T2 mapping is a qMRI technique that measures T2 relaxation of tissues, that reflects the locoregional water binding and that is prolonged by the presence of increased free water (20, 65–67). The T2 time is characteristic for each tissue and is affected by a number of physiologic and pathologic processes; in muscle imaging it has been used to explore the effects of exercise (65, 68, 69), and to specifically evaluate neuromuscular diseases, also after exercise (70–74). It must be underlined that, given known modification of T2 relaxation time after exercise, performing an MRI shortly after physical activity may significantly alter the signal; an appropriate resting time is therefore recommended.

T2 measurement is an attractive endpoint because it is primarily affected by early changes in muscle integrity, before fat infiltration occurs: in other words, the T2 value reflects the prodromal phases of the disease in the muscle. As a consequence, the T2 is an in vivo indicator of ongoing disease activity. Nonetheless its low specificity has to be kept in mind: an increase of T2 signal may be due to several unspecific events which all lead to intracellular or extracellular edema or a combination of both (29).

The difference between global T2 and water T2 must be specified. Global T2 signal is strongly influenced by fat infiltration, which is better described by fat-fraction evaluation techniques as Dixon (see above); the use of global T2 can thus lead to confusion and is therefore discouraged. What is really interesting in the setting of muscle diseases imaging is water T2, that reflects the ongoing pathologic changes in the muscle (75).

In DMD T2 was found to be increased in involved muscles in a series of studies (71, 76–82). Moreover, T2 and related variables were shown to increase over time and to strongly correlate with changes in functional performance (83, 84).

Studies combining qMRI and magnetic resonance spectroscopy (MRS) in DMD demonstrated interesting findings in the early detection of muscle involvement and in the assessment of steroid therapy effects. Hooijmans et al. showed that both T2 and PDE (phosphodiesters, see ³¹P-MRS section below) are already increased prior to fat replacement and remain elevated in affected muscles. Both measures could therefore serve not only as early markers of disease, but they could also potentially track response to therapy (85).

As a proof of principle, Arpan et al. in fact, demonstrated, through a combined quantitative MRI/MRS approach, that both quantitative MR techniques are able to effectively monitor the effects of corticosteroids on the lower extremity muscles of young boys with DMD. In treated subjects, T2 values and FF decreased when compared to untreated boys, and progressive fat replacement was slower during treatment, at least in a few muscles. The monitoring of T2 was able to detect therapeutic effects as early as 3 months after drug initiation (48).

Differently from DMD, BMD subjects did not show a T2 increase in affected muscles (86), likely in keeping with the minor degree of inflammation. Conversely, high T2 values were found in residual muscle tissue of myotonic dystrophy (DM) 1, with the highest values reported in the more advanced stages of disease, the increase being associated to a decrease in muscle force (72).

Finally, increased T2 values were also found in inflammatory myopathies, where they again reflect the degree of inflammation (58, 73, 87, 88).

Another available technique based on relaxometry uses T1 signal decay instead of T2. T1 values in muscle imaging are known to be increased in case of inflammation and strongly decreased where fat-replacement occurs (89–92); on the contrary, T2 values are elevated in both conditions (93).

Fast T1-mapping sequences have been applied in the study of myocardium (94), whereas only a few studies applied T1 mapping in skeletal muscle imaging in healthy subjects (69, 95), demonstrating a high repeatability.

T1 values, along with T2 and T2^* mapping, were found to be significantly increased in a small group of FSHD subjects (n = 7) and, in parallel, they correlated well with clinical severity scores (96). Gaur et al. failed to show a T1 abnormality in upper arms in DMD compared to controls (82).

Marty et al. recently explored muscle T1 relaxometry on healthy volunteers and compared its sensitivity with the standard three-points Dixon sequence to assess fat replacement in 30 BMD subjects: the measured T1 values showed a high repeatability, correlating well with FF as measured by Dixon sequence, and could discriminate involved muscles from normality. Additionally the acquisition time was quite fast, with a 10 s scan time per slice (97).

Despite being the most used and studied in the last decades, the applicability of ³¹P-based MRS in skeletal muscle imaging is nonetheless limited by the need of dedicated equipment, generally not available in a typical clinical setting (Figure 4).

Phosphorus (P) MR spectroscopy has long been used mainly because of its ability to detect signals from biological molecules important in the cellular energetic processes, including phosphocreatine (PCr), adenosine triphosphate (ATP), and inorganic phosphate (Pi) (126–129). Other molecules that are often studied are phosphomonoesters and phosphodiesters (PDE) that are products of degradation of membrane phospholipids (130, 131). Finally, though in an indirect way, also tissue pH can be estimated with this technique (132).

Some studies suggest that P-including metabolites or ratio of metabolites can be used as early disease markers, before significant fat replacement occurs (85, 133). In particular PDE-levels appear as a promising biomarker (134).

In DMD, MR spectroscopy has been applied both to evaluate cellular metabolism and to measure intramuscular lipid composition (85, 135–137). Studies dating back to late eighties and nineties documented decreased phosphocreatine/inorganic phosphate (PCr/Pi) and PCr, but increased Pi in gastrocnemius muscle of DMD children (128, 138). More recently Torriani et al. showed that PCr/Pi of posterior lower leg muscles was significantly lower in DMD children, suggesting that it may be used as a measure of bioenergetic reserve and as a marker of cellular oxidative metabolism (135). They also found elevated muscular pH, in line with previous experiences in literature (128, 138), and suggested that pH increase in DMD may be a response to dystrophin deficiency, possibly related to regenerative mitogenesis or altered ion transport. Importantly, Torriani et al. did not find any correlation between strength and motor functions and spectroscopy parameters, differently from a few small studies conducted previously that went in that direction (nonetheless the small sample sizes of those studies effectively prevented any significant conclusion) (137, 139). As authors suggested, in the context of small sized cohorts, metabolite profiles as defined by MRS may not be as sensitive as MR imaging and measures of fat replacement in predicting disease severity (135).

Hooijmans et al. (134) evaluated calf muscles in DMD subjects with ³¹P MRS at 7 Tesla and found that extensively fat replaced muscles had some metabolite ratios, including PDE/ATP, significantly increased compared to controls. Additionally, only the PDE/ATP ratio (and T2 values, see dedicated section above) was increased in muscles with only a minor fat replacement (85), suggesting PDE/ATP ratio may be used as a biomarker of mild/initial disease. The same group documented that in DMD PDE levels increase 2-fold in muscles with different levels of involvement over 2 years time, supporting the hypothesis that PDE-levels may increase very rapidly early in the disease process and remain elevated thereafter (134).

Phosphorus MRS of the upper limb in patients with DMD has also been attempted, being able to discriminate between patients and controls and also defining a specific pattern of disease evolution in non-ambulant subjects (49, 140).

In BMD Wokke et al. found higher PDE/ATP ratios in several not fat-replaced muscles, whereas PDE/ATP ratios were higher in all measured muscles with increased fat levels compared to healthy controls. Tissue pH was also found increased in all muscles of BMD subjects compared to healthy controls (133). Tosetti et al. found metabolites abnormalities also in mild BMD subjects, similar to previous findings in more severe BMD subjects (128, 141, 142), confirming an increased intracellular pH at rest in BMD (even though the meaning of such a finding is unclear) (143). In particular, the slight increase in the PDE peak could represent an early catabolic marker of disease (144), consistent with other muscular dystrophies (145).

The evaluation of pre- and post-exercise cellular energetics in DMD has not yet been performed; in BMD or DMD/BMD carriers anomalies were shown, but with equivocal results showing bidirectionally abnormal glycolysis, but no real impairment of oxidative phosphorylation (127, 128, 135, 141, 142, 144, 146).

Altered metabolic patterns have also been reported in myotonic dystrophy (147), LGMD (145), and FSHD, generally corresponding to either muscle degeneration or regeneration (148, 149). Furthermore, investigators have shown that ³¹P-containing metabolite concentrations correlated with the extent and rate of fat replacement and strength in FSHD (150).

Finally, also in subjects with inflammatory myopathies, ³¹P MRS showed an impaired phosphorous metabolism during exercise, that tended to normalize after treatment with steroids (90, 151). Other studies, however, have suggested a limited role for such metabolic abnormalities in inflammatory myopathies as sIBM (152).

Proton (¹H)-MRS does not require any additional hardware or software to be used in the most widely available MR scanning systems, hence it can easily implemented in clinical studies. Proton spectroscopy is now considered a reference standard for non-invasive fat quantification (153, 154). As fat replacement is a constant pathologic feature of dystrophies and most myopathies, the ability to reliably quantify fat is of primary importance: however, it must be reminded that for a correct application of MRS the water content within the muscle must be consistent throughout, a requisite that is obviously not respected in case of ongoing inflammation/edema, commonly occurring in myopathies (79, 155).

Numerous studies have demonstrated that ¹H-MRS is a useful method to quantify intramuscular lipid concentrations in dystrophic patients, especially in DMD (135, 137, 156, 157), also showing that MRS profiles may serve as biomarkers in disease progression in longitudinal studies (46). Arpan et al. evaluated with proton MRS the effects of corticosteroids in boys with DMD and demonstrated that MRS profiles can monitor the effects of therapy (48).

In subjects with IIM, proton spectroscopy showed that creatine (Cr) concentrations in normal appearing muscles on T1 and STIR images are increased compared to healthy controls, suggesting that changes in MRS profiles may precede pathological changes on conventional MRI (158).

In conclusion, to date, no pathognomonic proton MR spectroscopy pattern has been found to identify a specific myopathy, as fatty degeneration is the last step of them all (132). In such a context, spectroscopy of other nuclei may rather be a useful tool in the assessment of damage even before fat replacement and/or functional impairment occur, and may also track treatment response.

Sodium (Na) plays a vital role in cellular function and integrity; it is found both in the intracellular and extracellular space, but its extracellular concentration is 8–10 times higher than the intracellular one, the sodium-potassium (Na⁺-P⁺) pump being the responsible agent for the maintenance of the concentration gradient. In myocites the abnormalities in total sodium concentration and in its trans-membrane flux rates occur in several physiological and pathological conditions, including exercise, cellular proliferation, and myopathies (159–162).

As sodium is less represented in tissues than hydrogen, Na MRS has an average SNR that is extremely lower than that of proton MRS (additionally depending on organs). Such issue has been at least partially resolved by the progressive implementation of higher magnetic field scanners, up to 9.4 T (163), which increase the SNR.

With regard to DMD, the sodium concentration has been shown to be elevated in the myoplasm, with an osmotic effect causing intracellular muscle edema (164). In a pilot study, Glemser et al. investigated the effects of eplerenone in two DMD subjects and found that intracellular-weighted sodium signal and muscular sodium concentrations decreased with therapy (along with edema), suggesting a role of sodium quantification by MRS as a tracer of disease activity/response to therapy (165).

Myotonic dystrophy has also been linked to sodium channel conductance dysregulation, which may pathologically increase muscle fiber concentrations, influencing disease severity (161); in addition to this, the dysregulation was even incremented after exercise (162).

As aforementioned, the development of sodium spectroscopy at higher magnetic field strengths of 3T or 7T allows a more elevated SNR and thus a more precise quantification of intracellular ²³Na homeostasis in healthy volunteers and in myopathies. The potential of ²³Na MRI at ultra-high fields can be expected to be promising in the near future.

No studies have used carbon spectroscopy yet to examine either inherited or acquired myopathies. Nonetheless, this technique has been used to investigate skeletal muscle metabolism, mainly to determine glycogen levels in normal subjects and in altered glycogen storage diseases (166–170).

As already seen for sodium and phosphorus, carbon imaging is, however, quite challenging, as it requires dedicated equipment uncommonly found in a clinical setting.

Multi-voxel MRS is a technical advancement in spectroscopic studies, largely applied in the neuroradiological domain, which may possibly play a role in muscular imaging, given its ability to simultaneously image multiple voxels in a single acquisition. Traditional single-voxel spectroscopy, in fact, suffers the same issues as a routine muscular biopsy, as it focuses on a chosen region of interest, whereas the disease could involve an entire muscle or different parts of it in a heterogeneous manner. In this sense, the single-voxel approach could miss areas of disease pathology, showing false negative results.

Previously multiple-voxel MRS required a number of technical steps that significantly increased scan time, thus greatly limiting its applications: the recent implementations of various fast MRS techniques (as multiple-echo acquisition, echo-planar spectroscopy imaging, and parallel encoded MRS), however, has fortunately reduced the total scan time, making the multi-voxel approach more attractive, also in a clinical setting. As for the central nervous system, the obtained metabolites spectra can then be plotted and superimposed to anatomical images of muscles for a better overall depiction of the real involvement of rather large anatomic regions.

Even if functional MRI (fMRI) is immediately associated to functional studies of the brain, this technique has also been applied to other organs, including skeletal muscles, to assess microvasculature (187–189). The BOLD (blood oxygen level dependent) contrast used to build the fMRI signal, in effect, is based on alterations in hemoglobin oxygenation: the paramagnetic deoxyhemoglobin, in fact, distorts the local magnetic field, consequently influencing the local signal.

The skeletal muscle BOLD signal reflects the muscular microcirculation and can be regarded as a non-invasive tool to indirectly evaluate muscular function (188, 190, 191). BOLD fMRI has been applied to examine diseases causing altered perfusion of skeletal muscles (192, 193) and to assess the brain response to motor stimulation in myotonic dystrophy, but not at muscular level (194). Despite the known muscle degeneration detected in dystrophic patients, no altered microcirculation has been found in these patients so far.

Arterial spin labeling (ASL) is an advanced MR technique for measuring local perfusion without administration of intravenous gadolinium, mostly applied in brain imaging, particular in a pediatric setting. Skeletal muscle ASL has been validated several years ago (195), but a few more recent ASL studies always of normal muscles are available even at higher magnetic field up to 7 Tesla (196–198). Reports of clinical application of ASL in a pathologic context are scarce, including compartment syndrome and peripheral artery disease (199–201). Important limitations include known sensitivity of ASL technique to motion artifacts, low contrast-to-noise ratio, and relatively long acquisition times (202).

Contrast administration for MR muscle imaging is an off-label practice in clinical context of myopathies and its availability and applicability vary by jurisdiction and over time (203).

Gadolinium has been thoroughly used as a marker of fibrosis in cardiac imaging, where the replacement of myocardial cells is associated with the expansion of the interstitial space (204). Delayed contrast enhancement in cardiac imaging is a powerful marker to identify myocardial replacement by connective tissue. Yet a prerequisite for this approach is the absence of extracellular edema or any situation that could cause pathological enhancement of myocardial cells, such as necrosis or inflammation. In such cases, delayed enhancement is not reliable in documenting fibrosis.

Gadolinium enhancement could be applied to indirectly evaluate fibrosis in the skeletal muscle as well, but no studies are currently available in literature, probably due to several technical issues including the smaller extracellular space and lesser degree of fibrosis characteristic of skeletal muscle (29).

The applicability of electrical muscular stimulators (EMS) that are MRI-compatible has been recently demonstrated (205, 206). As a consequence, interest has risen regarding the possibility to study the movement-linked micro-changes in muscular tissue, using MR sequences that are synchronized with the electrical stimuli. Deligianni et al. demonstrated that both qualitative and quantitative parameters that have a significant dependence on stimulation levels (i.e., stimulation current) can be extracted from the integration of EMS-based stimulus and MRI (206). A wide range of triggered MRI sequences, including conventional cardiovascular sequences, can be used in combination with this technique; the possible scenarios of application of stimulus MRI are quite broad.

The human eye is capable to detect the visual patterns such as roughness and smoothness of an image, and specifically in diagnostic images, although with physiological limits that prevent the detection of subtle variations, such as those of thousands of gray levels, which is something a computer can effectively do.

Texture analysis is a computerized method that can detect the signal from each pixel of an image and describe the spatial variations of pixel signals, allowing a deeper insight into the image structure that may be impossible to detect visually; this way, texture analysis can indeed reflect in detail the microstructural heterogeneity of signal with a good sensitivity, in a non-operator-dependent way. This approach has been applied in several imaging modalities, to precisely characterize even the smallest variations of tissue homogeneity (207).

In 1999, Herlidou et al. demonstrated that texture analysis could discriminate between healthy subjects and dystrophic patients muscles with a sensitivity of 70%, and a specificity of 86% (208). Arpan et al. built T2 signal histograms of lower leg muscles in DMD and showed that they can represent an alternative approach to quantify disease involvement. A threshold was applied to quantify the percentage of affected areas: when compared to conventional method using mean T2 muscle signal, the supra-threshold pixels could better monitor involvement in dystrophic muscles. Therefore, the authors suggested that such a method may be more sensitive in the follow-up of minor pathologic linked to the disease evolution (77). For instance, T2w- and contrast-enhanced T1w-based histogram analysis has been implemented recently to assess the radiation changes in internal obturator muscles after radiotherapy for prostate cancer (209) and in paravertebral muscle of symptomatic lumbar canal stenosis (210).

Some studies applied histogram analysis to muscle ultrasonographic images, using different measures of tissue homogeneity, in healthy subjects (211, 212); others applied it to muscle disorders and found that disease involvement (here DMD) could effectively be well-described with such a technique and with a high level of detail (213). In a more recent study of muscle US, Sogawa et al. showed that texture analysis of ultrasonographic data can differentiate between neurogenic and myogenic diseases being promisingly useful in the non-invasive assessing of the underlying disease mechanisms (214).

Several limitations, however, apply. First, this analytic approach is computationally demanding and needs dedicated software. As already underlined by Carlier et al. texture analysis additionally depends on the used voxel size. A large voxel would, in fact, delete subtle differences and make the image too homogeneous with consequent loss of information, a small voxel would be largely influenced by rumor with possible confounding effect on the texture (29). Furthermore, the whole analyzed region has to be chosen above a certain size, to limit the effect of partial volume (209, 215). A harmonization of acquisition parameters between centers who apply such a technique is thus required; if correctly implemented and regulated, however, image texture analysis of MR or US images can be regarded as a promising diagnostic tool (216, 217).

In addition to its central importance in oncological imaging, positron emission tomography (PET) has nevertheless gained an increasing role in the detection of muscle inflammation, also in the clinical setting. A flogistic process of whatsoever origin is, in fact, associated with increased uptake of ¹⁸F-FDG. Yet, even if PET sensitivity in the detection of high-grade inflammation (e.g., acute infection) is high, the sensitivity in the context of inflammatory myopathies is relatively low. In such cases the combination of PET and MR imaging (i.e., T2-weighted sequences to detect edema) are recommended to increase the accuracy of diagnosis (218–220).

To date no study including a large cohort has been performed to highlight the value of ¹⁸F-FDG PET in the setting of non-inflammatory myopathies, the main reason being the high availability of conventional US and/or MRI imaging and the general tendency of physicians to avoid radiation exposure in non-oncologic subjects.

Liau et al. presented a case of a man with dermatomyositis in which PET/CT imaging revealed diffuse proximal muscle hypermetabolism, consistent with myositis inflammation (221). Renard et al. reported a case of DM in which the initial FDG-PET showed diffuse, increased, proximal predominant, muscular uptake, later normalized after steroid therapy (222). Others investigated the value of PET in different types of inflammatory myopathies, yet only in four cases: two subjects with low-grade myositis and one cases of sIBM were negative at PET examinations, whereas a case of DM showed diffuse proximal muscle hypermetabolism (223): the authors comment that FDG-PET is able to detect more severe myopathies when necrotizing/inflammatory changes occur, helping in demonstrating response to treatment. Nevertheless, as aforementioned, the role of PET in low-grade disease needs further assessment, due to reported low sensitivity in mild cases.

In 20 subjects, Tanaka et al. showed that FDG uptake can discriminate PM/DM from non-muscular diseases and is highly sensitive in detecting muscle inflammation in proximal muscles: moreover, the FDG uptake correlated well with muscle weakness and with the presence of inflammatory cells at biopsy (224). Other authors instead underlined how visual assessment of FDG uptake and the quantification of standardized uptake value (SUV) can have a significant role in the clinical management of PM/DM, allowing a better understanding of the diseases (225). In a larger cohort of 38 IIMs, Li et al. found that muscular FDG-PET uptake was generally higher in patients than controls, and that muscular uptake correlated with muscle strength and serum creatine kinase (CK) levels. Furthermore, PET examination, whose primary purpose was to rule out any malignancy in the context of IIMs, succeeded in detecting a tumor in roughly 20% of cases (226).

Also an interesting, large-cohort, single-photon emission computed tomography (SPECT) study applying ^99mTc-PYP (technetium pyrophosphate) in 90 IIMs subjects confirmed that radiotracer uptake was significantly higher than in healthy controls (227).

As per brain imaging, other PET tracers such as Pittsburg compound B (PIB) may also play a role in muscle imaging. A few experiences in the detection of amyloid plaques by PIB-PET in myocardium or in skeletal muscles in sIBM were remarkable (228, 229).

Finally, the recent implementation and increasing diffusion of new PET/MR scanners as the hybrid imaging technique of choice beyond the traditional PET/CT imaging, will probably improve the imaging process and promote an advance in knowledge also in the context of myopathies and muscle diseases in general (230).

Muscle MR scans are usually focused on certain ROIs for a number of reasons (firstly due to time constraints limiting too long acquisitions) based on the clinical suspicions. The technical implementation of whole-body MRI (WB-MRI) has overcome such “limitation,” making it possible to contemporarily study the entire body in a reasonable time (typically 1 h or less) (231, 232), assessing at a time fat and muscular volume (23, 233–236).

WB-MRI can effectively provide information that is enormously useful to clinical evaluation, detecting all areas of inflammatory involvement or fat replacement throughout the body, greatly helping in the diagnostic process (237). It consequently allows a better tailoring of eventual treatments, and, last but not least, a better monitoring of their efficacy, in a reasonably short time. WB-MRI can also serve as guidance for muscular biopsy possibly even more than conventional MRI, given the wide field of view, in particular for uncertain cases.

Conversely, the analysis of WB-MRI datasets may represent a difficulty, given the large amount of acquired data. To address such an issue, Hankiewicz et al. have proposed the use of a graphical technique (so called heat maps), already applied in other scientific disciplines, which provides a visual representation of each muscle of the body (Figure 5). Hence, a visualization of the overall condition of muscles of a subject at a glance is possible (23) and involved and/or spared muscles become immediately evident at visual inspection (23, 238). Such a reader-friendly visual presentation of data, in fact, could make more evident or reveal subtle relationships, tendencies, and involvement patterns (238), building a sort of a “fingerprint” of a subject, and, as a further step, of a specific disease. The so-built heat maps can thus be applied to promptly recognize a myopathic pattern and to facilitate uncertain diagnosis (236, 238, 239).

To date applied protocols for WB-MRI generally include traditional sequences as T1, T2 and STIR; nonetheless an effort is made to extend studies including DWI and other advanced techniques, with promising scenarios.