Introduction

To date, non-invasive muscle imaging is mostly performed with standard ¹H-MRI sequences like ¹H T₁-weighted and T₂-weighted short-TI inversion recovery (STIR) sequences for visual assessment of fat replacement and edema and Dixon-type sequences for quantification of proton-density fat fraction (PDFF). MRI measurements of ionic balance in skeletal muscles could offer added value in the assessment of neuromuscular diseases. To date, MRI measurements of sodium (²³Na) have already shown altered sodium concentrations in diseases like ion channelopathies or Fascioscapulohumeral muscular dystrophy^1,2. Potassium (³⁹K) MRI measurements were mostly performed in healthy subjects³. In vivo measurements of ion concentrations within myofibrillar degeneration are not yet available. Here, we performed a first combined ³⁹K/²³Na-MRI study in a cohort of ten patients with myofibrillar degenerations due to filamin C, desmin and LDB3 mutations.

Methods

Ten patients with genetically proven myofibrillar myopathies (6m, 4f, mean age 51.0±7.7; 6 filaminopathy, 2 desminopathy, 2 zaspopathies due to LDB3 mutations) and ten healthy controls (6m, 4f, mean age 50.0±11.0) were prospectively enrolled in our study after written informed consent according to the approval of the local ethics review board.
¹H-MRI was performed at a 3T system (Magnetom Vida, Siemens Healthcare GmbH, Erlangen, Germany) using an 18-channel flexible RF-coil and included a T₁-weighted turbo-spin echo (TSE) sequence, a T₂-weighted STIR sequence and a 6-point Dixon-type gradient echo. In the first two sequences both lower legs were examined, in the Dixon-type sequence only the less affected one as assessed by T₁-weighted sequence. The same leg was afterwards examined by ²³Na- and ³⁹K-MRI at a 7T system (Magnetom Terra.X, Siemens Healthcare GmbH) with a dual-tuned, quadrature birdcage ²³Na/³⁹K-RF-coil that included external reference phantoms with NaCl/KCl solutions ([Na⁺]/[K⁺] = 10/240,20/210,25/180,30/150,40/120 mM) using an acquisition-weighted Stack-of-Stars (AW-SOSt) sequence⁴. Sequence parameters are given in Table 1.
Intramuscular fat and edema were graded on ¹H-images using a four-point semi-quantitative scale (from 1:= no to 4:= complete fatty replacement/homogeneous edema²). Quantitative PDFF values were calculated based on 6-point-Dixon-type sequence assuming a 8-peak fat model, single T₂*-relaxation rate with a GraphCut algorithm. For ²³Na/³⁹K quantification ¹H-images were co-registered on ²³Na-images, segmented semi-automatically with Dafne and blood vessels extracted with a threshold within ²³Na-images^4,5. ²³Na- and ³⁹K-MRI data were corrected for partial-volume and relaxation effects and apparent tissue ²³Na and ³⁹K concentrations (aTSC and aTPC) calculated based on the external references and fat-corrected by the help of PDFF values.

Results

The patient cohort studied was very heterogeneous in terms of fatty muscle replacement both interindividually and intermuscularly and ranged from slight/intermediate changes to complete fatty replacement of all muscles. Less fatty replaced muscles displayed intermediate to severe edema (exemplarily ¹H-images for two different filaminopathy patients with different disease stage and one desminopathy patient see Fig. 1). In most of the cases the gastrocnemius medialis (GM) and the soleus muscle (SOL) were most affected (PDFF(GM) = 86.1%, 37.2±21.9%, 0.0002, PDFF(SOL) = 87.2%, 66.9±34.9%, 0.001; median, mean±std, p). aTSC were increased in less fatty-replaced muscles, aTPC were decreased (exemplarily aTSC/aTPC maps see Fig. 2, same patients as in Fig. 1). Subsequently, fat-corrected aTSC values were significantly increased in all muscle compartments, corresponding aTPC-values were decreased with significant values for all muscles evaluated except peroneus and gastrocnemius lateralis muscles (p-values SOL(TSC/TPC) = 0.0005/0.0002; see Fig. 3).

Discussion

Irrespective of the individual gene defect and stage of disease progression, all ten patients suffering from myofibrillar myopathies were found to have increased fat-corrected aTSC and decreased aTPC. The reason for these alterations are currently unknown. A possible mechanism for these altered tissue ion concentrations could be the involvement of the Na⁺/K⁺-pump in the degenerative muscle pathology due to the disease causing gene defects.

Conclusion

Our results highlight alterations of the ³⁹K/²³Na ion balance in the genetically and clinically heterogeneous group of myofibrillar myopathies that go beyond changes attributable to mere fatty-replacement of skeletal muscle tissue. Further studies on larger cohorts are needed to investigate whether different genotypes show differences in ³⁹K/²³Na-MRI or whether disease progression can be detected in early stages.

Acknowledgements

We are grateful to our patients for their participation. We thank the Imaging Science Institute (Erlangen, Germany) for providing us with measurement time at the 3T MRI system.