Shear Wave Elastography (SWE) for Monitoring of Treatment of Tendinopathies: A Double-blinded, Longitudinal Clinical Study

Abstract

We aimed to investigate the diagnostic accuracy with which shear wave elastography (SWE) can be used to monitor response to treatment of tendinopathies, and to compare it to conventional ultrasound (US)-imaging methods (B-mode US (B-US) and power Doppler US (PD-US)). A prospective Institutional Review Board-approved longitudinal study on 35 patients with 47 symptomatic tendons (17 Achilles-, 15 patellar-, and 15 humeral-epicondylar) who underwent standardized multimodal US and standardized clinical assessment before and after 6 months of treatment (tailored stretching exercise, sport break, and local Polidocanol) was carried out. All US studies were performed by radiologists blinded to the clinical symptoms on both tendon sides to avoid biased interpretations, by B-US, PD-US, and SWE, conducted in the same order, using a high-resolution linear 15 MHz probe (Aixplorer). Orthopedic surgeons who were in turn blinded to US imaging results used established orthopedic scores (Victorian Institute of Sports Assessment questionnaire for Achilles, Victorian Institute of Sports Assessment questionnaire for patellar tendons, and Disability Arm Shoulder Hand scoring system) to rate presence, degree, and possible resolution of symptoms. We analyzed the diagnostic accuracy with which the different US imaging methods were able to detect symptomatic tendons at baseline as well as treatment effects, with orthopedic scores serving as reference standard. B-US, PD-US, and SWE detected symptomatic tendons with a sensitivity of 66% (31 of 47), 72% (34 of 47), and 87.5% (41 of 47), respectively. Positive predictive value was 0.67 for B-US, 0.87 for PD-US, and 1 for SWE. After treatment, clinical scores improved in 68% (32 of 47) of tendons. Treatment effects were observable by B-US, PD-US, and SWE with a sensitivity of 3.1% (1 of 32), 28.1% (9 of 32), and 81.3% (26 of 32), respectively. B-US was false-positive in 68.8% (20 of 32), PD-US in 46.9% (15 of 32), and SWE in 12.5% (4 of 32) (SWE). Clinical scores and B-US, PD-US, and SWE findings correlated poorly (r = 0.24), moderately (r = 0.59), and strongly (r = 0.80). Unlike B-US or PD-US, SWE is able to depict processes associated with tendon healing and may be a useful tool to monitor treatment effects.