PURPOSE: To assess the capability of a commercial sonoelastography system to detect small tendon lesions by quantitative analysis of elastogram profiles.
MATERIALS AND METHODS: Strips of equine digital flexor tendons were used to model small human tendons. Two tendons were examined. From each tendon, six unmodified tendon strips (controls) and six tendon strips with a central defect of the same tendons were compared. The tendon strips were placed under a physiological tensile strain of 5%. Sonoelastographic visualization of the strain profile was performed. Regions of interest (ROI) were defined left and right of the tendon defects. Average tissue strains in these ROI were compared with tissue strain in controls.
RESULTS: In the first series of experiments, there was a significant (p = 0.011) difference in the strain profile in regions proximal and distal to the tendon lesions compared with the respective tendon areas in the control tendon strips. In a second series of experiments, similar trends were observed, but the differences were not significant (p = 0.824).
CONCLUSION: Even under carefully controlled experimental conditions using computational post-processing of sonoelastograms, tendon lesions could only be partially detected within elastograms from a clinical sonoelastography system. The ability to detect differences in some strain profiles indicates that tensile sonoelastography has the potential to identify small tendon lesions (such as those in the hand), but that substantial improvements with respect to quantitative analysis are required to make such measures diagnostically relevant.