Correlations between elastic modulus and ultrashort echo time (UTE) adiabatic T1ρ relaxation time (UTE-Adiab-T1ρ) in Achilles tendons and entheses

Abstract

Patients with psoriatic arthritis commonly have abnormalities of their entheses, which are the connections between tendons and bone. There are shortcomings with the use of conventional magnetic resonance imaging (MRI) sequences for the evaluation of entheses and tendons, whereas ultrashort echo time (UTE) sequences are superior for the detection of high signals, and can also be used for non-invasive quantitative assessments of these structures. The combination of UTE-MRI with an adiabatic-T_1ρ preparation (UTE-Adiab-T_1ρ) allows for reliable assessment of entheses and tendons with decreased susceptibility to detrimental magic angle effects. This study aimed to investigate the relationship between quantitative UTE-MRI measures and the biomechanical properties of Achilles tendons and entheses. In total, 28 tendon-enthesis sections were harvested from 11 fresh-frozen human cadaveric foot-ankle specimens (52±years old). Tendon-enthesis sections were scanned using the UTE-Adiab-T_1ρ and UTE-T₁ sequences on a clinical 3T scanner. MRI-based measures and indentation tests were performed on the enthesis, transitional, and tensile tendon zones of the specimens. Hayes’ elastic modulus showed significant inverse correlations (Spearman’s) with UTE-Adiab-T1ρ in all zones (R= − 0.46, − 0.54, and − 0.61 in enthesis, transition, and tensile tendon zones, respectively). Oliver-Pharr’s elastic modulus showed significant inverse correlations with UTE-Adiab-T1ρ in transition (R= − 0.52) and tensile tendon zone (R=− 0.60). UTE-T₁ did not show significant correlations with the elastic modulus. UTE-MRI and elastic modulus were significantly lower in the tensile tendon compared with the enthesis regions This study highlights the potential of the UTE-Adiab-T_1ρ technique for the non-invasive evaluation of tendons and entheses.

1. Introduction

Tendon and bone comprise a special interacting unit that is essential for locomotion and individual autonomy (Edwards et al., 2015). The enthesis, which is the region connecting tendon to bone, distributes stress over a wide area or volume. This region varies in composition depending on the location in the body, and fibrous and fibrocartilaginous entheses have been described. The structure of the enthesis also varies, both at the macroscopic and the microscopic levels, and there are significant spatial differences in collagen fiber orientation near the bone when compared with the portions closer to the tensile tendon or ligament.

Entheses are commonly involved in many diseases, including overuse syndromes as well as the early stages of osteoarthritis (OA) and seronegative spondyloarthropathies (Benjamin and Ralphs, 1998, 1997; Tan, 2006). In psoriatic arthritis (PsA), the enthesis is one of the primary targets of the disease (McGonagle, 2003; Watad et al., 2017). Noninvasive imaging is crucial for diagnosis and management, and magnetic resonance imaging (MRI) is one modality that is routinely used for the morphological evaluation of tendons and entheses (Hodgson et al., 2012; Pierre-Jerome et al., 2010; Weinreb et al., 2014). However, tendons and entheses typically demonstrate low signal-to-noise ratios due to the very short transverse relaxation times (T₂), and clinical MRI techniques can only be used for direct evaluation when the structures are markedly abnormal (Chang et al., 2015a; Du et al., 2010b; Juras et al., 2012). In contrast to conventional MRI sequences, ultrashort echo time (UTE) MRI techniques can detect considerable signals from both short- and long-T₂ components, permitting evaluation of tendons and entheses throughout the entire range of disease, from healthy to damaged (Chang et al., 2015a; Ma et al., 2022). UTE-MRI can also be used for quantitative assessment, such as through the measurement of apparent transverse relaxation time (UTE-T₂*) (Agergaard et al., 2021; Chang et al., 2015b, 2014; Chen et al., 2018b, 2018a; Filho et al., 2009; Jerban et al., 2017; Juras et al., 2012; Liu et al., 2019; Malmgaard-Clausen et al., 2021; Qiao et al., 2017; Robson et al., 2004; Xie et al., 2020), longitudinal relaxation time (UTE-T₁) (Chen et al., 2018a; Filho et al., 2009; Jerban et al., 2022b, 2019; Wright et al., 2012), longitudinal relaxation time in a rotating frame (UTE-T_1ρ) (Chen et al., 2018a; Du et al., 2010a), and magnetization transfer (MT) (Chen et al., 2018b, 2018a; Fang et al., 2022; Hodgson et al., 2011; Jerban et al., 2022b, 2019; Zhu et al., 2018).

T_1ρ has been reported to be sensitive to proteoglycan (PG) depletion (Duvvuri et al., 2001, 1997; Li et al., 2008, 2007, 2005; Regatte et al., 2006, 2003). However, the main confounding factor is the magic angle effect (Du et al., 2009; Erickson et al., 1993; Fullerton et al., 1985; Goodwin et al., 1998; Henkelman et al., 1994; Mlynárik et al., 2004; Mosher et al., 2001; Rubenstein et al., 1993; Shao et al., 2017; Xia et al., 1997). T_1ρ values significantly increase when the orientation angle of the collagen fibers in the tissue approaches 55° relative to the B0 field (MRI bore direction). T_1ρ values can vary considerably due to magic angle-related effects, with reports of 76% (Wu et al., 2020), 92% (Shao et al., 2017), 32% (Wang and Xia, 2013), and 11% (Li et al., 2011) variations, depending on the sequence that is utilized and the magnitude of the orientation angle alteration. The higher end of these T_1ρ variations can exceed changes caused by degeneration, which are typically in the range of 10–30% (Mosher et al., 2011), thereby complicating diagnosis and treatment monitoring based on T_1ρ. Notably, UTE-T_1ρ using a typical continuous-wave spin lock-pulse is also sensitive to the magic angle effect (Shao et al., 2017). On the contrary, employing a cluster of adiabatic inversion recovery spin-lock pulses for preparation before a 3D-UTE-Cones data acquisition (UTE-Adiab-T_1ρ) (Ma et al., 2018), can measure T_1ρ with low sensitivity to the magic angle effect (Ma et al., 2018; Wu et al., 2020). UTE-Adiab-T_1ρ has demonstrated significant inverse correlations with the tensile elastic modulus of the human anterior cruciate ligament (ACL) (n=13, R=−0.59, p<0.01)(Jerban et al., 2022a) as well as the indentation-based elastic modulus of tibiofemoral cartilage (n=40, R=− 0.54 for superficial layer and R=−0.37 for total thickness cartilage, p<0.01)(Namiranian et al., 2020).

This study aimed to investigate the correlations of UTE-Adiab-T_1ρ and UTE-T₁ with the biomechanical properties of Achilles tendons and entheses. We hypothesize that UTE-Adiab-T_1ρ will show significant correlations with the elastic modulus of the Achilles tendon and enthesis, at a similar level to previously examined musculoskeletal tissues (ligament and articular cartilage). This study will highlight the UTE-Adiab-T_1ρ technique for the non-invasive characterization of biomechanical properties.

2. Material and Methods

2.1. Sample preparation

Eleven fresh-frozen human cadaveric foot-ankle specimens (52±18 years old at the time of death) were provided by two whole-body donation companies (United Tissue Network, Phoenix, AZ, USA, and Medcure, Portland, OR, USA). Distal-posterior sides of the specimens were cut using a commercial bandsaw (B16, Butcher Boy Machines, Selmer, TN, USA) and a low-speed diamond saw (Isomet 1000, Buehler, IL, USA) into 4- to 6-mm thick sagittal sections including the Achilles tendon, enthesis, and the calcaneus bone. The total number of sections was 28 (2–3 sections per donor). All tendon-enthesis specimens were visually inspected and appeared grossly normal. The tendon-enthesis specimens were soaked in phosphate-buffered saline (PBS) for 2 hours to ensure adequate rehydration after potential drying throughout the sample preparation.

2.2. MRI scanning

MRI scans were performed on a 3T MRI scanner (MR750, GE Healthcare Technologies, WI, USA) using a standard GE eight-channel knee coil at room temperature. All specimens were placed in a plastic container filled with perfluoropolyether (Fomblin, Ausimont, NJ, USA) to minimize dehydration and susceptibility artifacts during MRI scans. The container was positioned carefully inside the knee coil to ensure that the tensile sections of the tendons were perpendicular to B0. All MR images were acquired in the sagittal plane of the tendon-enthesis specimens (i.e., the axial plane in the scanner).

To measure UTE-T₁, an actual flip angle imaging-variable flip angle (AFI-VFA) sequence (AFI: TE=0.032ms, TRs=20,100ms, FA=45°; VFA: TE=0.032ms, TR=20ms, FAs=5, 10, 20, 30°) (Ma et al., 2019) was performed with a total scan time of 35 minutes. To measure UTE-Adiab-T_1ρ values, five fat-saturated 3D-UTE-Cones sequences with adiabatic T_1ρ preparation (Ma et al., 2018) were performed with the following parameters: TR=500 ms; FA=10°, number of spokes=25; spin‐locking time (TSL)=0, 12, 24, 36, and 48ms; number of adiabatic full passage (AFP) pulses=0, 2, 4, 6, and 8. T1 and B1 maps calculated from the AFI-VFA sequence were used as prerequisites to correct the UTE-Adiab-T_1ρ measure. The total approximate scan time for UTE-Adiab-T_1ρ was 20 minutes. In addition to the UTE-MRI sequences, a clinical T1-weighted CUBE sequence (TR= 926 ms, TE=18 ms, and FA=90°) was performed for qualitative morphological evaluation of the specimens. Other imaging parameters for the performed sequences are summarized in Table 1.

Table 1:

Summarized MRI sequence parameters

Sequence	TRs/TEs/TSL (ms); FA (°)	Scan time (min)	Other parameters
AFI-VFA 3D-UTE-Cones T 1	AFI: TE=0.032; TRs=20 and 100; FA=45; VFA: TE=0.032; TR=20; FAs=5, 10, 20, 30;	35	FOV = 110 mm2 Matrix = 340×340 In-plane pixel size = 0.32×0.32 mm2 Band width = 125 kHz Slice thickness = 0.8 mm Number of slices = 100 (cover all specimens in one scan)
Fat-saturated 3D-UTE-Cones Adiab-T 1ρ	TR=500; TSL=0, 12, 24, 36, and 48; FA=10;	20
Clinical T1-weighted CUBE	TR=926; TE=18; FA=90;	5

2.3. Mechanical indentation testing

Specimens were washed with PBS after the MRI scans and were soaked in PBS for 30 minutes before mechanical tests. Skin, muscle, and fatty tissues were gently removed using a scalpel before performing mechanical tests. Indentation testing was performed using a commercial biomechanical testing system (MACH-1, Biomomentum, Quebec, Canada). The specimens were placed in 2-inch square Petri dishes filled with PBS to avoid dehydration during the mechanical testing. A 1.0-mm diameter plane-ended cylinder indenter was positioned perpendicular to the cut surfaces of the specimens. Specimens were tested along 7–10 rows of points depending on the size of the specimens. Each row of the measurement was manually set by the operator, starting from the bone-enthesis side (1 mm apart from the bone edge) and ending at the proximal edge of the tendon. The distance between rows was 2 mm, while the distance between points in each row was 1 mm. The indenter motion within and between rows was performed by horizontal displacement of the stage in the X and Y directions, respectively, using the operating software. The schematics of the tested points are shown in Figure 1a. Points in each row were labeled based on the distance from the bone edge and the length of each row (L). Specifically, points with a distance below L/3 from bone were referred to as the enthesis zone (red points in Figure 1a), points with a distance above L/3 and below 2L/3 from bone were referred to as the transition zone (yellow points in Figure 1a), and points with a distance above 2L/3 from bone were referred to as the tensile tendon zone (blue points in Figure 1a).

Figure 1.

(a) The schematics of the tested points on a representative Achilles tendon-enthesis specimen using indentation testing. Tested points are shown on a UTE-MRI image. (b) Experimental setup for indentation test performed on a representative Achilles tendon-enthesis specimen using a commercial biomechanical testing system (MACH-1; Biomomentum, Quebec, Canada) and a 1.0-mm diameter plane-ended cylinder indenter. Specimens were tested along 7–10 rows of points, each starting from the enthesis side and ending at the edge of the tendon. Points were manually set by an operator and labeled based on the distance from the bone edge with enthesis (distance below L/3 from bone, L is the length of each row, red points), transition (distance above L/3 and below 2L/3 from bone, yellow points), or tensile tendon (distance above 2L/3) zones. (c) Load-displacement curve indicating the loading and unloading processes at a displacement rate of 100μm/s (maximum indentation depth was 300 μm > 1/10 of the average specimen thickness). Overlaying load-displacement curves for two indentation processes showed the high repeatability level of the mechanical measurement. P_max and S refer to the maximum load and unloading slope required for the calculation of Hayes’s of Oliver-Pharr’s elastic modulus.

Figure 1b shows the indentation test setup measuring the mechanical properties of a representative Achilles tendon-enthesis specimen. After finding the contact between the indenter tip and the tissue, the actuator lowered the indenter at a rate of 100 μm/sec. The loading was followed by immediate unloading with an unloading rate of 100 μm/sec. Maximum indentation depth was set at 300 μm (<10% of the minimum thickness of the specimens) in order to avoid significant nonlinearities in the tissue deformation. The average mechanical properties for each point were measured from two consecutive indentation processes, with five seconds of rest time in between. Figure 1c shows the load-displacement curve for a representative indentation test (two consecutive indentations) to measure maximum load (P_max), which is required for the calculation of Hayes elastic modulus (E_Hayes) using Eq. 1 (Hayes et al., 1972). The unloading slope (S) of the load-displacement curve was used for required for the calculation of Oliver-Pharr’s elastic modulus (E _Oliver-Pharr) (Delaine-Smith et al., 2016; Oliver and Pharr, 1992), which is recommended for nonlinear materials and soft tissues. Overlaying the load-displacement curves for the two consecutive indentations illustrated a high repeatability level in the mechanical measurement process.

$$E_{\mathit{\text{Hayes}}}=\frac{(1-v^2)}{2ak}\frac{P_{max}}{\mathit{\text{depth}}}$$ Eq. 1

$$E_{\text{Oliver-Pharr}}=\frac{S}{2a}$$ Eq. 2

where ν, a, Pmax, S, depth, and k are Poisson’s ratio, the radius of the indenter, maximum load, unloading slope, indentation depth, and correction factor (k=1.25) introduced by Hayes et al. (Hayes et al., 1972), respectively. The correction factor selection was based on the provided chart in Hayes et al. (Hayes et al., 1972) study, which utilizes the indenter diameter, indentation depth, and Poisson’s ratio to guide the selection. A Poisson’s ratio of 0.45 was used in the calculations, which was expected to be appropriate as the loading rate was assumed to be relatively fast, and only a limited outward water flux was expected to occur in the meantime (Delaine-Smith et al., 2016).

2.4. MRI data analysis

Sections with MRI artifacts due to metal particles and air bubbles present around the specimens were excluded. In total, three enthesis zone sections, one transition zone section, and one tensile tendon zone section were excluded. Finally, 25, 27, and 27 datasets were used for enthesis, transition, and tensile tendon zone evaluations, respectively.

UTE-T₁ and UTE-Adiab-T_1ρ were calculated in a representative midsagittal slice for each specimen (Figure 1a) for 3×3-pixel regions of interest (ROIs). The center of each square ROI was selected visually by an experienced analyst to be matched with the indentation points, which are schematically illustrated in Figure 1a. Recorded pictures of the indentation points were used to ensure a rigorous ROI selection process. Single-component exponential fitting models (Eq.3 and Eq.4) were used to measure UTE-T₁ and UTE-Adiab-T_1ρ relaxation times.

$$\text{I}\left(\text{FA}\right)=M_{0}sin\left(B_1.FA\right)\frac{1-\mathit{\text{exp}}\left(-\text{TR}/T_1\right)}{1-\mathit{\text{exp}}\left(-\text{TR}/T_1\right)\text{cos}\left(B_1.FA\right)}+\mathit{\text{constant }}\left[9\right]$$ Eq. 3

$$\text{I}\left(\text{TSL}\right)=M_0\text{sin}\left(B_1.FA\right)\frac{\text{exp}\left(-\frac{TSL}{T_{1\rho }}\right).\left(1-\text{exp}(-\frac{TR-TSL}{T_1})\right)}{1-\text{exp}\left(-\frac{TSL}{T_{1\rho }}\right).\text{exp}\left(-\frac{TR-TSL}{T_1}\right).\text{cos}\left(B_1.FA\right)}+\mathit{\text{constant}}$$ Eq. 4

Where I, M0, B1, FA, and TR refer to the mean signal intensity, equilibrium magnetization, RF field magnitude, and nominal flip angle.

2.5. Statistical analysis

Since a perfect co-localization between indentation points and MRI ROIs was not possible, for each studied tendon-enthesis specimen (n=28 from 11 donors), arithmetic means of the MRI and indentation results within the enthesis, transition, and tensile tendon zones were used for statistical analysis. The one-sample Kolmogorov-Smirnov test determined that the MRI and indentation results were not normally distributed. Therefore, Spearman’s rank correlations were calculated between the mean UTE MRI quantifications (UTE-T1 and UTE-Adiab-T1ρ) and measured mechanical properties (peak indentation load and Hayes’s and Oliver-Pharr’s elastic modulus). To ensure that intra-specimen dependency does not affect the results. All correlation studies have been repeated using one average measure per donor. As an exploratory study, the Kruskal-Wallis test by ranks was used to examine the differences between the MRI and mechanical measures in enthesis, transition, and tensile tendon zones. P-values < 0.05 were considered significant. The Holm-Bonferroni method was used to correct the significance level for multiple comparisons between groups. All UTE MRI measurements and statistical analyses were performed using MATLAB (version 2021, The Mathworks Inc., Natick, MA, USA).

3. Results

Figure 2 shows a representative Achilles tendon-enthesis specimen imaged at 3T using clinical T₁-weighted CUBE and 3D-UTE-Cones (Figure 2b) sequences. The tendon and enthesis demonstrated no detectable signals on the clinical MR image (Figure 2a), while they showed high signals when imaged with 3D-UTE-Cones (Figure 2b). The abundant signal enabled UTE-T₁ and UTE-Adiab-T_1ρ quantifications in the targeted tissues. Pixel maps of UTE-T₁ and UTE-Adiab-T_1ρ generated over the tendon and enthesis regions are shown in Figures 2c and 2d, respectively.

Figure 2:

(a) A representative Achilles tendon-enthesis specimen imaged at 3T using a clinical T₁-weighted CUBE sequence (TR= 926 ms, TE=18 ms) where the tendon (blue arrow) and enthesis (red arrow) demonstrated no detectable signals. (b) UTE MRI (TR= 500 ms, TE=0.032ms) image of the same specimen with high signal acquired for both the tendon (blue arrow) and enthesis (red arrow). (c) Overlayed UTE-T₁ and UTE-Adiab-T_1ρ pixel maps within the tensile tendon and enthesis regions.

The average, standard deviation, and range of the UTE MRI quantifications and the indentation test results averaged within enthesis, transition, tensile tendon, and all zones are presented in Table 2.

Table 2:

The average, standard deviation (SD), and ranges of UTE-MRI measures and indentation test results (n=28).

	UTE-T1 (ms)	UTE-Adiab-T1ρ (ms)	Peak load (N)	Hayes Elastic modulus (MPa)	Oliver-Pharr Elastic modulus (MPa)
Enthesis	829±75 [696–953]	30.0±3.9 [22.8–39.4]	1.27±0.57 [0.31–2.59]	2.71.0±1.21 [0.66–5.51]	6.10.0±2.92 [1.21–11.07]
Transition	797±77 [658–932]	28.0±3.3 [21.9–35.1]	0.92±0.44 [0.26–1.78]	1.95±0.94 [0.55–3.77]	3.71±1.93 [0.79–7.79]
Tensile tendon	735±79 [588–863]	26.0±3.7 [19.8–32.7]	0.72±0.40 [0.18–1.76]	1.54±0.85 [0.38–3.74]	2.92±1.70 [0.63–6.91]
All	788±71 [667–918]	28.0±3.2 [22.4–34.5]	1.01±0.45 [0.26–2.09]	2.15±0.95 [0.55–4.43]	4.41±2.02 [0.95–8.77]

Table 3 presents Spearman’s correlation coefficients between elastic modulus (Hayes’s and Oliver-Pharr’s) and UTE-MRI measures. Correlation coefficients are calculated for averaged variables per section (n=28) and per donor (n=11). Hayes’ elastic modulus showed significant inverse correlations with UTE-Adiab-T1ρ in all zones (R= − 0.46, − 0.54, and − 0.61 in enthesis, transition, and tensile tendon zones, respectively). Oliver-Pharr’s elastic modulus showed a significant inverse correlation with UTE-Adiab-T1ρ only in transition (R= − 0.52) and tensile tendon zone (R=− 0.60). When averaged per donor, the UTE-Adiab-T1ρ correlations with elastic modulus were found to be significant only in the tensile tendon zone (R= − 0.79 and − 0.74 for Hayes’s and Oliver-Pharr’s elastic modulus, respectively). Correlations between UTE-T1 and elastic modulus were insignificant.

Table 3:

The Spearman’s correlation coefficients between elastic modulus and UTE-MRI measures. Correlation coefficients are calculated for averaged variables per section and per donor.

		Averaged data per section (n=25–27*)		Averaged data per donor (n=11)
		UTE-T1	UTE-Adiab-T1ρ	UTE-T1	UTE-Adiab-T1ρ
Enthesis	Hayes Elastic modulus	−0.36 (p=0.07)	−0.46 (p=0.02)	−0.27 (p=0.42)	−0.52 (p=0.10)
	Oliver-Pharr Elastic modulus	−0.30 (p=0.15)	−0.39 (p=0.04)	−0.19 (p=0.57)	−0.50 (p=0.12)
Transition	Hayes Elastic modulus	−0.29 (p=0.14)	−0.54 (p<0.01)	−0.37 (p=0.26)	−0.60 (p=0.04)
	Oliver-Pharr Elastic modulus	−0.30 (p=0.13)	−0.52 (p=0.01)	−0.35 (p=0.30)	−0.62 (p=0.04)
Tensile	Hayes Elastic modulus	−0.34 (p=0.08)	−0.61 (p<0.01)	−0.38 (p=0.25)	−0.79 (p<0.01)
	Oliver-Pharr Elastic modulus	−0.36 (p=0.06)	−0.60 (p<0.01)	−0.38 (p=0.25)	−0.74 (p=0.01)
All	Hayes Elastic modulus	−0.28 (p=0.15)	−0.49 (p=0.01)	−0.52 (p=0.10)	−0.66 (p=0.03)
	Oliver-Pharr Elastic modulus	−0.20 (p=0.32)	−0.39 (p=0.04)	−0.36 (p=0.27)	−0.48 (p=0.13)

^*n refers to the number of datapoints which was 25, 27, and 27 for enthesis, transition, and tensile tendon zones, respectively.

Figures 3 and 4 show the scatter plots and linear trendlines of Hayes’s and Oliver-Pharr’s elastic modulus versus UTE-Adiab-T_1ρ for averaged results per section and per donor, respectively.

Figure 3:

Scatterplots and linear trendlines of Hayes’s and Oliver-Pharr’s elastic modulus versus UTE-Adiab-T_1ρ values for averaged variables per section. R referred to Spearman’s correlation coefficient.

Figure 4:

Scatterplots and linear trendlines of Hayes’s and Oliver-Pharr’s elastic modulus versus UTE-Adiab-T_1ρ values for averaged variables per donor. R referred to Spearman’s correlation coefficient.

The mean UTE-MRI and indentation differences between the enthesis, transition, and tensile tendon zones are presented in Table 4. The mean UTE-T₁, UTE-Adiab-T_1ρ, and Hayes’s and Oliver-Pharr’s elastic modulus were significantly lower in the tensile tendon zone compared with the enthesis (p<0.01 using the Kruskal-Wallis test). Mean measures in the transition zone were lower than the enthesis zone but higher than the tensile tendon zone. However, the differences were not statistically significant, except for the UTE-T1 difference between the tensile and transition zones.

Table 4:

Mean UTE-MRI and indentation differences between the enthesis, transition, and tensile tendon zones. Statistical significance was examined with the Kruskal-Wallis test.

	Difference
	Transition/Enthesis	Tensile/Enthesis	Tensile/Transition
UTE-T1 (ms)	−32 (p=0.36)	− 94 (p<0.01)	− 62 (p=0.03)
UTE-Adiab-T1ρ (ms)	−1.9 (p=0.24)	−3.9 (p<0.01)	−2.0 (p=0.17)
Peak load (N)	−0.4 (p=0.08)	−0.5 (p<0.01)	−0.2 (p=0.25)
Hayes Elastic modulus (MPa)	−0.8 (p=0.08)	−1.2 (p<0.01)	−0.4 (p=0.25)
Oliver-Pharr Elastic modulus (MPa)	−2.4 (p=0.01)	−3.2 (p<0.01)	−0.8 (p=0.32)

4. Discussion

This study investigated the correlation between quantitative UTE-MRI measures (UTE-Adiab-T_1ρ and UTE-T₁) with indentation-based biomechanical properties of Achilles tendons and entheses. Significant correlations were found between UTE-Adiab-T_1ρ values and elastic moduli in the enthesis, transition, and tensile tendon zones. These findings, in addition to the reported low sensitivity to the magic angle effect (Ma et al., 2018; Wu et al., 2020), highlight the UTE-Adiab-T_1ρ technique as a useful tool for enthesis and tendon assessment.

Prior studies employing conventional, continuous-wave T_1ρ techniques have shown inverse correlations between T_1ρ values and macromolecular contents, particularly PG in cartilage (Jobke et al., 2013; Nishioka et al., 2012). The significant negative correlations between UTE-Adiab-T_1ρ and the mechanical properties of the entheses and tendons observed in this study are consistent with the prior results, as PG content is known to maintain tissue swelling pressure as a result of the negative charges on the molecule, which attract sodium and water(Maroudas et al., 1969). It should be noted that the relationship between UTE-Adiab-T_1ρ and PG/collagen content in entheses and tendons is yet to be investigated by comparing with the histological and biochemical ground truth measures. In addition, the strength of the correlations between UTE-Adiab-T_1ρ values and tissue elastic moduli (R=0.46–0.61) was in the range of the previously reported UTE-Adiab-T_1ρ correlations with the elastic moduli of ACL samples (R=0.60) (Jerban et al., 2022a) and tibiofemoral cartilage (R=0.33–0.54) (Namiranian et al., 2020).

We found higher UTE-T1, UTE-Adiab-T1ρ, and elastic modulus values in the enthesis zones compared with the tensile tendon zones. Higher UTE-T1 values in the enthesis zones also agreed with the results from previous investigations (Chen et al., 2018a). Notably, UTE-Adiab-T1ρ values for Achilles tendons and entheses have not yet been reported in the literature. Chen et al. (Chen et al., 2018a) used quantitative UTE measures and found higher UTE-T₂* yet lower magnetization transfer (MT) ratio values in the enthesis compared with the tensile tendon zones (Chen et al., 2018a). The range of the elastic modulus values obtained in this study differed from the previous study (Chen et al., 2018a), likely due to variations in indentation devices and mechanical testing parameters, including indenter shape, indenter diameter, indentation depth, and loading rate.

The limitations of this study can be summarized in six aspects. First, the number of samples was relatively small (28 sections from 11 donors), as is the nature of pilot ex vivo studies using rigorous reference standards. Moreover, the tendon-enthesis specimens were from eleven donors without any visible abnormality such as degeneration or tearing. The inclusion of more tendon-enthesis specimens with various degrees of degeneration would increase the generalizability of this study. Second, the elastic modulus measurement on the studied specimens was performed using indentation tests by lateral compression forces. Though this does not mimic the in vivo forces over the main length of the tendons, which is typically exposed to tensile loading, enthesis units do course around the calcaneus during dorsiflexion and plantarflexion of the ankle resulting in compression forces (Benjamin et al., 2006; Benjamin and Ralphs, 1998). Moreover, entheses and nearby tendon regions may include fibrocartilage differentiation from tendon cells involving changes in cell shape and expression of cartilage markers, especially type II collagen, and aggrecan (Benjamin et al., 2006; Benjamin and Ralphs, 1998). Third, the limited spaces between indentation points (i.e., 1–2mm) might result in high dependency between the original data points. However, since arithmetic means of the MRI and indentation results within the larger zones (enthesis, transition, and tensile tendon) were used for statistical analysis, the net effect would likely have minimal influence on the results. Fourth, the tendon and its enthesis are viscoelastic and nonlinear materials that may not be well examined using elastic modulus alone. The viscoelastic properties of the tissue might lead to insufficient recovery of the tested points before indenting the neighboring points. Also, using a lower Poisson’s ratio (<0.45) may be appropriate in future experiments because the assumption of a fast indentation with limited minor outward water flux during the loading in this study can be challenged. Although, the ratio between the tissue thickness and the indentation depth was higher than 10, assuming a zero-radial displacement at the interface of the tissue and Petri dish is challenging, which might result in Pmax underestimation. Future investigations should consider performing mechanical stress-relaxation and creep experiments on well-secured specimens to be compared with the quantitative MRI biomarkers. Fifth, specimens were soaked in PBS for 2 hours before scanning to ensure similar rehydration levels, which may affect the quantitative MRI measures and indentation results. However, the impact of the PBS soaking was assumed to be similar in all specimens and did not influence the comparisons presented in this study. Sixth, the long scan time of the proposed technique may limit its translation to clinical studies. Employing acceleration techniques (53), using less TSL points for T1ρ fitting, and utilizing reasonable constant values for T1 (excluding the T1 acquisition with ~20 minutes of scan time) could reduce the scan time to near 10 minutes for future in vivo investigations.

5. Conclusion

Correlations between UTE-Adiab-T_1ρ and indentation-based elastic modulus of Achilles tendons and entheses were investigated and were found to be significant in the enthesis, transition, and tensile tendon zones. This study highlights the potential of the UTE-Adiab-T_1ρ technique, which is known to be less sensitive to magic angle effects, for the non-invasive characterization of biomechanical properties in tendons and entheses.