rare genetic myopathy cohorts. Hoofddorp, The Netherlands, Sep- tember 2019. Neuromuscul Disord 30:938–47. https://​doi.​org/​10.​ Declarations 1016/j.​nmd.​2020.​08.​360 11. Malartre S, Bachasson D, Mercy G et al (2021) MRI and muscle Conflicts of interest The authors declare that they have no conflict of imaging for idiopathic inflammatory myopathies. Brain Pathol interest. 31:e12954. https://​doi.​org/​10.​1111/​bpa.​12954 12. Allenbach Y, Benveniste O (2018) Peculiar clinicopathological Consent to participate Informed consent was obtained from all indi- features of immune-mediated necrotizing myopathies. Curr Opin vidual participants included in the study. Rheumatol 30:655–663. https://​doi.​org/​10.​1097/​BOR.​00000​ 00000​000547 Open Access This article is licensed under a Creative Commons Attri- 13. Pilania K, Jankharia B (2021) Role of MRI in idiopathic inflam- bution 4.0 International License, which permits use, sharing, adapta- matory myopathies: a review article. Acta Radiol 63:200–213. tion, distribution and reproduction in any medium or format, as long https://​doi.​org/​10.​1177/​02841​85121​990305 as you give appropriate credit to the original author(s) and the source, 14. Del Grande F, Carrino JA, Del Grande M et al (2011) Magnetic provide a link to the Creative Commons licence, and indicate if changes resonance imaging of inflammatory myopathies. Top Magn Reson were made. The images or other third party material in this article are Imaging 22:39–43. https://​doi.​org/​10.​1097/​RMR.​0b013​e3182​ included in the article's Creative Commons licence, unless indicated 5b2c35 otherwise in a credit line to the material. If material is not included in 15. Pinal-Fernandez I, Casal-Dominguez M, Carrino JA et al (2017) the article's Creative Commons licence and your intended use is not Thigh muscle MRI in immune-mediated necrotising myopathy: permitted by statutory regulation or exceeds the permitted use, you will extensive oedema, early muscle damage and role of anti-SRP need to obtain permission directly from the copyright holder. To view a autoantibodies as a marker of severity. Ann Rheum Dis 76:681– copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. 687. https://​doi.​org/​10.​1136/​annrh​eumdis-​2016-​210198 16. Zheng Y, Liu L, Wang L et al (2015) Magnetic resonance imaging changes of thigh muscles in myopathy with antibodies to sig- nal recognition particle. Rheumatology (Oxford) 54:1017–1024. https://​doi.​org/​10.​1093/​rheum​atolo​gy/​keu422 13 974 Journal of Neurology (2023) 270:960–974 17. Rönnelid J, Barbasso Helmers S, Storfors H et al (2009) Use of a subacute-onset idiopathic inflammatory myopathy. Muscle Nerve commercial line blot assay as a screening test for autoantibodies 51:253–258. https://​doi.​org/​10.​1002/​mus.​24307 in inflammatory myopathies. Autoimmun Rev 9:58–61. https://​ 29. Landon-Cardinal O, Koumako C, Hardouin G et al (2020) Severe doi.​org/​10.​1016/j.​autrev.​2009.​03.​005 axial and pelvifemoral muscle damage in immune-mediated 18. Musset L, Miyara M, Benveniste O et al (2014) Analysis of necrotizing myopathy evaluated by whole-body MRI. Semin autoantibodies to 3-hydroxy-3-methylglutaryl-coenzyme Arthritis Rheum 50:1437–1440. https://​doi.​org/​10.​1016/j.​semar​ A reductase using different technologies. J Immunol Res thrit.​2020.​02.​009 2014:405956. https://​doi.​org/​10.​1155/​2014/​405956 30. Mercuri E, Pichiecchio A, Allsop J et al (2007) Muscle MRI in 19. Garibaldi M, Fionda L, Vanoli F et al (2020) Muscle involve- inherited neuromuscular disorders: past, present, and future. J ment in myasthenia gravis: expanding the clinical spectrum of Magn Reson Imaging 25:433–440. https://​doi.​org/​10.​1002/​jmri.​ myasthenia-myositis association from a large cohort of patients. 20804 Autoimmun Rev 19(4):102498. https://​doi.​org/​10.​1016/j.​autrev.​ 31. Tasca G, Monforte M, De Fino C et al (2015) Magnetic resonance 2020.​102498 imaging pattern recognition in sporadic inclusion-body myositis. 20. Udd B, Stenzel W, Oldfors A et al (2019) 1st ENMC European Muscle Nerve 52:956–962. https://​doi.​org/​10.​1002/​mus.​24661 meeting: the EURO-NMD pathology working group recom- 32. Spalkit S, Sinha A, Prakash M, Sandhu MS (2021) Dermatomy- mended standards for muscle pathology Amsterdam, The Neth- ositis: Patterns of MRI findings in muscles, fascia and skin of pel- erlands, 7 December 2018. Neuromuscul Disord 29:483–485. vis and thigh. Eur J Radiol 141:109812. https://​doi.​org/​10.​1016/j.​ https://​doi.​org/​10.​1016/j.​nmd.​2019.​03.​002 ejrad.​2021.​109812 (Epub 2021 Jun 6 PMID: 34118766) 21. Allenbach Y, Mammen AL, Benveniste O et al (2018) 224th 33. Zhang W, Zheng Y, Wang Y et al (2022) Thigh MRI in antisyn- ENMC international workshop: clinico-sero-pathological classi- thetase syndrome, and comparisons with dermatomyositis and fication of immune-mediated necrotizing myopathies Zandvoort, immune-mediated necrotising myopathy. Rheumatology (Oxford). The Netherlands, 14–16 October 2016. Neuromuscul Disord https://​doi.​org/​10.​1093/​rheum​atolo​gy/​keac2​69 (Epub ahead of 28:87–99. https://​doi.​org/​10.​1016/j.​nmd.​2017.​09.​016 print. PMID: 35512205) 22. De Bleecker JL, Lundberg IE, de Visser M (2013) 193rd ENMC 34. Fischer D, Walter MC, Kesper K et al (2005) Diagnostic value international workshop pathology diagnosis of idiopathic inflam- of muscle MRI in differentiating LGMD2I from other LGMDs. J matory myopathies 30 November–2 December 2012, Naarden, Neurol 252:538–547. https://​doi.​org/​10.​1007/​s00415-​005-​0684-4 The Netherlands. Neuromuscul Disord 23:945–951. https://​doi.​ 35. Barp A, Laforet P, Bello L et al (2020) European muscle MRI org/​10.​1016/j.​nmd.​2013.​07.​007 study in limb girdle muscular dystrophy type R1/2A (LGMDR1/ 23. Merlonghi G, Antonini G, Garibaldi M (2022) Immune-mediated LGMD2A). J Neurol 267:45–56. https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​ necrotizing myopathy (IMNM): a myopathological challenge. s00415-​019-​09539-y Autoimmun Rev 21:102993. https://d​ oi.o​ rg/1​ 0.1​ 016/j.a​ utrev.2​ 021.​ 36. Alonso-Pérez J, González-Quereda L, Bello L et al (2020) New 102993 genotype-phenotype correlations in a large European cohort of 24. Garibaldi M, Tasca G, Diaz-Manera J et al (2017) Muscle MRI in patients with sarcoglycanopathy. Brain 143:2696–2708. https://​ neutral lipid storage disease (NLSD). J Neurol 264:1334–1342. doi.​org/​10.​1093/​brain/​awaa2​28 https://​doi.​org/​10.​1007/​s00415-​017-​8498-8 37. Garibaldi M, Nicoletti T, Bucci E et al (2021) Muscle magnetic 25. Hollingsworth KG, de Sousa PL, Straub V et al (2012) Towards resonance imaging in myotonic dystrophy type 1 (DM1): refin- harmonization of protocols for MRI outcome measures in skeletal ing muscle involvement and implications for clinical trials. Eur J muscle studies: consensus recommendations from two TREAT- Neurol 29(3):843–854. https://​doi.​org/​10.​1111/​ene.​15174 NMD NMR workshops, 2 May 2010, Stockholm, Sweden, 1–2 38. Alonso-Jimenez A, Kroon RHMJM, Alejaldre-Monforte A et al October 2009, Paris, France. Neuromuscul Disord 22(Suppl (2019) Muscle MRI in a large cohort of patients with oculo- 2):S54–S67. https://​doi.​org/​10.​1016/j.​nmd.​2012.​06.​005 pharyngeal muscular dystrophy. J Neurol Neurosurg Psychiatry 26. Fischer D, Kley RA, Strach K et al (2008) Distinct muscle imag- 90:576–585. https://​doi.​org/​10.​1136/​jnnp-​2018-​319578 ing patterns in myofibrillar myopathies. Neurology 71:758–765. 39. Díaz-Manera J, Llauger J, Gallardo E et al (2015) Muscle MRI in https://​doi.​org/​10.​1212/​01.​wnl.​00003​24927.​28817.​9b muscular dystrophies. Acta Myol 34:95–108 27. Wong EH, Hui ACF, Griffith JF et al (2005) MRI in biopsy-nega- 40. Mohassel P, Landon-Cardinal O, Foley AR et al (2019) Anti- tive dermatomyositis. Neurology 64:750. https://​doi.​org/​10.​1212/​ HMGCR myopathy may resemble limb-girdle muscular dystro- 01.​WNL.​00001​44190.​30708.​D7 phy. Neurol Neuroimmunol Neuroinflam 6:e523. https://​doi.​org/​ 28. Van De Vlekkert J, Maas M, Hoogendijk JE et al (2015) Com- 10.​1212/​NXI.​00000​00000​000523 bining MRI and muscle biopsy improves diagnostic accuracy in 13