In vivo mechanical modeling of multilayer biological soft tissue

Abstract

Results: Validation studies with tissue mimicking agar gel phantoms showed low relative error of ~7% for two-layer phantoms and ~10% error for three layer phantoms when compared to known ground-truth values obtained using a commercial material testing system.

Conclusions: An ultrasound elastography-based technique has been developed into a reliable tool to characterize the nonlinear, anisotropic mechanical response of multi-layered soft tissue in vivo. The insights obtained from the in vivo and ex vivo studies provide a basis for application in surgical simulation, planning and diagnosis of diseased tissue.

Finally, the role of residual stresses on layer-specific tissue behavior was studied. Post compression testing, the tissue layers were separated to illustrate that the muscularis exists under tension and the mucosa-submucosa layer exists under compression while in vivo. The change in layer geometry was measured and introduced into the kinematics of the finite element model, using a multiplicative split of the deformation gradient tensor, to introduce residual stresses. The results showed that the muscularis (6.61± 4.38 kPa vs. 18.39± 14.41 kPa) and submucosa (3.81± 3.73 kPa vs. 12.16± 12.58 kPa) behaved softer and the mucosa behaved stiffer (0.55± 0.29 kPa vs. 0.35± 0.31 kPa) in the in vivo ‘no-load’ state as compared to the ‘stress-free’ state. The introduction of residual stresses also impacted the degree of anisotropy for the submucosa (anisotropic parameter 1.61±0.64 vs. 2.86±1.47) and mucosa (anisotropic parameter 1.22±0.42 vs. 2.77±0.95) from the in vivo ‘no-load’ state to the ‘stress-free’ state. The apparent softening of the muscularis and submucosa and stiffening of the mucosa in the in vivo ‘no-load’ state highlights the potential self-protective mechanisms present in the stomach to avoid injury.

The in vivo porcine stomachs paid particular attention to quantifying the inhomogeneity and anisotropy of tissue response. These in vivo studies illustrated that the initial shear moduli of the muscularis (5.69±4.06 kPa) was the highest followed by the submucosa (3.04±3.32 kPa) and the mucosa (0.56±0.28 kPa), a similar trend to the ex vivo results. Also, the cardiac muscularis was observed to be stiffer than the fundic muscularis (shear moduli of 7.96±3.82 kPa vs. 3.42±2.96 kPa), more anisotropic (anisotropic parameter of 2.21±0.77 vs. 1.41±0.38), and strongly distinguishable from its gastric fundic region counterpart based on a multivariate analysis. These results highlight the inhomogeneity and anisotropy of multilayer stomach tissue and are expected given the morphology of the underlying muscularis structure. The results also provide direct evidence of the drastic difference in in vivo and ex vivo material properties and the impact on the quality of material models obtained from ex vivo experiments. A statistical comparison of the initial shear moduli for each layer shows that the muscularis and submucosa are softer, while the mucosa is stiffer while in vivo.

Background: Mechanical characterization of the nonlinear, anisotropic behavior of soft tissue, particularly, when in vivo, accounting for internal stresses, is a challenging problem. For multi-layer organs such as the stomach, being able to discern accurate layer-specific properties adds significantly to this complexity. This thesis is aimed at addressing these challenges by developing an ultrasound elastography-based technique to measure and model the in vivo mechanical behavior of multi-layer organs, particularly the muscularis, submucosa and mucosa of the stomach.

Methods: To test tissue in vivo, a robotically manipulated compression head instrumented with a 30 MHz high-resolution ultrasound probe and load-cells was designed and developed. The layer-specific material properties of the tissue were iteratively optimized, as part of a finite element model, by minimizing an objective function that considers the difference between experimentally measured and computed axial and lateral displacements of the tissue as well as reaction forces at the indenter. The technique was successfully validated using tissue mimicking agar-based phantoms of varying material properties and geometries. Following validation, ex vivo studies were performed followed by in vivo studies on 10 porcine stomachs.

Following validation, intact ex vivo porcine stomach studies showed that the initial shear modulus of the muscularis (12.82±1.2 kPa) was the highest followed by the submucosa (10.42±1.14 kPa) and mucosa (0.12±0.04 kPa).