Whole-body muscle MRI of patients with MATR3-associated distal myopathy reveals a distinct pattern of muscular involvement and highlights the value of whole-body examination

Patients with genetically confirmed Matrin-3-associated distal myopathy (p.S85C/c.254C > G mutation) who underwent whole-body MRI in the course of routine diagnostic work-up were asked for permission to re-evaluate the obtained data for the present study. 15 patients (25–79 years of age, 7 female) out of 8 families were included in the current study. The majority of the patients (9 patients) were in mid-stage of the disease (41–55 years of age, disease duration 5–29 years), whilst each three were classified to be in disease onset (25–30 years of age, presymptomatic or myalgia only) and end-stage (67–79 years of age, disease duration 36–39 years, disease duration not known in one case) disease, respectively. While 4 of the patients (patients 4, 5, 10, 11) were previously described [29], 4 newly identified unrelated pedigrees were included in the current study. Table 1 summarizes the main characteristics of the patients.

Whole-body-Imaging was done on a 3.0 T MRI-system (Skyra; Siemens, Erlangen, Germany) using two flexible 18-channel transmit/receive surface body coils for neck, thorax, arms, abdomen and pelvis. A 36-channel angiography coil was used for the legs. The patient was placed in the supine position. Imaging included a T1-Weighted Dixon Turbo Spin Echo sequence in an axial plain and 8 mm slice thickness. The total amount of time required for the MRI examination was 45 min. No sedation or contrast agent was used.

The grade of affection of the individual muscle was determined using the five-point semi-quantitative grading scale established by Fischer et al. [10]. Briefly, stage 0 refers to a normal appearance, while decreased T1-signal intensity to a variable extent is found in stages 1–4 (1—slight, non-confluencing; 2—beginning confluence in less than 50% of the muscle; 3—confluence in more than 50% of the muscle; 4—replacement of the entire muscle).

Determination was conducted by an experienced radiologist (DS) who was blinded to the clinical data. A complete summary showing the individual patients’ determined Fischer-grades can be found in Suppl. Table 1.

To establish a comprehensive MRI-pattern of muscular affection, patients were first stratified in regard to disease duration (disease onset, mid-stage and end-stage disease). For pattern definition, only advanced disease stages (e.g. mid- and end-stage) were used. For each individual muscle, average as well as median Fischer grade and standard deviation was determined. Using these parameters, the grade of affection was classified into four categories—not/rarely affected (average Fischer grade 0–1.4, median Fischer grade 0), mild/variable affected (average Fischer grade 1.5–2.4, median Fischer grade 1–2, standard deviation > 1.5), moderately affected (average Fischer grade 2.5–3.4, median Fischer grade 3) and early/severely affected (average Fischer grade ≥ 3.5, median Fischer grade 4).

To test the reliability of the obtained MRI-pattern, all patients with MATR3-myopathy included in this study were compared to the MRI-pattern of muscular affection identified in this study. The derived MATR3 MRI-pattern was dichotomized, only distinguishing between unaffected (not or mild/variable affected) and affected (moderately and early/severely affected) muscles. In patients, all muscles with a Fischer grade > 1 were considered to be affected. For each patient, the mismatch to the defined MATR3-pattern was determined. In this regard, a mismatch was defined as either a muscle that was affected in the individual patient but not in the defined MATR3-pattern or vice versa a muscle that appeared to be not affected in the individual patient but was defined affected in the MATR3-pattern. A mismatch ≤ 20% of all muscles studied in the individual patient was supposed a good congruency to the MATR3-pattern. Hence, the portion of patients fulfilling this criterion was identified.

Patterns of muscular involvement for the individual distal myopathies were defined on the basis of previous findings in distal myopathies. The work of Bugiardini et al. served as the basis for pattern definition [28]. Furthermore, a selective literature review was done using the terms “MRI”, “magnetic resonance imaging”, “MR” and “imaging” in connection with “distal myopathy” as well as the respective disease entities. A comprehensive list of the literature used for pattern definition can be found in Suppl. Table 2. In addition to the known “classical” distal myopathies (obligatory distal phenotype), several myopathies with comparably high prevalence and potential predominant distal affection were included (Facioscapulohumeral muscular dystrophy 1, Myotonic dystrophy 1, sporadic Inclusion body myositis, Hereditary myopathy with early respiratory failure).

The established literature-based pattern was then compared to MATR3-pattern and mismatch was determined. Again, a mismatch ≤ 20% was defined as relevant congruency.

Muscular involvement was mainly symmetrical. A predominant affection of the lower extremities with an emphasis on the lower legs was observed. In lower legs, a predominant involution of soleus muscle as well as medial head of gastrocnemius muscle was detected. However, the anterior and lateral compartment (e.g. tibialis anterior, extensor digitorum longus and peroneus longus muscle) were also involved in most of the cases, though to a lesser extent. A variable degeneration of lateral head of gastrocnemius muscle was seen. However, the deep posterior compartment was mainly unaffected. In thighs, a selective affection of parts of the hamstring muscles (semimembranosus muscle and the long head of the biceps femoris muscle) was observed whereas the anterior and medial compartments were mostly spared. Imaging of the trunk showed early and severe involvement of the gluteus minimus muscle. In contrast, gluteus medius muscle showed a variable grade of affection while gluteus maximus muscle appeared to be not relevantly involved. Evaluation of the paraspinal musculature revealed a scattered degeneration predominantly affecting thoracic parts of the axial musculature (e.g. thoracis longus, semispinalis thoracis and spinalis thoracis muscles). The cervical and lumbar parts of paraspinal musculature appeared only mildly affected. Except for the deltoid muscle, that—especially in the ventral part—showed a moderate involution, no relevant involvement of the musculature of shoulder and arm was seen. Representative examples of whole body imaging findings in MATR3-myopathy are depicted in Fig. 1 while the observed grade of alteration for the individual muscles as determined by Fischer grading are shown in Fig. 2.

At disease onset, there was already a slight (Fischer Grade 1 and 2) involvement in muscles that appeared to be severely affected in advanced disease stages. Especially, gluteus medius muscle as well as the lower limb muscles (e.g. soleus and peroneus longus muscles and the medial head of the gastrocnemius) showed mild involution in some of the patients (Fig. 2b, upper panel). During the course of the disease, a contemporaneous degeneration of the typically affected muscles was observed (Fig. 2b, middle panel), while not or rarely affected muscles appeared to be spared up to late stages of the disease, showing only mild involution that could be ascribed to age-related changes (Fig. 2b, lower panel).

A comprehensive pattern of muscular involvement in MATR3-associated myopathy was defined on the basis of the Fischer grading system (for details see “Methods” section). Three muscles, namely soleus muscles, the medial head of gastrocnemius muscle and gluteus minimus muscle were identified as early/severely affected. Tibialis anterior, extensor digitorum, peroneus longus and semimembranosus muscles as well as the thoracic parts of paraspinal musculature were categorized as being moderately affected. A mild/variable involution was identified for the lateral head of gastrocnemius and biceps femoris muscles, gluteus medius muscle, deltoideus muscle and cervical/lumbar parts of paraspinal musculature. All other muscles were defined as been not or rarely affected. A comprehensive summary of the defined MATR3-pattern is given in Fig. 3a.

To evaluate the practicability of the defined MATR3-pattern, mismatches to the findings in the individual patients were identified (Fig. 3b, Table 2). For all patients, a median mismatch of 6 muscles (12% of all examined muscles) was determined. The best match to the defined pattern was observed in the group of individuals with mid-stage disease, showing a median mismatch of 2 muscles (4%). While patients in end-stage disease showed a rather good match to the defined MATR3 pattern (median mismatch 7 muscles), a consistently poor match was seen in the disease onset group (median mismatch 10 muscles).

Overall, in 11 out of 15 patients (73%) a mismatch less than 20% of all examined muscles was observed. None of the patients in the disease onset group showed a sufficient match with the MATR3-pattern. A sufficient congruency (defined as mismatch < 20%) was observed for 89% of the patients in mid-stage and 100% of the patients in end-stage disease. An even stricter match to the defined MATR3-pattern (mismatch < 10%) was observed in 5 out of 9 patients in mid-stage disease (55%), while only one specimen (33%) in late-stage disease fulfilled this criterion.

To determine the diagnostic practicability of the established MATR3-pattern, it was compared to literature-based lower-limb MRI patterns of other distal myopathies (Fig. 4a). A relevant discrimination (e.g. mismatch > 20%) of the MATR3-pattern to the majority of distal myopathies (16 out of 22 disease entities) was seen. In contrast, relevant similarities were observed in 6 distal myopathies (TIA1, late-stage TTN, MYOT, SQSTM1/TIA1, KLHL9, ADSSL1). Comprehensive information regarding affected muscles other than lower limb muscles were lacking for most of these distal myopathies, as only a few underlying studies applied whole-body MRI (2 out of 15, Table 3). Based on the available literature, a frequent affection of gluteal muscles was observed in MYOT- associated distal myopathy. However, a selective involution of gluteus minimus muscle as in MATR3-associated myopathy was not described (Table 3). Muscular involution of the paraspinal musculature was again solely reported in MYOT-associated distal myopathy. While a selective involvement of the thoracic paraspinal segments has not been described, there has been no explicit discrimination between lumbar, thoracic and cervical segments in this study (Table 3). Furthermore, respective changes were not present in MR-studies from patients with MYOT- and SQSTM1/TIA1-associated distal myopathy overseen in the neuromuscular research centre in Halle (Fig. 4b).