To determine if mono-exponential T2, mono-exponential UTE-T2*, or bi-exponential UTE-T2* correlate with biomechanical properties in human Achilles tendons.

Samples: Five human cadaveric ankles (donor age 81 ± 16.5 years, mean ± standard deviation) were used for this study.

Protocol: Whole, intact Achilles tendons were scanned on a 3T clinical scanner (Discovery MR 750; GE Healthcare, Milwaukee, WI) using an 8-channel transmit-receive extremity coil with tendon orientation parallel to B₀. Our quantitative MR protocol consisted of a 2D-CPMG sequence (TR = 2000 ms, TE = 10, 20, 30, 40, 50, 60, 70, 80 ms, FOV 10 cm, matrix = 320 x 256, slice thickness = 1 mm, scan time ~8.5 min per tendon, Fig 1A) and 5 sets of an isotropic, dual-echo 3D-UTE-Cones sequence covering 10 echoes (TR = 20 ms, TE = 0.03, 0.1, 0.2, 0.4, 0.6, 0.8, 2, 4, 10, 15, 20, 30 ms, FOV 10 cm, matrix = 168 x 168, slice thickness 0.6 mm, scan time ~5 min per set/~25 min per tendon, Fig 1B). Thereafter 15 cm of Achilles tendons were dissected with the calcaneal attachment maintained. The tendons were bluntly separated along physiologic planes when possible, yielding anterior (soleus) and posterior (gastrocnemius) halves (Fig 1C). Dissected tendons were then imaged using an isotropic 3D-FSE sequence (0.4 mm3) from which cross-sectional area measurements were made for biomechanical analyses.

Image Analysis: Using mid-sagittal images, a fellowship trained musculoskeletal radiologist manually drew regions of interests over the entire tendon (global analysis) as well as the proximal, mid, and distal thirds of each tendon (regional analysis). Mono-exponential T2, mono-exponential UTE-T2*, and bi-exponential UTE-T2* values were calculated using a semi-automated MATLAB (The Mathworks Inc., Natick, MA) routine as previously described (Fig 1D-G)¹.

Biomechanics: The calcanei were potted into a low melting point alloy (Cerrobend, Satterlee Co., Minneapolis, MN), allowing for bone fixation at an appropriate testing angle without heat damage during fixation. Tendons were marked at the surface using India ink and the proximal free end was freeze-clamped using liquid nitrogen cooled grips (Fig 2A) inside a material testing device (Instron model 1122; Norwood, MA). Tendons were preconditioned to reduce the effects of hysteresis before being tensilde loaded to failure. Multi-angle high definition recorded video was used for strain measurements and determination of the precise location of failure (Figs 2B-D). Ultimate tensile strain (%), ultimate tensile stress (MPa), elastic modulus (MPa) and energy to failure (J) were determined².

Statistics: Descriptive statistics were calculated. Spearman correlation was performed between multi-parametric values and biomechanical measures using both global analysis values and regional analysis values at locations of failure.

Eight separate halves were imaged and tested. For two Achilles tendons no physiologic plane was discernable between the soleus and gastrocnemius tendons at dissection and the Achilles was tested as a whole. Mean T2 value was 12.3 ± 4.8 ms and mean mono-exponential UTE-T2* value was 3.5 ± 2.0 ms. Mean short T2* was 2.2 ± 1.0 ms and mean short fraction was 85%. Mean tendon failure was 5.8 ± 3.1 cm above the center of the calcaneal insertion. Mean ultimate tensile strain was 14 ± 4%, mean ultimate tensile stress was 15.6 ± 6.6 MPa, mean elastic modulus was 113.4 ± 40 MPa, and mean energy to failure was 5473.1 ± 2916.7 J. Spearman correlations were significant for ultimate tensile strain and global mono-exponential UTE-T2* values (rho = 0.90, p = 0.002, Fig 3A), global short/long T2* fractions (rho = -0.97 and 0.97, respectively, p < 0.001, Fig 3B), and regional mono-exponential UTE-T2* values (rho = 0.93, p = 0.001). CPMG T2 demonstrated no significant correlations with biomechanical measures (p > 0.3). Stress, elastic modulus, and energy to failure demonstrated no significant correlations with MR measures (p > 0.2).In this pilot study, very high and significant correlation coefficients were seen between mono-exponential T2* and bi-exponential T2* fractions obtained using the UTE-Cones sequence and ultimate tensile strain. Ultimate tensile strain represents the percentage change in tendon length prior to failure and high strains have been previously associated with tendon degeneration³. Our results suggest that non-invasive MRI of the Achilles tendon may serve as a surrogate measure. Specifically, the larger the UTE-T2* value (positive correlation), the larger the long T2* fraction (positive correlation), or the smaller the short T2* fraction (negative correlation), the more deformation can occur in the tendon prior to failure.