Introduction 2D strain imaging is a potentially important method for investigating strain distributions in the musculoskeletal system [Drakonaki, 2012]. Inducing tensile strain in the Achilles tendon in vivo can be performed by passive dorsiflexion of the foot manipulated by a clinician or mechanical apparatus and can be more repeatable and sensitive to changes than axial strains found by palpation [Brown, 2012]. Lateral strains are prone to measurement errors, as tensile force direction may not be aligned with the image-lateral direction and apparent strains may be affected by tissue boundaries
Principal strain is a simple measure derived from the axial, lateral and shear strain components found from 2D strain imaging. The maximum principal strain images highlight slip boundaries and produce mean strain measurements that are more sensitive to changes in the tendon as it heals from rupture.
Maximum and minimum principal strains are defined in equation 1.
Where εx is the image-lateral strain, εz is the image-axial strain and εxz is the shear strain (equation 2)
The aim of this study was to show that quantitative principal strain measurements of Achilles tendons during passive dorsiflexion can offer clinically relevant measures of tendon health.
Methods Lateral and principal strain images were generated from ultrasound RF sequences using the AutoQual algorithm [Brown, 2013] for a longitudinal study of Achilles tendon rupture. This had 21 patients performing controlled repeatable passive dorsiflexion at 0, 1, 3, 6, 8, 12 and 24 weeks post rupture. Means (εx,εp) and standard deviations (Sx,Sp ) of strain values were taken from the manually segmented tendon regions of interest (ROI) at the rupture site.
Mean image signal to noise ratios (iSNR - equation 3) and Spearman rank correlation coefficient of the time series were calculated for comparison. Mean P-values to test the similarity of patent groups between weeks were also generated using Welch’s t-test.
Results Mean principal strain values were consistently higher than lateral strains but also had higher amounts of variation, leading to lower iSNR values. However, over the longitudinal study of tendons healing from rupture, the principal strains provided a stronger correlation against time and an improved week-by-week distinctiveness of strain values shown by lower P-value (Table 1).
Discussion Principal strain values showed an improvement to Spearman rank correlations and less overlap of patient groups week by week. The higher signal to noise ratio could indicate an increased sensitivity to change within the rupture zone in weeks 1–3, and would be greatly affected by selection of ROI.
The principal strain image (Figure 1) also highlights slip boundaries, which could increase accuracy of ROI selection for all types of strain analysis.
This imaging technique is highly sensitive to the tissue mechanical changes during Achilles tendon healing and shows promise as a monitoring and predictive tool for Achilles tendon health, including tendinopathy.
References Drakonaki et al. Br J Radiol. 2012;85(1019):1435–1445.
Brown et al. ISBI. 2012:1104–1107.
Brown et al. J Biomech. 2013;46(15):2695–2700.