Once the model is calibrated, it can be used to study the emergent behaviour of tissue under various scenarios. We applied it to investigate liver regeneration and the recurrence of Hepatocellular Carcinoma (HCC). Below, I highlight the key results:
- We implemented our model to test another degree of Partial Hepatectomy (PH). Since 50% PH has now been successfully and frequently used for surgical removal of liver associate tumors, we have performed an in silico 50% PH. Our model shows that here, similar to the 70% PH, hepatocytes enlarge their volume and also proliferate but with a reduced replication rate.
- We used our model to predict the potential recurrence of the HCC tumour in the remnant liver after a 70% PH. The outcome of our simulations is in accordance with clinical observations, which comes to reinforce our confidence in the applicability of this approach in other scenarios.
- Following the previous study, and based on the fact that HCC growth rate is likely affected by host factors, we have performed an exploratory sensitivity analysis by varying ±10% the input variables that feed our model. We found that there are three parameters that according to our model most likely make an impact on the tumour growth: the hepatocyte oxygen uptake constant, cancer cells cycle duration and cancer cells oxygen uptake constant. We take this as a proof of concept of how knowing different parameters that determine tumour development predict such behaviour in this 3-D off-lattice agent-based model.
- We have also studied the tumour growth rate based on the initial position of the residual tumour clone. We observed that tumour volume varies depending on the location of the residual tumour clone. It grows faster when located at the periphery of the liver. This result suggests that, although the geometry of the liver does not influence when studying liver regeneration, it might make an impact on the HCC recurrence.
It is important to mention that even though at its current stage, our model can qualitatively and quantitatively capture characteristics of the processes of liver regeneration and hepatocellular carcinoma recurrence, it is not calibrated to any particular type of patient specific parameters (age, sex, pre-existing conditions,…). This is obviously a handicap for the model’s direct application in clinical practice. However, the algorithm can be fed with specific patient factors, and in this way the model would easily transform into a workbench for hypothesis testing.