0
Research Papers

# A Study of the Anisotropy and Tension/Compression Behavior of Human Cervical Tissue

[+] Author and Article Information
Kristin M. Myers

Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Room 3-347A, Cambridge, MA 02139kmyers@alum.mit.edu

Simona Socrate

Harvard-MIT Division of Health Sciences & Technology, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Room E25-406, Cambridge, MA 02139ssocrate@mit.edu

Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Room 3-347A, Cambridge, MA 02139apaskal@alum.mit.edu

Michael House

Maternal/Fetal Medicine, Tufts Medical Center, 800 Washington Street, Boston, MA 02111mhouse@tuftsmedicalcenter.org

Overnight equilibration is necessary to avoid confounding effects arising from the hydration of the specimen following excision. Refer to (9) for an extensive discussion.

The tolerance of the load cell is 0.001 N.

Maintaining tissue hydration is important for materials, such as cartilage and cervical tissue, where the diffusion of interstitial fluid through the extracellular matrix plays a relevant role in the tissue response to deformation. Care was taken to mimic physiologic conditions by matching the in-vivo pH and saline concentration. Based on assays of the bathing solution, there was no detectable loss of macromolecular tissue components to the medium.

This period of re-equilibration is generally sufficient to restore the specimens to a virgin state. Evidence is provided in our previous study (9), where specimen responses were found to be repeatable after re-equilibration.

Differences between these sets of parameters provide an indirect measure of the degree of anisotropy of the local collagen alignment.

Note that, for ease of presentation, we refer to all dissipative/time-dependent deformation mechanisms as viscous. Cervical tissue is highly hydrated, and diffusion of the interstitial fluid driven by pressure gradients plays an important role in the time-dependent nature of the tissue response, which is in effect poro-visco-elastic (18,22). However, in a 1D modeling framework, a distinction between the effects of these mechanisms would be entirely arbitrary. Adding an additional dissipative element in the rheological model, to address the effects of interstitial fluid diffusion, would introduce additional model parameters and compromise the uniqueness of the fit of the model to the experimental data.

We found that an explicit integration scheme was more cost effective than an implicit scheme, since fitting values of the power-law exponent, $n$, of the viscous constitutive relation 4 were typically low $(n<3)$, thus allowing sufficiently large time steps within the bounds of stability limits.

Differences between tensile and compressive responses are often referred to as “tension-compression nonlinearity” in the articular cartilage literature. We find this terminology misleading, and believe it is important to illustrate that the behavior of the material in uniaxial loading is continuous, with a single modulus at small strains.

Model fits to the higher strain levels were not attempted with this simplistic model.

For pregnant tissue, the substantial degree of softening with successive loading cycles indicate a tendency of the collagen network to creep, and the equilibrium right branch of the model was not needed.

J Biomech Eng 132(2), 021003 (Jan 05, 2010) (15 pages) doi:10.1115/1.3197847 History: Received November 28, 2007; Revised May 29, 2009; Published January 05, 2010; Online January 05, 2010

## Abstract

The cervix plays a crucial role in maintaining a healthy pregnancy, acting as a mechanical barrier to hold the fetus in utero during gestation. Altered mechanical properties of the cervical tissue are suspected to play a critical role in spontaneous preterm birth. Both MRI and X-ray data in the literature indicate that cervical stroma contains regions of preferentially aligned collagen fibers along anatomical directions (circumferential/longitudinal/radial). In this study, a mechanical testing protocol is developed to investigate the large-strain response of cervical tissue in uniaxial tension and compression along its three orthogonal anatomical directions. The stress response of the tissue along the different orthogonal directions is captured using a minimal set of model parameters generated by fitting a one-dimensional time-dependent rheological model to the experimental data. Using model parameters, mechanical responses can be compared between samples from patients with different obstetric backgrounds, between samples from different anatomical sites, and between the different loading directions for a single specimen. The results presented in this study suggest that cervical tissue is mechanically anisotropic with a uniaxial response dependent on the direction of loading, the anatomical site of the specimen, and the obstetric history of the patient. We hypothesize that the directionality of the tissue mechanical response is primarily due to collagen orientation in the cervical stroma, and provides an interpretation of our mechanical findings consistent with the literature data on preferential collagen alignment.

<>

## Figures

Figure 1

(a) The anatomical location of the cervix in relation to the fetus and female pelvic organs. (b) An idealized schematic of the stroma, fascia, and mucosa layers of the cervix.

Figure 2

A schematic representation of the cervical stroma (the fascia and mucosa are not represented) and the three seamless zones of preferentially aligned collagen fibers. X-ray and MRI data on human cervical tissue indicate that the middle layer of stroma contains collagen fibers preferentially aligned in the circumferential direction and the outer and inner layers contain collagen fibers preferentially aligned in the longitudinal direction (5-6). (Adapted from Ref. 5.)

Figure 3

(a) Specimen preparation of compression cubes and tension strips. (b) Anatomical directions of the cervix. Multiple samples are cut at different radial distance from the inner canal to capture different fiber orientations. Cervical samples are only excised from the stroma region. (c) Directions of “preferential” collagen fiber orientation (5-6). Circles indicate fibers in the circumferential direction (out of plane).

Figure 4

Figure 5

(a) Compression specimen in loading fixture. (b) Video extensometer view of the compression test. The right specimen image is the right-angle prism view. The arrows indicate the directions transverse to the loading axis. (c) Tension specimen in the loading fixture. (d) Video extensometer view of the tension specimen.

Figure 6

The one-dimensional rheological mechanical model for the axial stress-strain response of cervical tissue in uniaxial loading. (a) Schematic arrangement of model elements: material parameters for each modeling component are circled. (b) Explicit time integration scheme.

Figure 7

True stress and lateral stretch response for two specimens from the same nonpregnant cervical slice (NPND Patient 3) for the load-unload tests. The specimens were excised from different radial locations on the slice, and each specimen was tested along two different directions of uniaxial compression (longitudinal and circumferential). Sample 1, left, was excised from the middle stroma region, and Sample 2, right, was excised from the outer stroma region. For each sample, a difference in peak stress and stretch is recorded between the responses in the two loading directions. Further, by comparing Sample 1 and Sample 2, differences in the stress and stretch responses can be ascribed to differences in the direction of preferential collagen alignment. (The lateral stretch for the radial direction represents the average from the longitudinal and circumferential tests.)

Figure 8

True stress and lateral stretch response for the ramp-relaxation tests for the same two specimens presented in Fig. 7 (NPND Patient 3). Each specimen was excised from different radial locations on the slice, and each specimen was tested along two different directions of uniaxial compression. The peak stresses and stretches are different when comparing the different loading directions for a single sample. However, the equilibrium stresses are the same when comparing loading directions. (The lateral stretch for the radial direction represents the average from the longitudinal and circumferential tests.)

Figure 9

True stress and lateral stretch response for two specimens from the same cervical slice (NPPD Patient 2) for the load-unload tests. Each specimen was excised from different radial locations on the slice, and each was tested along two different directions of uniaxial compression. The same anisotropic trends noted for the NPND specimens in Fig. 7 are also present for these two specimens. (The lateral stretch for the radial direction represents the average from the longitudinal and circumferential tests.)

Figure 10

True stress and lateral stretch response for the ramp-relaxation tests for the same two specimens presented in Fig. 9 (NPPD Patient 2). Each specimen was excised from different radial locations on the slice, and each was tested along different directions of uniaxial compression. The peak stresses and stretches are different when comparing the different loading directions for a single sample. However, the equilibrium stresses are the same when comparing loading directions. (The lateral stretch for the radial direction represents the average from the longitudinal and circumferential tests.)

Figure 11

True stress and lateral stretch response for two specimens from the same pregnant cervical slice (PCS Patient 2) for the load-unload tests. Each specimen was excised from different radial locations on the slice, and each specimen was tested along two different directions of uniaxial compression. It was evident that the pregnant tissue underwent some damage during each testing step and was unable to completely recover its virgin conditions by re-equilibration between tests. For example, Sample 1 was tested twice in the circumferential direction, and the amplitude of the stress response for the second circumferential test was smaller than the amplitude of the stress response for the initial circumferential test. Therefore, sequential testing along different directions is not a reliable approach to determine anisotropy effects in pregnant tissue.

Figure 12

True stress and lateral stretch response for the ramp-relaxation tests for the same two specimens presented in Fig. 1 (PCS Patient 2). Each specimen was excised from different radial locations on the slice, and each was tested along different directions of uniaxial compression. (The lateral stretch for the radial direction represents the average from the longitudinal and circumferential tests).

Figure 13

The true stress and lateral stretch response for a single tension strip from NPND Patient 1. (a) True stress response for the load-unload tests. Note that three load-unload cycles are followed by a stress-relaxation test at each strain level. Therefore, when plotting the load-unload stress-strain cycles on the same axes the softening effects between cycles are magnified. (b) True stress response for the ramp-relaxation tests. (c) Lateral Stretch response for the load-unload tests. (d) Lateral stretch response for the ramp-relaxation tests.

Figure 14

The true stress response for a single tension strip from NPPD Patient 3. The specimen slipped from the grips during the 35% true strain load-unload test therefore the last ramp-relaxation test was omitted. (a) True stress response for the load-unload tests. (b) True stress response for the ramp-relaxation tests. (c) Lateral Stretch response for the load-unload tests. (d) Lateral stretch response for the ramp-relaxation tests.

Figure 15

The true stress response for a single tension strip from PCS Patient 1. (a) True stress response for the load-unload tests. There was an error in data acquisition for the intermediate load-unload test conducted to 55% true strain, therefore, these data have been omitted from the graph. Furthermore, the inset of the load-unload graph illustrates the small-strain regime response of the pregnant tissue. (b) True stress response for the ramp-relaxation tests. The specimen slipped from the grips in the stress-relaxation test conducted to a true strain of 75%, therefore, these data have been omitted from the graph. (c) Lateral Stretch response for the load-unload tests. (d) Lateral stretch response for the ramp-relaxation tests.

Figure 16

Tension and compression data along the circumferential direction for two specimens taken from a single cervical slice from NPND Patient 1. In the large-strain regime, the tension-compression stress response is asymmetric. However, in the small-strain regime, the uniaxial loading stiffness matches well (see inset).

Figure 17

Model fit for Sample 1 in Fig. 9 (NPPD Patient 2). The model captures stress behavior well along both directions of axial loading. One set of best-fit material parameters for each loading direction is able to capture both the load-unload and ramp-relaxation stress behavior. Sample 1 was excised from the middle stroma region and believed to have collagen fibers preferentially aligned in the circumferential direction. Differences between best-fit material parameters indicate that the tissue is anisotropic.

Figure 18

Model fit for a nonpregnant tension specimen (NPPD Patient 2). The model captures the stress behavior well in tension. One set of best-fit material parameters is able to capture both the load-unload and ramp-relaxation stress behavior.

## Discussions

Some tools below are only available to our subscribers or users with an online account.

### Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related Proceedings Articles
Related eBook Content
Topic Collections