Experimentation and modelling of the human oesophagus

Abstract

The oesophagus is a primarily mechanical organ which transports food from the pharynx to the stomach through a series of muscular contractions. As of yet, the passive mechanical properties of the human oesophagus have not been established [1]. This is essential for proper investigation into how clinical diseases affect the oesophagus’ primary function. Further to this, mechanical data on the oesophagus has applications within tissue engineering [2], and computational models formed from this data can be used to enhance medical device design and surgical simulations [3]. Currently, the vast majority of ex vivo experimentation on the oesophagus has been carried out using animal tissue [2]. While this may give a good representation of how human tissue might behave, it cannot be used to accurately model the human oesophagus for applications within medicine. The oesophagus is one of the few visceral organs that can be separated into two distinct layers; the mucosa-submucosa and the muscularis propria. Currently, outside of this author’s work [4], the human oesophagus has not been studied in regard to its layerdependent properties. This work investigates the layer-dependent properties through a series of increasing deformation level cyclic tests performed under uniaxial tension. First, the oesophagus was explanted by means of dissection. Next, the organ was separated into the two main layers. Specimens were then tested in both the longitudinal and circumferential directions to observe any anisotropy. The cyclic tests were conducted at two different strain rates and included two cycles per deformation level to observe the pre-conditioning behaviour. The results displayed anisotropic behaviour, with greater stiffness in the longitudinal direction than the circumferential direction for both of the layers. The mucosa-submucosa was found to be stiffer than the muscularis propria layer in both directions. The observed behaviours were then modelled using two different models. For the first, the first loading path of each deformation level was extracted and an anisotropic, hyperelastic matrix-fibre model based on histological observations was used to simulate the hyperelastic behaviour. Next, the viscoelastic behaviour and stress-softening of the tissue were captured through modelling of the cyclic testing results. Both models produced a good fit with the experimental data.