Quantitative muscle MRI captures early muscle degeneration in calpainopathy

To evaluate differences in qMRI parameters of muscle diffusion tensor imaging (mDTI), fat-fraction (FF) and water T2 time in leg muscles of calpainopathy patients (LGMD R1/D4) compared to healthy controls, to correlate those findings to clinical parameters and to evaluate if qMRI parameters show muscle degeneration in not-yet fatty infiltrated muscles. We evaluated eight thigh and seven calf muscles of 19 calpainopathy patients and 19 healthy matched controls. MRI scans were performed on a 3T MRI including a mDTI, T2 mapping and mDixonquant sequence. Clinical assessment was done with manual muscle testing, patient questionnaires (ACTIVLIM, NSS) as well as gait analysis. Average FF was significantly different in all muscles compared to controls (p < 0.001). In muscles with less than 8% FF a significant increase of FA (p < 0.005) and decrease of RD (p < 0.004) was found in high-risk muscles of calpainopathy patients. Water T2 times were increased within the low- and intermediate-risk muscles (p ≤ 0.045) but not in high-risk muscles (p = 0.062). Clinical assessments correlated significantly with qMRI values: QMFM vs. FF: r = − 0.881, p < 0.001; QMFM versus FA: r = − 0.747, p < 0.001; QMFM versus MD: r = 0.942, p < 0.001. A good correlation of FF and diffusion metrics to clinical assessments was found. Diffusion metrics and T2 values are promising candidates to serve as sensitive early and non-invasive methods to capture early muscle degeneration in non-fat-infiltrated muscles in calpainopathies.


Study population
In total, 19 individuals with genetically and/or histologically confirmed calpainopathy (LGMD R1/D4) (10 females, aged 25-69 years; mean age 39.8 years, 9 males; aged 26-56 years; mean age 36.2 years) and 19 age-and gender-matched healthy volunteers (10 females, aged 26-60 years; mean age 39.2 ± 12.6 years) participated in this study.Clinical information of calpainopathy patients is given in Table 1.In all individuals with only a monoallelic mutation in the genetic testing, diagnosis of calpainopathy was proven by myopathic changes in muscle biopsy specimen and reduced calpain-protein abundance (investigated by immunoblotting).Exclusion criteria for healthy volunteers were medical history of neuromuscular diseases (NMD) and injuries in lower extremity within 12 months prior to study enrolment.
The institutional and committee for approving the experiments was the ethics committee of the Faculty for Medicine of the Ruhr University Bochum (Ruhr University Bochum No. 15-5281).The study was performed in accordance with the guidelines of the ethics committee and line with the Declaration of Helsinki (DVH), the relevant national and international recommendations according to Good Clinical Practice (GCP), §7 Health Professions Act North Rhine-Westphalia and §15 of the Professional Code of the Medical Association of Westphalia-Lip.Informed consent was obtained from all participants.

Clinical assessments
Muscle strength was evaluated using both, the Medical Research Council (MRC) scale and Quick Motor Function Test (QMFT) by an experienced clinician (JF: 5 years of experience, RR: 8 years of experience, AKG: 16 years of experience) 24 .Hip flexion, knee flexion, knee extension, ankle dorsiflexion, and ankle plantarflexion were assessed with a Chatillon Dynamometer® (Chatillon DFE II Dynamometer, Chatillon Force Measurement, AMETEK, USA) for force measurements.Daily life activities were enquired using the ACTIVLIM and the NSS 24,25 .An experienced medical technical assistant performed the following tests in ambulant individuals: the 6-MWT, time to walk 10 m, and timed up-and-go test.

MRI acquisition and sequences
The participants laid in a feet-first supine position.Cushions were used to support participants' knees and sandbags placed around their feet to prevent motion.Scans were obtained using a Philips 3.0T Achieva MR system and a 16CH Torso XL coil.The thigh region from hip to knee was split into two axial fields of view (FOV) of 480 × 276 × 150 mm 3 along the z-axis with a 30 mm overlap and the proximal edge positioned in the crotch.The calf region was scanned with one axial FOV of 480 × 276 × 150 mm 3 .The proximal edge of the FOV was positioned 60 mm below the tibial plateau perpendicular to the tibial bone.The protocol consisted of a 4-point Dixon sequence (voxel size 1.5 × 1.5 × 6.0 mm 3 ; TR/TE 210/2.6,3.36, 4.12, 4.88 ms; flip angle 8°, SENSE: 2), a multi-echo spin-echo (MESE) sequence for quantitative water mapping including 17 echoes and Cartesian kspace sampling (voxel size 3.0 × 3.0 × 6.0 mm 3 ; TR/TE 4598/17 × ∆7.6 ms; flip angle 90/180°, SENSE: 2), and a diffusion-weighted spin-echo EPI (voxel size 3.0 × 3.0 × 6.0 mm 3 ; TR/TE 5000/57 ms; SPAIR/SPIR fat suppression; SENSE: 1.9; 42 gradient directions with eight different b-values (0-600 s/mm 2 ) 26 .A noise scan was obtained using the same imaging parameters as the DWI, but without RF power and gradients (only acquisition channels open).Scanning time per stack (each FOV) was approximately 12 min.

Data pre-processing
Data were pre-processed as described before using QMRITools (www.qmrit ools.com) 26 .In brief, the diffusion data were denoised using a PCA method 27 .To correct for subject motion and eddy currents both legs were registered separately.Then the tensors were calculated by taking IVIM into account and using an iWLLS algorithm 27,28 .The IDEAL method was used for the Dixon data considering a single T2* decay and resulting in a separated water and fat map 30 .The derived water maps were used for the manual segmentation.The T2-mapping data were analysed using an extended phase graph (EPG) fitting approach 31 .

Muscle segmentation and tractography
Eight thigh muscles (vastus lateralis, vastus medialis, rectus femoris, semimembranosus, semitendinosus, biceps femoris, sartorius, and gracilis) and seven calf muscles (gastrocnemius medialis and lateralis, soleus, tibialis anterior, peroneus, extensor digitorum and tibialis posterior) were first segmented in patients and controls using an automated segmentation tool and subsequently optimized by an experienced rater (JF) in both legs 32 .The rater checked the automated segmentation results and manually corrected the muscle shape if necessary.Automated segmentation was not precise enough in patients due to loss of muscle structure in fatty infiltrated muscles.Patient muscles were segmented manually on all 25 acquired slices of the Dixon water images (3D-slicer 4.4.0,https:// www.slicer.org) 33 .
The segmentations were then registered to T2 and DTI data to correct for small motions between sequences and image distortions using sequential rigid and b-spline transformations (elastix, https:// elast ix.lumc.nl 34 .Average values over all slices of water-T2 time and proton density fat fraction (FF) were obtained.SNR was calculated as the local average signal divided by the local noise sigma 35 .For analysis of diffusion data, the preprocessed diffusion data were divided based on the muscle segmentation.Secondly, whole muscle tractography was performed within each diffusion muscle volume.(MRIToolkit) 36 .The following fiber tracking stop parameters were used: maximum angle 15°, step size 1.5 mm, FA range 0.1-0.6 35,36 .The DTI parameters fractional anisotropy www.nature.com/scientificreports/(FA), mean diffusivity (MD), axial diffusivity (λ 1 ) and radial diffusivity (RD) were extracted for each individual muscle using tract-based sampling.

Statistical analysis
FF was compared between LGMD patients and healthy controls in a general linear model with patient/control, body side, and muscle (to control degrees of freedom for multiple test points per subject) as fixed factors for all leg muscles.For the analysis of the diffusion metrics and T2 times, we chose a disease-specific approach in terms of muscle involvement, based on the degree of muscle fat infiltration shown by individual muscle FF in this study and literature review 18 .We chose to analyse qMRI changes in different groups of muscles that have a different risk of degeneration and not in a patient/control approach.
We identified the three following muscle groups with different risks of fat infiltration in calpainopathies 20 : I. high risk: biceps femoris, semimembranosus, semitendinosus, gastrocnemius medialis and soleus II.intermediate risk: vastus medialis, vastus lateralis, rectus femoris, sartorius, gracilis, gastrocnemius lateralis and peroneal group III.low risk: extensor digitorum, tibialis anterior and tibialis posterior To assess differences in diffusion metrics and T2 time between low-fat muscles of controls and LGMD patients, the cut-off value for low-fat muscles was defined as: highest mean fat fraction in healthy controls + 1 SD (~ 8%) 17 .
Additionally, muscles with low diffusion data quality, which was defined by a SNR of lower than 10 before denoising, were excluded for analysis of diffusion parameters 39 .After performing Levene's test for equality of variances, two-sided t-tests for independent samples between LGMD and control group were conducted to evaluate the following hypotheses for muscles with < 8% FF: I.In high-risk muscles we hypothesize that the known fiber atrophy and myocellular damage in calpainopathies 40 could lead to an increase of FA, a decrease of MD and RD and an increase in T2.II.In intermediate-risk muscles, there may be significant differences in T2 and diffusion metrics between the patient and control group.III.In low-risk muscles no significant changes of T2 and diffusion metrics between study groups can be detected.
To evaluate correlations between clinical assessments and qMRI values mean FF, FA, MD, and T2 of all thigh and calf muscles were correlated to the 6-MWT, time to walk 10 m, and timed up-and-go test using Pearson's correlation coefficients.Furthermore, mean qMRI values of thighs and calves were correlated to the results of ACTIVLIM, NSS, and QMFM using Spearman's rank correlation coefficients.Additionally, the qMRI values of the corresponding muscle groups were correlated to the force measurements by dynamometry and MRC scale, i.e., quadriceps muscle and knee extension.
All statistical analyses were performed using IBM SPSS V28.The significance level for all tests was set at p < 0.05.

Results
All scans and clinical assessments were successfully performed.Example images of the applied MRI sequences are shown in Fig. 1.
The average FF was significantly different in all patient muscles than in controls (Main effect: p < 0.001; Bonferroni-corrected post-hoc t-tests for each muscle: p ≤ 0.047).The mean FF per muscle of LGMD patients is displayed in Fig. 2. The tibialis posterior was less affected in LGMD patients, while semimembranosus and semitendinosus showed the highest FF.Average mean FF values per muscle of all LGMD patients are illustrated in Fig. 3.
When comparing diffusion metrics between patient and control group in muscles with less than 8% FF and SNR higher than 10, a significant increase of FA and decrease of RD was found in high-risk muscles of LGMD patients (see Table 2 and Fig. 3).MD showed a nonsignificant decrease in LGMD group (p = 0.121).In low-and intermediate-risk muscles, a significant increase of FA was observed, while the other diffusion metrics showed no differences in comparison to the control group.When comparing T2 between study groups we found an increase in all previously defined muscle groups which was significant in the low-and intermediate-risk groups (p ≤ 0.045) but not for high-risk group (p = 0.062).
Correlations between clinical assessments and qMRI values in LGMD patients are displayed in Tables 3, 4 and Supplement Table S1.A significant correlation between 6-MWT and FF of thigh muscles was observed (r = − 0.691, p < 0.05).For the other qMRI values weak to moderate correlations were found (see Table 3).Timed-up-and-go showed a strong correlation to FA and T2 in thigh muscles (p < 0.01), while 10 m walking-time correlated strongly with FF, FA and MD in thigh muscles and FF and FA in calf muscles.QMFM showed strong correlation with FF, FA, and MD in the thigh and a moderate to strong correlation with those parameters in calf muscles (p < 0.01; see Fig. 4).ACTIVLIM showed higher correlations to FF, FA and MD than NSS (|r|≥ 0.631 vs. |r|≥ 0.372; see Table 4).No significant correlations of questionnaires and QMFM to T2 were found.The force measurements of knee extensors, knee flexors and dorsi extensors showed moderate to good correlation to FF, FA and MD.
The highest correlations were found for FF (see Supplement Table S1).

Discussion
In recent years, qMRI has shown to be a useful tool in diagnostic management and disease monitoring of NMD.The advantages of the combination of MRI measurements and clinical assessments in the disease monitoring of patients with LGMDR1 were first shown by Fardeau and colleagues 41 .In recent years, several semiquantitative MRI studies identified a pattern of fat distribution in calpainopathies showing a predominant involvement of the hamstrings and the soleus and gastrocnemius medialis 19,20,40 .Using Dixon FF, we confirmed those observations in a cohort of 19 calpainopathy patients presented in this study.As Barp et al. described before, relative sparing of quadriceps, sartorius, and gracilis in the thigh and anterolateral compartment of the calf was identified 20 .The least fat-infiltrated muscle was the tibialis posterior displaying a fat fraction of lower than 20% even in highly affected individuals (see Fig. 2) 42 .FF is known to correlate with MFM and force measurements by MRC scale in LGMDR1 and has shown superior sensitivity over standard functional evaluation 21,41 .In this study, significant moderate to strong correlations to different clinical outcome measures were observed not only for FF but also for diffusion metrics FA and MD.Since ACTIVLIM correlated more strongly to FF than NSS, ACTIVLIM questionnaire might be more suitable in monitoring of LGMD.This may be explained by the fact that several items in NSS address orofacial weakness which in LGMD is usually not observed.The highest correlation was identified between FF in thigh muscles and the QMFM suggesting that although being a general assessment QMFM can best reflect degree of fat infiltration.In contrast to the findings of Arrigoni and colleagues, in a cohort of 11 patients with LGMDR1, in our study moderate to strong correlations of diffusion metrics FA and MD to clinical outcome measures were observed 22 .These different results may be explained by methodological reasons: We used a different post-processing algorithm to obtain our DTI data for the whole muscle from tractography.This approach might be more sensitive to detect degenerative changes, as diffusion data can be analysed from the whole muscle volume.In contrast, Arrigoni et al. used an ROI-based method in which the different muscle compartments were segmented on a single slice only.Even with careful placement of the ROI, this method may over-or underestimate DTI values in the case of inhomogeneous muscle degeneration.This can be particularly important when correlating to clinical as well as functional data, as the whole muscle is important for these parameters.Furthermore, in our study all www.nature.com/scientificreports/muscles were segmented separately to avoid inclusion of muscular fascia and connective tissue.Correlations of FA and MD to clinical assessments have been described before and although those findings must be interpreted with caution due to expected changes of diffusion metrics in highly fat-infiltrated muscles, it has been shown that diffusion metrics can provide additional information on muscle status 18,37 .Arrigoni et al. suggested that diffusion metrics and T2 values should only be examined in preserved muscles 22 .Williams et al. analysed the effects of fat infiltration in muscle on DTI parameters and concluded that at least 76% of healthy muscle tissue should be present in a muscle to correctly calculate diffusion parameters, otherwise FA is overestimated 44 .
However, Otto et al. were able to show in a model for fat infiltration in muscle tissue that with adequate fat suppression only the water signal of the muscle is used for DTI and sufficiently sensitive results for musclespecific DTI values can also be achieved in fat-infiltrated muscle 19 .We used the same method of fat suppression for our study.To avoid possible confounding effects of fat infiltration in this study diffusion metrics and T2 values were only evaluated in low-fat muscles.Fatty infiltration in myopathies is not just a confounder, it is also the visible aspect of advanced and non-reversible muscle damage.Therefore, sensitive qMRI biomarkers should capture microstructural changes before fatty infiltration occurs as those would ideally capture a stage where therapies could still be effective 45 .As described, calpainopathies show a typical disease pattern with predominant involvement of hamstrings, gastrocnemius medialis and soleus.Consequently, differences in diffusion metrics and T2 times in this study were assessed in muscle groups that were previously defined as dependent on the risk of fat-infiltration.A significant increase of FA was found in all muscle groups of patients while MD remained unchanged.This may reflect fiber atrophy which is supported by a significant decrease of RD found in this study.A reduction in RD is a sign of a reduced fiber diameter 46 .An increase of FA with an accompanying decrease of MD was also interpreted as a sign of fiber atrophy in NMD 19 .However, Berry et al. showed in simulations that FA changes are more pronounced and precede MD changes with a decreasing muscle fiber size which may explain that no significant differences of MD were found in this study 47 .
A potentially more sensitive approach to measuring muscle fibre diameter was recently published by Tan et al.Using tissue-specific multi-compartment modelling obtained from an orientation invariant dictionary, which models muscle fibres as cylinders with radial constrained diffusivity, myofiber diameter can be measured in µm (AFD-Apparent Fiber Diameter) 48 .
In patients with healthy and denervated shoulder muscles, AFD has been shown to measure myofiber diameter with high sensitivity.The FA increase, RD decrease and unchanged MD shown by Tan et al. in the denervated muscles due to fiber atrophy support that changes of diffusion metrics in non-fat infiltrated muscles of LGMD patients in this study are related to early fiber atrophy.Focal fiber atrophy in calpainopathies has been described in muscle biopsies of still asymptomatic individuals suggesting that fiber atrophy occurs early in course of disease 49 .
Another histopathological finding in muscle biopsy of patients with calpainopathies (but also other muscular dystrophies) is a higher number of eosinophils in comparison to healthy controls which may reflect role of inflammation in pathophysiology 50 .In this study an increase in T2 values in all groups of non-fat infiltrated LGMD muscles was found as an indication of either an inflammatory process or a sign of active muscle degeneration.Correlations of T2 values to clinical assessments were lower than of the other qMRI parameters.Thus, elevated T2 does not translate immediately to clinical function and may precede an impairment of those prior to changes in DTI metrics.However, increase of T2 values can very well distinguish between healthy controls and low-fat muscles of LGMD patients.Therefore, diffusion metrics and T2 values can capture changes in not yet fatty-infiltrated muscles in calpainopathies.

Limitations
There are some limitations in this study.Due to the early onset of the disease (median disease duration: 19 years) most patients in this study had advanced fat infiltration.Consequently, analysis of low-fat muscles was limited due to a low number of preserved muscles and high level of fat-infiltration of predominantly affected muscles in our cohort.Furthermore, being a cross-sectional cohort study, our study did not reflect the role and changes of qMRI values during disease progression.Longitudinal studies are needed in the future which ideally should include less affected patients.

Conclusion
Dixon FF values confirm the predominant involvement of hamstrings in the thigh and soleus and gastrocnemius medialis in the calf with relative sparing of anterior muscle group.A good correlation of FF but also diffusion metrics to clinical assessments were found.Diffusion metrics and T2 values are promising candidates to capture early muscle degeneration in non-fat-infiltrated muscles in calpainopathies.

Figure 1 .
Figure 1.Example images of the applied MRI sequences.mDixon fat fraction (FF), fractional anisotropy (FA), mean diffusivity (MD) and T2 maps for thigh and calf muscles of two representative LGMD patients, and a healthy control (HC).

Figure 3 .
Figure 3. qMRI data in low-fat muscles.Overview of mean fat fractions of all LGMD patients in thigh and calf muscles (a).High-risk muscles are coloured in yellow and orange, intermediate-risk muscles are coloured in green and low-risk muscles are coloured in blue.Muscle groups with significant differences of FA and MD in muscles with FF < 8% and SNR > 10 and T2 in muscle groups with FF < 8% between study groups are coloured in red (b) (increase/decrease in patient group: burgundy / pink).

Table 1 .
Demographic and clinical data of patients with calpainopathy.*Diagnosis confirmed via muscle biopsy and reduced Calpain-3 protein abundance investigated by immunoblotting.

Table 2 .
Mean differences and p-values for two-sided t-tests for independent samples between patients and controls for the different muscle groups.Significant results are marked with an *. p < 0.05 was defined as statistically significant.