Ultrasound elastography: compression elastography and shear-wave elastography in the assessment of tendon injury

The Achilles tendon is the strongest and thickest tendon in the human body, has a superficial location and good differentiation from surrounding tissues, making it particularly suitable for USE assessment [4]. A total of 16 studies (11 using CE and 5 using SWE) were found in the assessment of abnormal Achilles tendon.

Regarding CE, it was demonstrated almost perfect intra-observer repeatability and variable inter-observer repeatability [64‐66]. Regarding SWE, it revealed high reproducibility and moderate to high repeatability [64, 67, 68].

Normal Achilles tendons have a homogenous stiff pattern, whereas pathological ones exhibit a more heterogenous and reduced stiffness pattern [12, 16, 27, 31‐34, 52‐54], except for one study, using CE, that found symptomatic tendons to be harder than the asymptomatic ones [28]. Studies using CE showed moderate to perfect correlation of USE with conventional US in Achilles tendinopathy evaluation [27, 28, 31, 32] and excellent correlation with functional scores, greater than conventional US [34]. The application of CE demonstrated superior [34] or comparable results [27] in terms of specificity, sensitivity and accuracy compared to conventional US in the diagnosis of clinically symptomatic Achilles tendinopathy. A study by Klauser et al. [31] showed perfect agreement of CE with histology (Figs. 3 and 4), along with moderate agreement between conventional US and histology.

The use of SWE revealed high specificity [12, 53] in the detection of tendinopathy and the combined use of conventional US is recommended to increase the sensitivity of the diagnosis [52]. Coombes et al. described that a low Achilles V_s is correlated with higher age, self-reported pain and disability over the preceding week, and with decreased loading capacity of the Achilles tendon [53]. Studies in athletes demonstrated that tendon softening may predict pain and tendinopathy [38, 69‐71]. Using conventional US, Fredberg et al. identified athletes with the greatest risk of developing tendon disorders [72]. Hence, USE may be useful in the detection of subclinical tendinopathy, providing time for prophylactic procedures (reduction of the load and eccentric training, for example). Once rehabilitation is considered the standard treatment for restoring the original mechanical properties of the tendon, it is crucial to assess the relation between clinical functional outcomes and the mechanical properties of the Achilles tendon.

In case of a tendon rupture, loss of tension, haematoma or effusion contribute to a significant decrease in stiffness, as verified by Chen et al. [54] and Aubry et al. [12], who identified signal-void areas at SWE. Domenichini et al. state that dynamic SWE is recommended for rupture diagnosis, since a totally ruptured tendon will not become taut when stretched [73].

USE was studied after the surgical repair of Achilles tendon rupture [17, 29, 30, 55]. In this scenario, there is an increase in tendon stiffness and heterogeneity according to the physiological healing process, probably due to the structural disorganisation of collagen fibres, with predominance of type III collagen instead of type I as usual. An increased stiffness in the contralateral tendon was also reported [74, 75] and that can be due to overload during the rehabilitation period or a sign of predisposition to tendinopathy [17]; however, the explanation remains unclear. Using SWE, Zhang et al. showed that, in addition to increased tendon stiffness, there was a positive correlation between the degree of tendon functionality and its elasticity, suggesting that lower elastic modulus values might predict poor mechanical properties, and functional outcomes, with worse healing of repaired tendons [55]. It is known that a repaired Achilles tendon never regains its normal US appearance. Therefore, USE is useful for monitoring tendon repair and rehabilitation, eventually avoiding re-rupture from early tendon loading [76].

Gehmert et al. carried out a laboratory study in rabbits demonstrating, through USE evaluation, that the application of mesenchymal stem cells in a ruptured Achilles tendon can totally restore their elastic properties [30]. Another recent laboratory study by Yamamoto et al. with Achilles transected tendons in rabbits described that, at the healing site, there was a gradual decrease of the strain ratio and the tendon became significantly harder over time [36]. They also demonstrated a correlation between the strain ratio and mechanical and histological properties of the healing tendon tissue. According to these findings, USE may reflect biological healing processes and, in that way, contribute to define the appropriate time to return to previous activity with lower risk of re-rupture.

USE has also been studied in systemic diseases. A stiffness reduction of the Achilles tendon was observed in patients with ankylosing spondylitis [32], in diabetic patients with foot ulcers [33] and in acromegaly patients [35].

The reproducibility of the CE technique in healthy patellar was characterised with good intra- and inter-observer reliability by Porta et al. [77]. Excellent intra-observer reliability and moderate to excellent inter-observer reliability were reported for SWE [56, 68, 78].

Conventional US shows hypoechogenic regions in a tendon with patellar tendinopathy. However, asymptomatic individuals can have abnormal tendon morphology, including hypoechogenic regions at conventional US that do not help to predict the subsequent development of symptoms [79, 80].

Five studies using USE were found (two using CE and three using SWE) and the results were discrepant concerning elasticity characterisation. Three studies had results favouring a decrease in stiffness [37, 38, 52] and two studies, using SWE, described an increase of V_s and, therefore, stiffness in patellar tendinopathy [53, 56]. There are important methodological differences that can explain this disparity. It should be emphasised that different regions of the same tendon were assessed and different sizes of ROIs were used: Dirrichs et al. measured V_s using an ROI of 1 mm in diameter placed at the most rigid area of the tendon [52], Coombes et al. used a larger ROI and measured V_s at the middle and proximal parts of the tendon [53] and Zhang et al. [56] measured V_s at a proximal part. It has been shown that the elastic modulus is influenced by the size of the ROI (Kot et al. reported that the elastic modulus of the patellar tendon increased with increasing size of the ROI [81]), as well as by the assessed tendon portion [15, 77, 82].

The patellar tendon is connected to two hard and fixed structures (patellar bone and tibial tuberosity), contrary to the other studied tendons, which are connected on one side to a softer and more compliant structure (muscle). This fact is stated by Porta et al. as an explanation of their results, in which the healthy patellar tendon was characterised by a soft pattern using CE [77]. Therefore, it is not clear which pattern and elasticity values are able to differentiate a healthy patellar tendon from a patellar tendinopathy.

Zhang et al. assessed 13 athletes with unilateral patellar tendinopathy using SWE and found a correlation between increased stiffness of painful tendon and the intensity of pain and degree of dysfunction [56]. Using CE, Ooi et al. described symptomatic patellar tendons in athletes as softer than asymptomatic ones and better related to functional scores than conventional US. The same study showed that USE increased conventional US sensitivity and accuracy in the diagnosis of patellar tendinopathy [38].

Patellar tendinopathy is common among active athletes [83], so USE may be employed to assist sports medicine clinicians, providing a more effective rehabilitation in athletes with patellar tendinopathy.

Repetitive microtrauma is considered the main factor that leads to epicondylar tendinopathy. In lateral epicondylitis, extensor carpi radialis brevis (ECRB) is the most commonly affected muscle, but supinator and other wrist extensors can also be involved. Any activity involving excessive and repetitive use of these muscles can result in multiple microtears, leading to a cycle of tendon degeneration and repair [84‐86]. The use of conventional US to differentiate the tissue affected by degenerative disease from the healthy tissue can be challenging because, often, both have the same echogenicity [86, 87]. Seven studies (six using CE and one using SWE) showed a decrease in stiffness pattern on symptomatic tendons [39‐44]. For instance, Fig. 5 represents an elastogram and US images of a normal and a pathologic epicondylar tendon. The focal hypoechoic area in the deep part of the ECRB tendon, corresponding to collagen degeneration with fibroblastic proliferation, is the most common US finding of lateral epicondylitis [88]. Using CE on lateral epicondylitis, De Zordo et al. found slightly more focal lesions and a more frequent involvement of the lateral collateral ligament and the peritendinous fascia compared to conventional US [39]. Lateral collateral ligament tear and large intra-substance tears are related to poor outcomes when using non-operative treatment. In these cases, autologous blood injection, platelet-rich plasma cell therapy and surgery are approaches to be considered [89].

Park et al. described that USE findings were related to pain during resistive middle finger extension on physical examination, meaning that the superficial group was affected, and, therefore, the combination with conventional US may add information to differentiate the injured group fibres [41].

Three studies concluded that USE was superior to conventional US in terms of accuracy and/or sensitivity [39, 41, 42]. Klauser et al. studied epicondylar tendons in cadavers and concluded that the combination of USE and conventional US provided better correlation with histology than the use of both modalities alone [43, 44].

Correlations of USE findings with clinical symptoms scores [52], symptom duration and history of corticoid injection were found [41]. Studies addressing elasticity changes on follow-up and rehabilitation treatments were not found for epicondylar tendinopathy. The functional outcomes and the relation with USE should be further evaluated in future research.

In our research, five studies used CE and seven studies used SWE in cases of rotator cuff tendinopathy or supraspinatus tendon tear. Good to very good intra- and inter-observer repeatability were reported for measurement of V_s in the supraspinatus muscle and tendon [58, 90].

Globally, a stiffness decrease and a heterogeneous pattern of softness are described in affected tendons compared to normal ones (Fig. 6) [46‐48, 58, 60, 91]. This makes USE particularly interesting when conventional US cannot differentiate between healthy and pathological tissue. The findings on USE were correlated to magnetic resonance imaging (MRI) [45, 46, 49, 62] and conventional US [45, 46, 60]. Rosskopf et al. demonstrated that V_s decreased with increasing fat content and supraspinatus muscle atrophy, and was increased in the final stage of fatty infiltration [58]. Krepkin et al. found that V_s was negatively correlated with T2* weighted values measured on MRI sequences and tear size of the degenerated supraspinatus tendon, reflecting the tendon quality [62]. Seo et al. also stated that CE was able to quantify the severity of the fat atrophy of the supraspinatus tendon, reporting excellent accuracy and inter-observer reliability [45]. Since supraspinatus fatty infiltration and muscle atrophy can predict repeated tears rates and poor functional outcomes, it must be acknowledged that USE may confer prognostic utility in surgical repair decision [92, 93]. Hatta et al. found that quantitative SWE assessment of the supraspinatus muscle was highly correlated with extensibility of the musculotendinous unit on cadaveric shoulders. Therefore, USE could offer a non-invasive method to predict rotator cuff extensibility during the preoperative planning phase for a rotator cuff repair [61]. Hou et al. also hypothesised that SWE may be useful in the preoperative setting for patient selection and surgical planning, as it can reflect tendon quality or post-operative failure rates [60].

Muscle stiffness can be related to rotator cuff tendinopathy. An increased stiffness on the upper trapezius muscle assessed with SWE was associated with rotator cuff tendinopathy in a study of volleyball players by Lin et al. [57]. These authors suggested that athletes with increased shear modulus on the upper trapezius may have higher risk of developing rotator cuff tendinopathy and, consequently, USE applied in the upper trapezius can be employed for prevention/rehabilitation purposes. Another study found that deltoid muscle softening had a correlation with tendinopathy severity assessed by conventional US [60].

The evaluation of 39 patients with calcifying tendinopathy using SWE allowed to define a non-dark pattern that was a predictor of symptomatic relief after fine needle aspiration. These results reflect the utility of USE in the evaluation and management of rotator cuff calcific tendinosis [57].

Tendinopathy of the long head of biceps tendon was characterised with CE in one study. The authors found that the affected tendons had softening areas on USE, with a positive correlation with conventional US [50].

The quadriceps tendon was also characterised with CE in patients undergoing chronic haemodialysis and it was described as being thinner and softer with greater colour heterogeneity on the elastogram compared to the controls [51].