In this study, the damage evolution of liver tissue was quantified at the microstructural level under tensile, compression, and shear loading conditions using an interrupted mechanical testing method. To capture the internal microstructural changes in response to global deformation, the tissue samples were loaded to different strain levels and chemically fixed to permanently preserve the deformed tissue geometry. Tissue microstructural alterations were analyzed to quantify the accumulated damages, with damage-related parameters such as number density, area fraction, mean area, and mean nearest neighbor distance (NND). All three loading states showed a unique pattern of damage evolution, in which the damages were found to increase in number and size, but decrease in NND as strain level increased. To validate the observed damage features as true tissue microstructural damages, more samples were loaded to the above-mentioned strain levels and then unloaded back to their reference state, followed by fixation. The most major damage-relevant features at higher strain levels remained after the release of the external loading, indicating the occurrence of permanent inelastic deformation. This study provides a foundation for future structure-based constitutive material modeling that can capture and predict the stress-state dependent damage evolution in liver tissue.

Introduction

Improvements to automobile safety have historically relied on investigations focused on the effects of force, acceleration, and concentrated stresses on dummies, postmortem human surrogates(PMHS), and animals during various crash conditions [17]. However, the use of dummies is limited so that it cannot accurately represent human tissue deformation and response; further, the use of PMHS and animals are costly and present a potential ethical challenge. More recently, research in this field has shifted toward the development of computational models that simulate impact and assess traumatic injury (TI) [817]. Specifically, much effort has been directed toward the development of finite element (FE) models, which represent a promising approach to robustly simulate the precise deformations during blunt impact without the use of human and animal tissues, dummies, and automobiles [12,1824]. Over the years, there has been much progress in the development of FE models for impact and injury predictions. FE models of various human body regions including the lower extremities, abdomen, pelvis, thorax, neck, and head have been generated as well as full body models such as total human model for safety and global human body models consortium [2432]. However, to this end, a noticeable gap exists in the understanding of injury thresholds for these soft tissues. With increasing requirements for model predictability, computational models that include the true structure–property relationships of tissues and provide injury thresholds would be extremely valuable for FE simulations of organ deformation and injury during accident scenarios and would contribute to the optimization of automobile design and safety measures [33].

The quantitative study of injury to tissues has been the focus of several reports but is difficult to study [32,3436]. Damage to tissue or organs is often documented in qualitative ways using terms such as fractures and lacerations [37], without a clear delineation of what is acceptable or unacceptable damage [38,39]. This contrasts with the field of material science, which rigorously tests the damage and failure mechanisms of inorganic materials such as metals, ceramics, and polymers. For example, microfeatures of damage such as microcracks, macrocracks, and voids (or holes) [4042] have been thoroughly characterized in reports by McVeigh and colleagues, who conduct shear deformation tests on high strength steels to quantify the ductility of these materials [42]. Similarly, Agarwal et al. documented void growth and nucleation in aluminum alloys under various stress states [40]. These studies and many others provide essential information detailing the failure mechanisms and limits of many materials, contributing to the assignment of safety factors crucial in the determination of material performance and durability. In the biological field, damage characterizations may be more complex as ultrastructural damage and physiological effect must be correlated to make accurate injury predictions; however, efforts to describe the progression of damage, void growth, and void nucleation, or the appearance of new voids, represent a necessary step toward establishing injury criterion.

The purpose of the current study was to investigate the damage progression of the liver under tensile, compression, and shear states by utilizing an interrupted mechanical test, an essential methodology in damage mechanics, to study damage evolution in porcine livers. The liver was chosen since it is one of the most vulnerable abdominal organs in automobile crashes [4348], accounting for 38% of all abdominal injuries with a mortality rate that can reach up to 67% if damage to the hepatic veins occurs [34]. This study evaluated microstructural images of chemically fixed tissues at various strain levels and under different modes of force application (tension, compression, and shear) and quantified accumulating damage voids. The data showed that all three modes of deformation lead to void growth and nucleation, and eventual failure with tensile forces having the most pronounced effect. Additionally, the presented data demonstrate the utility of interrupted mechanical testing as a way to quantify microdamage features not only for liver tissue but also for all soft tissues.

Materials and Methods

Sample Preparation and Experiment Design.

Porcine livers from healthy juvenile pigs (∼6 months old) were obtained from a local abattoir immediately following death and were placed in phosphate buffered saline (PBS, Fisher Scientific, Pittsburgh, PA) at 4 °C for transportation to the laboratory. All tests were performed within 12 h of extraction. Samples were cut into various sizes depending on the testing modes (tension, compression, or shear (Supplemental Fig. 2 which is available under the “Supplemental Data” tab for this paper on the ASME Digital Collection). Tension samples (2 samples per liver per strain rate for a total of 18 samples) were cut into dogbone shapes using a stamp with total length of 18 mm, gage length of 10 mm, width of 2 mm, and thickness of 1 mm. The radius of the fillet between the grip area and gage length was 2 mm. These samples were cut parallel to the surface of the liver. Compression samples were extracted using an 8 mm diameter biopsy punch by slowly twisting the punch while applying slight downward pressure, which resulted in a sample that was orthogonal to the surface of the liver. We utilized 2 samples per liver for each compression strain rate for a total of 24 liver compression samples. Once these samples were removed, they were trimmed to a disk shape, with a razor blade, of thickness 5 mm. Shear samples (2 samples per liver per strain rate for a total of 18 samples) were cut, with a stamp, into a square shape of 10 mm × 10 mm with a thickness of 5 mm. These samples were taken parallel to the surface of the liver. In all of these samples, the capsule was removed and testing was done solely on the liver parenchyma. Furthermore, directionality was not considered in our extracted samples, because liver tissue has been shown to exhibit an isotropic mechanical response to mechanical testing [32,49,50]. Control histology samples (n = 2 per liver) were taken in a similar fashion using the shear sample stamp and used as a global control for all tests. Before each test, samples were measured in triplicate using digital calipers (Mitutoyo, Inc., Kawasaki, Japan) and the averaged dimensions were used to ensure accurate calculations of stress and strain.

All mechanical tests were performed on a Mach-1 device (Biomomentum, Laval, QC, Canada), which is a stepper motor-driven micromechanical testing machine equipped with a 100 N load cell that has a resolution of approximately 0.005 N. Displacement resolution is 0.5 μm. The load cell is rigidly attached to the vertically aligned stepper motor stage. One end of the sample was connected to the load cell and the other to the rigid base of the machine. The Mach-1 motion program integrates control and data acquisition (National Instruments PCLe-6320, Austin, TX) and is used to set up the testing program. The sampling rate is user selectable (10, 20, 100, 1000 Hz) and was set to 1000 Hz for all tests. Further, the generated data were not passed through filters. During the tensile, compressive, and shear mechanical testing, tissue samples were immersed in PBS solution at room temperature by a water tank custom-designed for the testing machine. Preliminary failure tests were conducted for all testing modes to determine the desired strain levels for each of interrupted mechanical tests to represent the transition, linear, and subfailure regions, respectively. Additionally, samples were preconditioned at 5% strain for ten cycles to generate the steady-state material response of the soft tissue. The detailed test settings and protocols for tensile, compressive, and shear interrupted mechanical tests are described in Secs. 2.2, 2.3, and 2.4, respectively. The testing setups for each mechanical test are shown in (Supplemental Fig. 1 which is available under the “Supplemental Data” tab for this paper on the ASME Digital Collection) without the water bath.

Fig. 1
Fig. 1
Close modal

Interrupted Tensile Testing.

After examining the nonlinear tensile stress–strain curves of liver tissues, true strain levels of 10%, 20%, and 30% were chosen to represent the transitional region, the linear region, and the subfailure region, respectively. Note that for the interrupted tensile testing at each strain level (10%, 20%, and 30%), the sample size was set as N = 6 for a total of 18 samples in tension. For interrupted tensile testing, dogbone-shaped samples were clamped using custom-made grips with a sinusoidal tooth design that was able to minimize sample slippage [51]. A preload of 0.5 g was used to set the reference status, and the displacement rate was set at 10 mm/min. Each sample was preconditioned at 5% strain for ten cycles, and then pulled to the target strain level; the sample was then held at the target strain level, and the PBS solution was immediately replaced with 10% neutral buffered formalin (Fisher Scientific, Pittsburgh, PA). After 3-h on-grip fixation, the deformed geometry and morphology of the specimen were found to be permanently preserved, and the tissue sample was prepared for histology.

To validate the inelastic nature of major microfeatures observed in the deformed tissue, a verification test was carried out, in which additional livers were obtained and two dogbone-shaped samples were removed for each strain level from each liver, as done previously in the interruption tests. For the verification test, the liver tissue samples were pulled to the target strain and then returned to the reference status. The PBS was immediately replaced with 10% neutral buffered formalin, and the specimen at the reference status (undeformed) was subjected to 3-h fixation before being prepared for histology.

Interrupted Compressive Testing.

The compressive testing was carried out with the Mach-1 by mounting a top plate to the machine's linear positioner and a bottom plate to the base of the machine. A cylinder-shaped tissue sample (Φ = 8 mm, thickness = 5 mm) was mounted between two plates for the unconfined compression testing. Similarly, 0.5 g was chosen as the preload and 10 mm/min was set as displacement rate. The target strain levels were 10%, 20%, 30%, and 40% (the sample size was set as N = 6 for each target strain), of which 40% fell into the subfailure region of the stress–strain curve. After the ten-cycle preconditioning (5% strain), the tissue sample was compressed to the target strain level. Once the target strain was reached, the linear positioner was stopped by the program, and the PBS was changed to 10% neutral buffered formalin for 3-h tissue fixation under deformed status. Similar to the interrupted tensile testing, the compressed shape and morphology of liver tissue were found to be preserved after chemical fixation.

As with the tension tests, a compression verification test was carried out, in which additional livers were obtained and two cylinder-shaped samples were removed for each strain level from each liver, as done previously in the compression interruption tests. For the verification test, the liver tissue samples were compressed to the target strain and then returned to the reference status (undeformed). The tissue sample at the reference status was then chemically fixed by 10% neutral buffered formalin for 3 h and prepared for histology.

Interrupted Shear Testing.

A pair of custom-made shear plates was incorporated onto the Mach-1. One plate was mounted onto the Mach-1 linear positioner, and the other plate was mounted on the base of the Mach-1 machine. The sample was then mounted between the two shear plates; one small drop of cyanoacrylate glue was applied for preventing sample slippage from plates due to the vertical device setting. The shear test was performed by moving the plate mounted to the linear positioner upward while maintaining the position of the sample on the plate mounted to the base. After the ten-cycle preconditioning, the tissue sample was loaded to a target shear angle (0.8 rad, 0.9 rad, and 1.0 rad; N = 6 for each strain angle) and held at that position. The PBS was immediately replaced with 10% neutral buffered formalin and the fixation lasted for 3 h while the tissue was in the deformed status.

Again, a shear verification test was carried out, in which additional livers were obtained and two square-shaped samples were removed for each strain level from each liver, as done previously in the shear interruption tests. For the verification test, the liver tissue samples were compressed to the target strain and then returned to the reference status (undeformed).

Histology and Image Analysis.

For histology, the fixed tissue samples were removed from 10% neutral buffered formalin, dehydrated with graduated concentrations of ethanol, embedded in paraffin, and cut into 7 μm sections that were parallel to the direction of loading. The tissue sections were then stained with hematoxylin and eosin, which showed cell nuclei in black color and extracellular matrix (ECM) in pink color. Control samples were also processed to obtain a reference for each of the measured properties. All images were taken using a Leica Microsystems DB2500 microscope (Leica Microsystems, Wetzlar, Germany) at 10x magnification. ImageAnalyzer v.2.2-0 software (MSU CAVS) was used for microstructural analysis of histological images. The parameters obtained from each image during microfeature analysis included the following: void number density, area fraction, mean area, and mean nearest neighbor distance (NND) (Supplemental Fig. 3 which is available under the “Supplemental Data” tab for this paper on the ASME Digital Collection). Number density represents the total number of voids in the observation area; area fraction represents the ratio of the total void area to the total observation area; mean area represents the average size of the damage voids; and mean NND represents the average distance between neighboring voids.

Stress and Strain Calculations.

Data from our mechanical tests were processed in Microsoft Excel (Microsoft, Redmond, WA) to create true stress–true strain curves. Raw unfiltered data were obtained in load and displacement values for all tests. These values were then used to calculate the engineering stress and strain as shown in Eqs. (1) and (2). Note that the shear angle was calculated based on the travel distance of the shear plate and the sample thickness. The engineering stress and strain values were then converted to the true stress–true strain values as shown in Eqs. (3) and (4)
$σengineering=forceareaundeformed$
(1)
$εengineering=lengthdeforemed−lengthundeforemedlengthundeformed$
(2)
$σtrue=σengineering*(1+εengineering)$
(3)
$εtrue=ln(1+εengineering)$
(4)

Statistical Analysis.

Mean ± standard deviation (STDEV) was used to present the experimental data. Statistical analyses were performed using one-way analysis of variances. Dunnett's test was used for pairwise comparison, as well as comparison versus the control group (SigmaStat 3.0, SPSS Inc., Chicago, IL). Data were considered significantly different at p < 0.05.

Results

Tension, compression, and shear tests all revealed nonlinear stress–strain behavior, as expected for soft tissues (Figs. 1(a)1(c)). The hematoxylin and eosin histology of liver tissue at the stress-free status (undeformed state) is shown in Fig. 1(d), in which the lobules of the liver could be distinguished, and the boundary lines of individual lobules are shown as gap-like lines (white arrow). The control samples, or the native liver state, were used as a baseline measure of the natural voids in the liver tissue primarily found near the sinusoids and blood vessels, thus an increase in the number of voids, area of the voids, or a decrease in NND would be indicative in tissue damage. Histology of the interrupted tensile, compressive, and shear tests revealed that damage microfeatures, as well as their evolutions, varied in different loading states.

Evolution of Damage Microfeatures in TensileInterrupted Testing

Damage Pattern.

When the liver tissue was loaded to 10% tensile strain, small round-shaped voids started showing up inside the liver lobule (Fig. 2(a)); those voids were small in size and distributed homogeneously in the lobule, suggesting they may not be damage microfeatures but instead are recoverable structural deformation. When the global deformation reached 20% strain, thin cracks were found to emerge with an alignment pattern following local tissue anatomic texture (Fig. 2(b)). As an example, relatively large cracks had formed inside the lobules (Fig. 2(b), white arrows), accompanied by an array of lengthy, thin cracks (Fig. 2(b), black arrows). Major lacerations developed when the liver tissue reached 30% tensile strain, and the large tears crossed the boundary lines of multiple lobules (Fig. 2(c), white arrows). These major lacerations can reach a length of over 200 μm.

Fig. 2
Fig. 2
Close modal

Quantification of Microfeatures.

In the interrupted tensile test, an increasing trend was found in the number density, area fraction, and the mean area as the strain increases from the control to 30% strain (Fig. 3 and Table 1). On the other hand, mean NND showed a declining trend with increasing strain, which is likely due to an increase in the number of voids, and even merging to form larger voids (Fig. 3 and Table 1). For number density, area fraction, and mean NND parameters, 10%, 20%, and 30% strains were all found to be significantly different (p < 0.05) from the control. For the mean area parameter, only the 20% and 30% strains were significantly different (p < 0.05) from the control.

Fig. 3
Fig. 3
Close modal
Table 1

Microfeature parameters showed damage evolution quantitatively in interrupted tensile mechanical tests with STDEV

ObjectNumber density (/mm2)Area fractionMean area (μm2)Mean NND (μm)
Control8.66 ± 6.6611.13 ± 8.600.0004 ± 0.000333.11 ± 2.65145.36 ± 35.58
10%83.11 ± 17.97104.33 ± 22.590.0072 ± 0.003869.44 ± 32.2646.98 ± 6.93
20%118.99 ± 21.75149.38 ± 27.300.0099 ± 0.002880.46 ± 13.7638.54 ± 3.23
30%184.11 ± 37.37231.08 ± 46.900.0173 ± 0.005392.25 ± 26.6331.96 ± 3.16
ObjectNumber density (/mm2)Area fractionMean area (μm2)Mean NND (μm)
Control8.66 ± 6.6611.13 ± 8.600.0004 ± 0.000333.11 ± 2.65145.36 ± 35.58
10%83.11 ± 17.97104.33 ± 22.590.0072 ± 0.003869.44 ± 32.2646.98 ± 6.93
20%118.99 ± 21.75149.38 ± 27.300.0099 ± 0.002880.46 ± 13.7638.54 ± 3.23
30%184.11 ± 37.37231.08 ± 46.900.0173 ± 0.005392.25 ± 26.6331.96 ± 3.16

Evolution of Damage Microfeatures in Compression Interrupted Testing

Damage Pattern.

As the liver tissue was compressed to 10% strain, small thin fissure cracks were noticed (Fig. 4(a)). At 20% compressive strain, the density of fissure cracks increased (Fig. 4(b), white arrow) along with condensed tissue areas, which is displayed in a dark red color (Fig. 4(b), black arrow). As the deformation reached 30% compressive strain, more condensed tissue areas (Fig. 4(c), black arrow), fissure cracks, and small voids (Fig. 4(c), white arrows) were observed within the lobules. When the compressive force reached 40% strain, fissure cracks and voids had grown and coalesced (Fig. 4(d), white arrows).

Fig. 4
Fig. 4
Close modal

Quantification of Microfeatures.

Similar to the interrupted tensile tests, the interrupted compressive tests showed an increasing trend in the number density and the area fraction as the strain increased from the control to 40% compressive strain, as well as a decreasing trend in the mean NND (Figs. 5(a)5(c), and Table 2). The mean area had a decreasing trend from the control to 10% and 20% compressive strains but then it increased from 20% to 30% to 40% strains (Fig. 5(d)), likely reflecting the competition between the overall tissue compression and the growth of damage voids. For the compression data, number density and area fraction parameters (30% and 40% strain levels) were found to be significantly different (p < 0.05) from the control samples. There was no significant difference found between the other parameters.

Fig. 5
Fig. 5
Close modal
Table 2

Microfeature parameters showed damage evolution quantitatively in interrupted compressive mechanical tests with STDEV

ObjectNumber density (/mm2)Area fractionMean area (μm2)Mean NND (μm)
Control8.66 ± 6.6611.13 ± 8.600.0004 ± 0.000333.11 ± 2.65145.36 ± 35.58
10%43.67 ± 22.4554.80 ± 28.200.0027 ± 0.001447.83 ± 6.9462.72 ± 24.34
20%62.33 ± 35.7378.19 ± 44.900.0035 ± 0.002242.66 ± 4.3156.31 ± 12.36
30%98.50 ± 37.38123.99 ± 46.890.0075 ± 0.003162.18 ± 14.7844.37 ± 9.00
40%143.83 ± 35.67118.00 ± 44.280.0119 ± 0.004764.24 ± 15.7536.97 ± 2.75
ObjectNumber density (/mm2)Area fractionMean area (μm2)Mean NND (μm)
Control8.66 ± 6.6611.13 ± 8.600.0004 ± 0.000333.11 ± 2.65145.36 ± 35.58
10%43.67 ± 22.4554.80 ± 28.200.0027 ± 0.001447.83 ± 6.9462.72 ± 24.34
20%62.33 ± 35.7378.19 ± 44.900.0035 ± 0.002242.66 ± 4.3156.31 ± 12.36
30%98.50 ± 37.38123.99 ± 46.890.0075 ± 0.003162.18 ± 14.7844.37 ± 9.00
40%143.83 ± 35.67118.00 ± 44.280.0119 ± 0.004764.24 ± 15.7536.97 ± 2.75

Evolution of Damage Microfeatures in Interrupted Shear Testing

Damage Pattern.

When the liver tissue was subjected to a shear deformation of 0.8 rad, a few cracks (Fig. 6(a), white arrows) and regions of condensed tissue were observed (Fig. 6(a), black arrow). As the shear deformation was increased to 0.9 rad, large irregular-shaped voids were observed along with smaller voids and thin cracks (Fig. 6(b), white arrow). Regions of condensed tissue were also seen to increase (Fig. 6(b), black arrow). When the shear deformation reached 1.0 rad, major lacerations emerged (Fig. 6(c), white arrow) along the direction of the shear force. Regions of condensed tissue were also observed (Fig. 6(c), black arrow).

Fig. 6
Fig. 6
Close modal

Quantification of Microfeatures.

The interrupted shear tests demonstrated trends like the tensile interruption tests. As the strain moves from control to 1.0 rad, the number density, area fraction, and mean area parameters all show an increasing trend, while the mean NND parameter shows a decreasing trend as the strain increases (Fig. 7 and Table 3). For the number density, area fraction, and mean area, the values obtained at 0.9 and 1.0 rad were found to be significantly different (p < 0.05) from the control. Additionally, the mean NND parameter at 0.8 rad was also found to be significantly different (p < 0.05) from the control.

Fig. 7
Fig. 7
Close modal
Table 3

Microfeature parameters showed damage evolution quantitatively in interrupted shear mechanical tests with STDEV

ObjectNumber density (/mm2)Area fractionMean area (μm2)Mean NND (μm)
Control8.66 ± 6.6611.13 ± 8.600.0004 ± 0.000333.11 ± 2.65145.36 ± 35.58
0.8 rad45.78 ± 29.8157.50 ± 37.390.0025 ± 0.001740.61 ± 6.8285.16 ± 49.34
0.9 rad104.22 ± 56.46131.01 ± 70.900.0064 ± 0.003946.24 ± 6.4747.38 ± 22.66
1.0 rad152.22 ± 33.29190.99 ± 41.790.0110 ± 0.003557.09 ± 9.8635.97 ± 3.47
ObjectNumber density (/mm2)Area fractionMean area (μm2)Mean NND (μm)
Control8.66 ± 6.6611.13 ± 8.600.0004 ± 0.000333.11 ± 2.65145.36 ± 35.58
0.8 rad45.78 ± 29.8157.50 ± 37.390.0025 ± 0.001740.61 ± 6.8285.16 ± 49.34
0.9 rad104.22 ± 56.46131.01 ± 70.900.0064 ± 0.003946.24 ± 6.4747.38 ± 22.66
1.0 rad152.22 ± 33.29190.99 ± 41.790.0110 ± 0.003557.09 ± 9.8635.97 ± 3.47

Verification Experiments Revealing the Damage Nature of Major Microfeatures.

For tensile verification experiment, long cracks and major lacerations were observed in the unloaded liver tissues at 20% and 30% tensile strains (Figs. 8(b) and 8(c)). For compressive verification experiment, long fissure cracks and large voids were also visualized in the unloaded liver tissues at 30% and 40% compressive strains (Figs. 8(f) and 8(g)). For shear verification experiment, large irregular-shaped voids and major lacerations were observed in the unloaded liver tissues at 0.9 rad and 1.0 rad shear strains (Figs. 8(i) and 8(j)). The verification experiments of the three loading modes (Fig. 8 and Supplemental Table 1 which is available under the “Supplemental Data” tab for this paper on the ASME Digital Collection.) all showed that major microfeatures had remained after the release of the external loading, indicating the damage nature of those major microfeatures (permanent inelastic deformation).

Fig. 8
Fig. 8
Close modal

Discussion

In this study, damage evolution in the liver tissue under tension, compression, and shear loading modes were investigated via interrupted mechanical testing methodology for the first time. This interrupted mechanical testing methodology was able to capture the distinct stress-state dependent damage progression patterns of the porcine liver. For tension tests, the damage microfeatures were long cracks (20% strain) and major lacerations that crossed the boundary lines of liver lobules (30% strain) (Fig. 2). The degree and severity of the accumulated damages at 30% strain revealed that liver tissue is sensitive to large tensile deformation. For compression tests, fissure cracks and major voids formations occurred when compressive strains reached 30% and 40%; moreover, these microfeatures were relatively small and spatially confined, possibly indicating a tolerance for compressive deformation (Fig. 4). For shear tests, large voids with irregular shapes were seen at 0.9 rad and major lacerations were observed at 1.0 rad, thus implying a weak tolerance to shear deformation.

Both histology and quantification of microfeatures showed that, for tension and shear loading modes, damage progression was dominated by void growth and nucleation, which dramatically disrupted the structural integrity of the tissue. In addition, the majority of damage voids were located in the liver lobule for all three mechanical loading modes. This phenomenon could be explained by the structural makeup of the liver and the hepatic ECM [52]. Liver tissue is composed of many lobules separated by connective tissues, primarily dense collagen network. Each lobule consists of a central vein surrounded by six hepatic portal veins and six hepatic arteries, which are connected by many capillary-like tubes called sinusoids. These sinusoids are designed to slow down blood flow and allow the liver cells to filter the blood coming from the digestive tract, before passing it to the rest of the body; however, there is little structural support in these components except a loose collagen network, making it susceptible to damage as seen in the results. As load increases, the damage voids tend to propagate, coalesce, and increase in size. The histological results also showed that, for tension and shear at higher strains, some damage voids eventually extend outside the lobule into the connective tissue lining and the neighboring lobule. The observed damage progression behavior reflects the heterogeneous nature of liver lobe composition and ultrastructure, of which the damage occurs at the weaker portion first (inside the lobule) and then the stronger materials, i.e., lobule boundary consisting of connective tissue lining.

This study indicates that damage patterns under high tensile strains exhibited the most pronounced lacerations, which complements reports that describe tensile stress as the main contributor to liver lacerations frequently observed in traumatic abdominal injuries [49]. The damage evolution under compression loads is different so that it endures compression loads much better than tension and shear. This may be due to the interstitial water (water bounded in tissue ground substance such as proteoglycans) that bears compressive loads analogous to the responses observed in cartilage, i.e., proteoglycan aggregates and bound water bear much of the repetitive compressive loads [53,54].

This study presents new quantitative information regarding the evolution of damage under three modes of load application (tension, compression, and shear); however, there is still a need to correlate these mechanistic descriptions of damage to the physiological injuries/functional disruptions of a living liver. For example, the lobule tissue laceration could be an underlying cause of traumatic liver hemorrhage and irreversible organ failure. Further studies are needed in order to accurately determine injury thresholds critical for the successful implementation of FE models toward improving automobile safety.

As with any experiment, there are several limitations in this study. One limitation of this study is that it uses ex vivo tissue samples. As these tissues are easier to experiment on in the lab setting, they do not fully represent the living tissue condition. As in the case of the liver, there is no blood flow or active perfusion in these ex vivo tissue samples, which could alter the damage effects at each loading condition, especially compression where it might increase the damage propagation at lower levels of stress and strain. Future studies should look to address how the perfusion pressure inside the liver affects the tissue as it is subjected to an external load. Further, similar to all other soft tissue testings, our experimentation is subject to human error throughout the sample preparation (e.g., sample dissection) and testing phases (e.g., caliper measurements), although efforts were made to minimize these potential errors. The fact that displacement rate, as opposed to the normalized strain rate, was used across testings is also a limitation of this study. This limitation does not affect our conclusions since the current testings are in a low strain-rate or quasi-static regime. Another limitation involves the fixation and histology process, which are practical approaches for this study but could induce a small degree of shrinkage into the samples.

Additionally, porcine liver exhibits significant differences from human liver tissue. Porcine livers exhibit five lobes, whereas human livers exhibit only four. Additionally, hepatic lobules in porcine livers are more well-defined than in humans, presumably producing a biomechanical response that is different from human tissue. Porcine livers are also ultrastructurally different in design and ECM makeup. Even with all of these differences, porcine livers are regularly used as a model for describing the result of TI to the liver. Although not a perfect system, the use of porcine tissue in this study helps to establish a methodology without the use of human tissue, which is more difficult and more expensive to acquire. The goal of this study is to establish a methodology that future studies can use with human tissue and thus provide a more accurate description of damage evolution in liver tissue.

Conclusions

In this study, interrupted mechanical testing was successfully applied to investigate the damage progression of liver tissues under tension, compression, and shear loading conditions. The liver tissue showed different patterns of damage evolution under the three different loading conditions. By capturing damage evolution of different stress states, this study provides solid experimental data for physics-based computational modeling. In other words, this study was able to quantitatively obtain damage parameters that can be later used in the development of structure-based constitutive models. These results not only have an impact on TI simulation of the human body in civil scenarios such as car crashes or sports injuries, but they also have potential applications in visual training programs for battlefield medicine, in which detailed injuries (e.g., liver lacerations) and high fidelity visual presentations are a must to train the medical personnel to better handle and treat injured soldiers.

Acknowledgment

This material is based on the work supported by the Department of Energy, Southern Regional Center for Lightweight Innovative Design (SRCLID), under award number DE-EE0002323 (LW, JL). This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. JL is supported in part by 1R01EB022018-01.

Funding Data

• National Institutes of Health (1R01EB022018-01).

• U.S. Department of Energy (DE-EE0002323).

References

1.
Hardy
,
W. N.
,
Mason
,
M. J.
,
Foster
,
C. D.
,
Shah
,
C. S.
,
Kopacz
,
J. M.
,
Yang
,
K. H.
,
King
,
A. I.
,
Bishop
,
J.
,
Bey
,
M.
, and
Anderst
,
W.
,
2007
, “
A Study of the Response of the Human Cadaver Head to Impact
,”
Stapp Car Crash J.
,
51
, pp.
17
80
.
2.
King
,
A. I.
,
Viano
,
D. C.
, and
Mizeres
,
N.
,
1995
, “
Humanitarian Benefits of Cadaver Research on Injury Prevention
,”
J. Trauma Acute Care Surg.
,
38
(
4
), pp.
564
569
.
3.
Kuppa
,
S. M.
,
Klopp
,
G. S.
,
Crandall
,
J. R.
,
Hall
,
G.
,
Yoganandan
,
N.
,
Pintar
,
F. A.
,
Eppinger
,
R. H.
,
Sun
,
E.
,
Khaewpong
,
N.
, and
Kleinberger
,
M.
,
1998
, “
Axial Impact Characteristics of Dummy and Cadaver Lower Limbs
,”
16th International Technical Conference on the Enhanced Safety of Vehicles
, Windsor, ON, Canada, May 31–June 4, pp. 1608–1617.
4.
Backaitis
,
S. H.
, and
Mertz
,
H. J.
,
1993
,
Hybrid III: The First Human-like Crash Test Dummy
,
Society of Automotive Engineers
,
Warrendale, PA
, p.
830
.
5.
Deng
,
Y.-C.
,
1989
, “
Anthropomorphic Dummy Neck Modeling and Injury Considerations
,”
Accid. Anal. Prev.
,
21
(
1
), pp.
85
100
.
6.
Foster
,
J.
,
Kortge
,
J.
, and
Wolanin
,
M.
,
1994
, “
Hybrid III-A Biomechanically-Based Crash Test Dummy
,”
SAE
Paper No. 770938.
7.
Viano
,
D. C.
,
1988
, “
Limits and Challenges of Crash Protection
,”
Accid. Anal. Prev.
,
20
(
6
), pp.
421
429
.
8.
Jost
,
R.
, and
Nurick
,
G.
,
1999
, “
Finite Element Modelling of the Human Body in Vehicle Side Impact
,”
Int. J. Crashworthiness
,
4
(
1
), pp.
31
37
.
9.
Bibin
,
L.
,
Anquez
,
J.
,
de la Plata Alcalde
,
J. P.
,
Boubekeur
,
T.
,
Angelini
,
E. D.
, and
Bloch
,
I.
,
2010
, “
Whole-Body Pregnant Woman Modeling by Digital Geometry Processing With Detailed Uterofetal Unit Based on Medical Images
,”
IEEE Trans. Biomed. Eng.
,
57
(
10
), pp.
2346
2358
.
10.
Delotte
,
J.
,
Behr
,
M.
,
Baque
,
P.
,
Bourgeon
,
A.
,
De Peretti
,
F.
, and
Brunet
,
C.
,
2006
, “
Modeling the Pregnant Woman in Driving Position
,”
,
28
(
4
), pp.
359
363
.
11.
Song
,
E.
,
Trosseille
,
X.
, and
Baudrit
,
P.
,
2009
, “
Evaluation of Thoracic Deflection as an Injury Criterion for Side Impact Using a Finite Elements Thorax Model
,”
SAE
Paper No. 2009-22-0006.
12.
Tropiano
,
P.
,
Thollon
,
L.
,
Arnoux
,
P. J.
,
Huang
,
R. C.
,
Kayvantash
,
K.
,
Poitout
,
D. G.
, and
Brunet
,
C.
,
2004
, “
Using a Finite Element Model to Evaluate Human Injuries Application to the HUMOS Model in Whiplash Situation
,”
Spine
,
29
(
16
), pp.
1709
1716
.
13.
1999
, “
The Pedestrian Model in PC-Crash: The Introduction of a Multi Body System and Its Validation
,”
SAE
Paper No. 1999-01-0445.
14.
Fressmann
,
D.
,
Münz
,
T.
,
Graf
,
O.
, and
Schweizerhof
,
K.
,
2007
, “
FE Human Modelling in Crash–Aspects of the Numerical Modelling and Current Applications in the Automotive Industry
,”
LSDYNA Anwenderforum
, DYNAmore GmbH, Frankenthal, Germany, pp. F-I-23–F-I-34.
15.
Zhao
,
J.
, and
Narwani
,
G.
,
2005
, “
Development of a Human Body Finite Element Model for Restraint System R&D Applications
,” 19th International Technical Conference on the Enhanced Safety of Vehicles (
ESV
), Washington, DC, June 6–9, Paper No. 05-0399.
16.
Kirkpatrick
,
S. W.
,
MacNeill
,
R.
, and
Bocchieri
,
R. T.
,
2003
, “
Development of an LS-DYNA Occupant Model for Use in Crash Analyses of Roadside Safety Features
,” Transportation Research Board 82nd Annual Meeting, Washington, DC, Jan. 12–16, Paper No.
03-4450
.
17.
King
,
A.
, and
Chou
,
C.
,
1976
, “
Mathematical Modelling, Simulation and Experimental Testing of Biomechanical System Crash Response
,”
J. Biomechanics.
,
9
(
5
), pp.
301
317
.
18.
Horgan
,
T.
, and
Gilchrist
,
M.
,
2004
, “
Influence of FE Model Variability in Predicting Brain Motion and Intracranial Pressure Changes in Head Impact Simulations
,”
Int. J. Crashworthiness
,
9
(
4
), pp.
401
418
.
19.
Jakeman
,
L. B.
,
Guan
,
Z.
,
Wei
,
P.
,
Ponnappan
,
R.
,
Dzwonczyk
,
R.
,
Popovich
,
P. G.
, and
Stokes
,
B. T.
,
2000
, “
Traumatic Spinal Cord Injury Produced by Controlled Contusion in Mouse
,”
J. Neurotrauma.
,
17
(
4
), pp.
299
319
.
20.
Kleiven
,
S.
, and
Hardy
,
W. N.
, eds.,
2002
, “
Correlation of an FE Model of the Human Head With Local Brain Motion-Consequences for Injury Prediction
,”
SAE
Paper No. 2002-22-0007.
21.
Mao
,
H.
,
Zhang
,
L.
,
Yang
,
K. H.
, and
King
,
A. I.
,
2006
, “
Application of a Finite Element Model of the Brain to Study Traumatic Brain Injury Mechanisms in the Rat
,”
Stapp Car Crash J.
,
50
, pp.
583
600
.
22.
Richens
,
D.
,
Field
,
M.
,
Hashim
,
S.
,
Neale
,
M.
, and
Oakley
,
C. A.
,
2004
, “
Finite Element Model of Blunt Traumatic Aortic Rupture
,”
Eur. J. Cardio-Thoracic Surg.
,
25
(
6
), pp.
1039
1047
.
23.
Snedeker
,
J. G.
,
Bajka
,
M.
,
Hug
,
J.
,
Szekely
,
G.
, and
Niederer
,
P.
,
2002
, “
The Creation of a High‐Fidelity Finite Element Model of the Kidney for Use in Trauma Research
,”
J. Visualization Comput. Animation.
,
13
(
1
), pp.
53
64
.
24.
Miller
,
K.
,
2000
, “
Constitutive Modelling of Abdominal Organs
,”
J. Biomech.
,
33
(
3
), pp.
367
373
.
25.
Li
,
Z.
,
Kim
,
J.-E.
,
Davidson
,
J. S.
,
Etheridge
,
B. S.
,
Alonso
,
J. E.
, and
Eberhardt
,
A. W.
,
2007
, “
Biomechanical Response of the Pubic Symphysis in Lateral Pelvic Impacts: A Finite Element Study
,”
J. Biomech.
,
40
(
12
), pp.
2758
2766
.
26.
Iwamoto
,
M.
,
Kisanuki
,
Y.
,
Watanabe
,
I.
,
Furusu
,
K.
,
Miki
,
K.
, and
Hasegawa
,
J.
,
2002
, “
Development of a Finite Element Model of the Total Human Model for Safety (THUMS) and Application to Injury Reconstruction
,”
International Research Council on Biomechanics of Injury
(
IRCOBI
), Munich, Germany, Sept. 18–20, pp.
31
42
.
27.
Iwamoto
,
M.
,
Nakahira
,
Y.
, and
Kimpara
,
H.
,
2015
, “
Development and Validation of the Total Human Model for Safety (THUMS) Toward Further Understanding of Occupant Injury Mechanisms in Precrash and During Crash
,”
Traffic Inj. Prev.
,
16
(
Suppl. 1
), pp.
S36
S48
.
28.
Schwartz
,
D.
,
Guleyupoglu
,
B.
,
Koya
,
B.
,
Stitzel
,
J. D.
, and
Gayzik
,
F. S.
,
2015
, “
Development of a Computationally Efficient Full Human Body Finite Element Model
,”
Traffic Inj. Prev.
,
16
(
Suppl. 1
), pp.
S49
S56
.
29.
Vavalle
,
N. A.
,
Davis
,
M. L.
,
Stitzel
,
J. D.
, and
Gayzik
,
F. S.
,
2015
, “
Quantitative Validation of a Human Body Finite Element Model Using Rigid Body Impacts
,”
Ann. Biomed. Eng.
,
43
(
9
), pp.
2163
2174
.
30.
Kimpara
,
H.
,
Nakahira
,
Y.
, and
Iwamoto
,
M.
,
2016
, “
Development and Validation of THUMS Version 5 With 1D Muscle Models for Active and Passive Automotive Safety Research
,”
IEEE 38th Annual International Conference of the Engineering in Medicine and Biology Society
(
EMBC
), Orlando, FL, Aug. 16–20, pp. 6022–6025.
31.
Shao
,
Y.
,
Zou
,
D.
,
Li
,
Z.
,
Wan
,
L.
,
Qin
,
Z.
,
Liu
,
N.
,
Zhang
,
J.
,
Zhong
,
L.
,
Huang
,
P.
, and
Chen
,
Y.
,
2013
, “
Blunt Liver Injury With Intact Ribs Under Impacts on the Abdomen: A Biomechanical Investigation
,”
Plos One.
,
8
(
1
), p.
e52366
.
32.
Umale
,
S.
,
Deck
,
C.
,
Bourdet
,
N.
,
Dhumane
,
P.
,
Soler
,
L.
,
Marescaux
,
J.
, and
Willinger
,
R.
,
2013
, “
Experimental Mechanical Characterization of Abdominal Organs: Liver, Kidney & Spleen
,”
J. Mech. Behav. Biomed. Mater.
,
17
, pp.
22
33
.
33.
Nahum
,
A. M.
, and
Melvin
,
J.
, eds., 2012,
Accidental Injury: Biomechanics and Prevention
,
Springer
, New York.
34.
Kemper
,
A. R.
,
Santago
,
A. C.
,
Stitzel
,
J. D.
,
Sparks
,
J. L.
, and
Duma
,
S. M.
,
2010
, “
,”
,
54
, pp.
15
26
.
35.
Sparks
,
J. L.
,
Bolte
,
I. V. J. H.
,
Dupaix
,
R. B.
,
Jones
,
K. H.
,
Steinberg
,
S. M.
,
Herriott
,
R. G.
,
Stammen
,
J. A.
, and
Donnelly
,
B. R.
,
2007
, “
Using Pressure to Predict Liver Injury Risk From Blunt Impact
,”
Stapp Car Crash J.
,
51
, p.
401
.
36.
Untaroiu
,
C. D.
,
Lu
,
Y.-C.
,
Siripurapu
,
S. K.
, and
Kemper
,
A. R.
,
2015
, “
Modeling the Biomechanical and Injury Response of Human Liver Parenchyma Under Tensile Loading
,”
J. Mech. Behav. Biomed. Mater.
,
41
, pp.
280
291
.
37.
Schueller
,
G.
,
2012
, “
Liver and Bile Ducts
,”
,
Springer
, Berlin, pp.
31
54
.
38.
Gourgiotis
,
S.
,
Vougas
,
V.
,
Germanos
,
S.
,
Dimopoulos
,
N.
,
Bolanis
,
I.
,
Drakopoulos
,
S.
,
Alfaras
,
P.
, and
Baratsis
,
S.
,
2007
, “
Operative and Nonoperative Management of Blunt Hepatic Trauma in Adults: A Single-Center Report
,”
J. Hepato-Biliary-Pancreatic Surg.
,
14
(
4
), pp.
387
391
.
39.
Ahmed
,
N.
, and
Vernick
,
J. J.
,
2011
, “
Management of Liver Trauma in Adults
,”
J. Emerg. Trauma Shock.
,
4
(
1
), p.
114
.
40.
Agarwal
,
H.
,
Gokhale
,
A.
,
Graham
,
S.
, and
Horstemeyer
,
M.
,
2003
, “
Void Growth in 6061-Aluminum Alloy Under Triaxial Stress State
,”
Mater. Sci. Eng.: A.
,
341
(
1–2
), pp.
35
42
.
41.
Gall
,
K.
,
Horstemeyer
,
M. F.
,
Degner
,
B. W.
,
McDowell
,
D. L.
, and
Fan
,
J.
,
2001
, “
On the Driving Force for Fatigue Crack Formation From Inclusions and Voids in a Cast A356 Aluminum Alloy
,”
Int. J. Fract.
,
108
(
3
), pp.
207
233
.
42.
McVeigh
,
C.
,
Vernerey
,
F.
,
Liu
,
W. K.
,
Moran
,
B.
, and
Olson
,
G.
,
2007
, “
An Interactive Micro-Void Shear Localization Mechanism in High Strength Steels
,”
J. Mech. Phys. Solids.
,
55
(
2
), pp.
225
244
.
43.
Baker
,
A.
,
Perry
,
E.
, and
Fossard
,
D.
,
1986
, “
Traumatic Rupture of the Stomach Due to a Seat Belt
,”
Injury.
,
17
(
1
), pp.
47
48
.
44.
Viano
,
D. C.
,
King
,
A. I.
,
Melvin
,
J. W.
, and
Weber
,
K.
,
1989
, “
Injury Biomechanics Research: An Essential Element in the Prevention of Trauma
,”
J. Biomech.
,
22
(
5
), pp.
403
417
.
45.
Feliciano
,
D. V.
, and
Rozycki
,
G.
,
1994
, “
The Management of Penetrating Abdominal Trauma
,”
,
28
, pp.
1
39
.
46.
Tinkoff
,
G.
,
Esposito
,
T. J.
,
Reed
,
J.
,
Kilgo
,
P.
,
Fildes
,
J.
,
Pasquale
,
M.
, and
Meredith
,
J. W.
,
2008
, “
American Association for the Surgery of Trauma Organ Injury Scale I: Spleen, Liver, and Kidney, Validation Based on the National Trauma Data Bank
,”
J. Am. Coll. Surg.
,
207
(
5
), pp.
646
655
.
47.
Gaspar
,
B.
,
Negoi
,
I.
,
Sorin
,
P.
,
Hostiuc
,
S.
,
Ganescu
,
R.
, and
Beuran
,
M.
,
2014
, “
Selective Nonoperative Management of Abdominal Injuries in Polytrauma Patients: A Protocol Only for Experienced Trauma Centers
,”
Maedica
,
9
(
2
), p.
168
.
48.
Negoi
,
I.
,
Paun
,
S.
,
Stoica
,
B.
,
Tanase
,
I.
,
Vartic
,
M.
,
Negoi
,
R. I.
,
Hostiuc
,
S.
, and
Beuran
,
M.
,
2016
, “
Latest Progress of Research on Acute Abdominal Injuries
,”
J. Acute Dis.
,
5
(
1
), pp.
16
21
.
49.
Clemmer
,
J.
,
Prabhu
,
R.
,
Chen
,
J.
,
Colebeck
,
E.
,
Priddy
,
L. B.
,
Mccollum
,
M.
,
Brazile
,
B.
,
Whittington
,
W.
,
Wardlaw
,
J. L.
,
Rhee
,
H.
,
Horstemeyer
,
M. F.
,
Williams
,
L. N.
, and
Liao
,
J.
,
2016
, “
Experimental Observation of High Strain Rate Responses of Porcine Brain, Liver, and Tendon
,”
J. Mech. Med. Biol.
,
16
(
3
), p.
1650032
.
50.
Roan
,
E.
, and
Vemaganti
,
K.
,
2007
, “
The Nonlinear Material Properties of Liver Tissue Determined From No-Slip Uniaxial Compression Experiments
,”
ASME J. Biomech. Eng.
,
129
(
3
), pp.
450
456
.
51.
Clemmer
,
J.
,
Liao
,
J.
,
Davis
,
D.
,
Horstemeyer
,
M. F.
, and
Williams
,
L. N.
,
2010
, “
A Mechanistic Study for Strain Rate Sensitivity of Rabbit Patellar Tendon
,”
J. Biomech.
,
43
(
14
), pp.
2785
2791
.
52.
Bissell
,
D.
, and
Choun
,
M.
,
1988
, “
The Role of Extracellular Matrix in Normal Liver
,”
Scand. J. Gastroenterol.
,
23
(
Suppl. 151
), pp.
1
7
.
53.
Roughley
,
P. J.
,
2001
, “
Articular Cartilage and Changes in Arthritis: Noncollagenous Proteins and Proteoglycans in the Extracellular Matrix of Cartilage
,”
Arthritis Res.
,
3
(
6
), p.
342
54.
Iozzo
,
R. V.
,
2000
,
Proteoglycans: Structure, Biology and Molecular Interactions
,
CRC Press
, Boca Raton, FL.