The imaging of tendons with ultrasound has been utilized in the clinical setting to assist in the diagnosis of tendinopathy, monitor the efficacy of treatments, and assess the risk of developing symptoms. However, while imaging shows the presence and extent of structural changes within the tendon, the clinical interpretation of the images requires context in regard to the features of pain (location and distribution) and the aggravating factors (loading related to increases in pain).
This is due to the limited relationship between structural disorganization and pain, similar to other musculoskeletal conditions. Clinically, tendon abnormalities on imaging do not confirm that pain and dysfunction are generated by the tendon.
An addition to the old saying that a picture is worth a thousand words—but only as an adjunct to the clinical picture.
This clinical commentary discusses the imaging features of the pathological tendon, accuracy and sensitivity of ultrasound, the role of imaging in the clinical and research settings (prediction of symptoms and monitoring), and future directions.
Features of “Normal” and “Pathological” Tendons on Imaging allows for the visualization of the internal architecture of the tendon. Normal tendon primarily contains type I collagen that is hierarchically arranged into parallel aligned fibrils, fibers, and fascicles. This uniform alignment of fibers can be visualized in normal tendons using ultrasound imaging at the fascicle level.
Four main histological changes are observed in tendinosis.
- the primary change being increases in cell numbers that exhibit an altered, more metabolically active phenotype. Rather than flattened tenocytes that are arranged parallel between fibers, increased numbers of rounded, activated tenocytes are interspersed throughout the pathological tendon.
- While determining changes in cell number and phenotype are beyond the resolution of clinical imaging modalities, the consequence of these active metabolic cells can be detected. A shift in the proteoglycan content, from the small leucine-rich proteoglycans (eg, decorin) to the larger hydrophilic proteoglycans (eg, aggrecan), results in an increase in bound water and tendon thickening.
- These changes have been described on ultrasound as increases in tendon dimensions and heterogeneous or diffuse changes in echogenicity. Previous work in asymptomatic patellar tendons suggests that abnormal tenocyte morphology and changes in proteoglycan content are the primary changes in tendinosis.13 Fibrillar disorganization is another feature of tendinosis in which fibers are present in a haphazard arrangement, somewhat due to the change from type I collagen to type II and III collagen. The parallel arrangement of normal tendon fibers generates a single ultrasound reflection when the ultrasound probe is perpendicular to the long axis of the tendon. Multiple reflections and shadowing are generated by fibrillar disorganization and lack of parallel-aligned fibers, which are represented by an area of hypoechogenicity on ultrasound
- Weinberg et al98 reported that Doppler flow was only observed in tendons that contained an area of hypoechogenicity and not in abnormally thickened tendons with normal echogenicity. This association between the presence of blood vessels and areas of matrix disorganization suggests that the infiltration of blood vessels may be opportunistic. The infiltration of blood vessels and accompanying nerves has previously been implicated as a source of pain, with moderate associations reported between Doppler signal and the presence and location of pain, with Doppler signal also being associated with poorer clinical outcomes. However, increased Doppler signal is present in asymptomatic tendons, suggesting that blood vessels and accompanying nerves are not the primary source of pain.
It is also important to note that the reliability of detecting Doppler signal is poor with exercise affecting the presence/absence of Doppler signal.
Accuracy, Sensitivity, and validity of Imaging.
There has been considerable debate on the clinical utility of imaging in tendinopathy. There have been numerous studies that have investigated the accuracy (number of correct imaging diagnoses, both abnormal and normal imaging, divided by the total number of cases) and sensitivity (number of correct abnormal imaging diagnoses divided by the total number of symptomatic cases) of a number of imaging modalities in detecting clinical tendinopathy.
A central issue with studies that test the accuracy of imaging is the lack of a valid clinical gold standard. The diagnosis of tendinopathy can be complex, due to a confusing clinical picture with few clinical diagnostic tests for tendinopathy. As yet, there is no consensus as to which clinical exam may serve as the gold standard for the diagnosis of tendinopathy.A central issue with studies that test the accuracy of imaging is the lack of a valid clinical gold standard. The diagnosis of tendinopathy can be complex, due to a confusing clinical picture with few clinical diagnostic tests for tendinopathy. As yet, there is no consensus as to which clinical exam may serve as the gold standard for the diagnosis of tendinopathy.
But caution is required when interpreting the results of studies, as asymptomatic participants are frequently included, and their inclusion may result in an overestimation of the accuracy and sensitivity of imaging in detecting clinical tendinopathy.
Participant selection in these studies is critical, and future studies need to reflect the use of imaging in the clinical setting (ie, imaging to differentiate tendinopathy from other pain conditions in the same anatomical region).
Can Imaging Predict the Onset of Pain or Clinical Outcome?
A feature and criticism of the use of imaging in individuals with tendinopathy is the poor correlation with the presence of pain and pain severity. Abnormal imaging has been reported in various tendons in as many as 59% of asymptomatic individuals. This reflects the complex nature, and our limited understanding, of tendon pain.
Tendon pain is not solely driven by local tissue changes, yet there is likely to be an interaction between the local tissue and the peripheral and central nervous system.
Despite the poor relationship between pathological changes and pain, local tissue changes and the use of imaging in visualizing these changes may be important as a prognostic tool. Fredberg et al followed 54 asymptomatic Danish elite soccer players for development of Achilles tendon pain over 12 months. The authors reported that players with substantial imaging changes (thickening and hypoechoic region greater than 2 mm in the transverse planes) at baseline had a 3-fold increase in the relative risk of developing symptoms (95% confidence interval [CI]: 1.6, 4.9). Similarly, Malliaras and Cook reported that abnormal ultrasound imaging of the patellar tendon increased the relative risk of developing pain 15-fold in volleyball players (95% CI: 1.9, 111.4). Conversely, Giombini et al, who investigated the Achilles, patellar, and quadriceps tendons in 37 asymptomatic elite fencers, reported that hypoechogenicity predicted future symptoms only in the patellar tendon, not the quadriceps or Achilles tendon. Sonoelastography has been utilized to investigate the development of Achilles tendon pain in elite Australian football players. Structural and mechanical property changes of the Achilles tendon, such as intratendinous delaminations, hypoechogenicity, and neovascularization, as well as soft tendon properties, were present in 52.4% of asymptomatic elite Australian football players (unpublished data from Ooi et al).
Monitoring of Tendon Structure: What Can We Expect From the Tendon?
To assess the efficacy of treatments, outcome measures of pain and function (Victorian Institute of Sport AssessmentAchilles [VISA-A] for the Achilles tendon), return to activity, and structure on imaging have been used. However, as mentioned earlier, assessment of tendon structure has been limited to subjective grading or measurements of tendon dimensions. Recent studies that have semi-quantified aspects of tendon structure have provided information on how tendons respond structurally to various treatments. Shalabi et al reported a significant decrease in Achilles tendon volume and intratendinous signal following a 3-month eccentric loading program.
While improvements in tendon structure were observed, tendon volume and intratendinous signal did not return to normal. Similarly, long-term follow-up (mean of 4.2 years) of the same cohort reported no significant difference in tendon volume compared to baseline measures, despite improvements in pain and function.
Furthermore, previous research has found that Achilles tendon structure on UTC is no different after a 16-week eccentric loading program, despite improvements in the VISA-A score. The findings of these studies suggest that improvements in the tendon are not necessary to facilitate clinical improvement after an eccentric exercise program.
A systematic review by Drew et al reported that improvements in pain and function with eccentric exercise were not mediated by changes in tendon structure. In this context, while improvement in, or normalization of, tendon structure is a positive result, it is not necessary for improvement in pain and function, suggesting that the pathological tendon might have adapted to become, and remain, load tolerant.
Pathological Achilles and patellar tendons demonstrate an increased mean cross-sectional area of aligned fibrillar structure compared to structurally normal tendons. Therefore, it appears that the pathological tendon maintains sufficient amounts of aligned fibrillar structure by increasing tendon dimensions. Interventions such as eccentric exercise may not be efficacious in remodeling the area of pathology; rather, these loading protocols may cause adaptation and increase the loading capacity of the surrounding aligned fibrillar structure
In the context of previous research, stability in tendon structure accompanied by improvements in pain and function can be considered a positive outcome.Future research is needed to investigate whether tendons that remain stable, in terms of extent of pathology, have a better clinical outcome than those that worsen over time.
Future Directions for Research
Load is a critical factor in the development of tendinopathy, and numerous investigators are using imaging to improve our understanding of how tendons respond to load. Grigg et al reported an acute response in the Achilles tendon, where anteroposterior diameter was reduced immediately after eccentric exercise and was restored after 24 hours. Ultrasound tissue characterization has been used to detect a short-term temporal response in the Achilles tendon and superficial digital flexor tendon of the horse in which a loss of aligned fibrillar structure was observed 48 hours after maximal load that returned to baseline at 3 to 4 days. This is consistent with other studies that have reported a short-term response in hydration level using off-resonance saturation MRI,90 tendon volume and
Proposed Clinical Role of Imaging in Tendinopathy
In answering the question we posed in the title, imaging is not telling us the entire picture in tendinopathy. The role of imaging in the clinical setting may be somewhat limited, as it does not directly relate to symptoms. Imaging allows for the visualization of structure; it does not represent the entire clinical picture and should not be used as the sole diagnostic criterion in determining whether the clinical presentation is generated by the tendon. While imaging may not tell us the entire picture, just like pain on palpation gives us little information on muscular strength and endurance, it provides important information about structure. While not discussed as part of this review, imaging may be useful in differential diagnosis As has been consistently stated throughout this narrative review, imaging needs to be placed in the context of the overall clinical picture, and, despite the risk of potential bias, this may be enhanced by clinicians performing the imaging themselves (specifically ultrasound).
However, caution is advised, with the potential for imaging in the clinician’s hands to confuse the clinical picture and lead to poor outcomes due to technical insufficiencies.
As stated above, ultrasound is highly user dependent, with specifically trained musculoskeletal radiologists able to produce high-quality images that may provide more clinically relevant information than those produced by clinicians with less experience in imaging.
In testing the efficacy of treatments or monitoring the improvement of clinical symptoms using imaging, a potential shift away from focusing on improving tendon structure may be needed.
As clinical improvements have been shown not to be mediated by structural changes, stability in tendon structure may be a positive outcome in the clinical context of reduced pain and dysfunction. Currently, we are limited by the reliance of conventional imaging modalities on subjective interpretation. The development of new imaging techniques that utilize more quantifiable parameters, such as UTC or sonoelastography, will hopefully enhance our ability to diagnose, predict the development of symptoms, and monitor the efficacy of treatments.