Introduction: Understanding of the pancreas is impacted by its fragility. This in part has led to the under-utilisation of pancreata for transplantation. Gaseous oxygen perfusion (persufflation) is a technique which has shown promise in improving pancreas preservation. However, one key challenge still remains: the lack of knowledge about the interaction between the organ and its preservation.
In this pilot study, we describe an unbiased mathematical approach to understand the ScubaTx persufflation device, the pancreas and their interaction to help identify and describe physical and/or biological factors that may impact the quality of the preservation.

Method: Using the fully automated ScubaTx  device, a pancreas (or suitable analogue), was preserved to gather data regarding key parameters (i.e. flow/pressure). By utilising a combination of lumped parameter modelling, and system identification techniques, we derive equations that describe aspects of the device and pancreas; the amalgamation of which enabled the study of their interaction. Refinement of the equation was performed through experimentation and data-model fusion. The setup used can be seen in Figure 1.

Results:Lumped-parameter fluid and grey-box modelling enabled the creation of an accurate steady-state model of the device and the pancreas. The derived equations that describe the system were obtained through specification or experimentation and calibrated to accurately reflect the preservation setup and embedded into a digital model. Changes in the device, organ or the preservation process, such as leaks or blockages in the vasculature or the device's pneumatics, were accurately detected as anomalies in the model.
System internal resistance was derived from the base equations. This gives a combined measure of the pancreas-machine interaction that can provide an indication of the state of the graft's vascular biology. A change in resistance is reflected in the model and can allude to changes in the device setup or specific resultant biological interactions. Varying the resistance of an organ analogue allowed for the identification of the device's relevant parameters and improved the calibration process and the accuracy of the model.

Conclusion: Exploiting techniques from engineering domains allowed us to develop a model that describes the ScubaTx device setup, the pancreas under preservation and their interaction. We are able to observe vascular occlusions such as blockages or leaks and hope that in future works we can generate enough data to classify the various anomalies. We are working further towards a deeper understanding of the vascular-pancreas and how it interacts with preservation devices. Moving forward, we hope to deepen our understanding of the biology of the pancreas to better understand organ preservation, disease and transplant survival.