Poster Presentation NSW State Cancer Conference 2023

Computational modelling and simulations as tools for treatment personalisation in ovarian cancer (#209)

Marilisa Cortesi 1 2 , Dongli Liu 1 , Elyse Powell 3 , Kristina Warton 3 , Caroline Ford 3
  1. School of Clinical Medicine, Gynaecological Cancer Research Group, Randwick, NSW, Australia
  2. Electrical Electronic and Information Engineering "G. Marconi", University of Bologna, Cesena, FC, Italy
  3. School of Clinical Medicine, Gynaecological Cancer Research Group, Randwick, NSW, Australia

Personalised medicine in ovarian cancer is hampered by the lack of knowledge regarding which individual features affect drug response and how they vary among subjects/ therapeutic agents. Computational models (CMs) are effective at identifying these complex patterns and can be used as tools to enable personalised medicine [1].

In this work, we investigated the use of CMs for treatment personalisation in high grade serous ovarian cancer (HGSOC) through the development and validation of a CM of ovarian cancer metastasis.

The CM was developed combining a finite-element and an agent-based model. This structure enables the simulation of both cell behaviour and of the diffusion of nutrients and drugs through the virtual tissue. The CM comprises 3 different cell types: fibroblasts, mesothelial and cancer cells, and their spatial organisation mimics the structure of the omentum, one of the most common metastasis sites for HGSOC. Simulations of this CM result in temporal evolution of the behaviour/status of each simulated cell and the corresponding local drug and nutrients concentration.

The behaviour of 4 HGSOC cell lines was characterised through the evaluation of adhesion, invasion and response to common chemotherapies cisplatin, carboplatin and paclitaxel. The response to the same drugs was also evaluated in primary cells derived from ascites fluid of 3 patients. Whenever appropriate, experiments were conducted in a 3D organotypic model of the omentum lining [2].

The CM was initially calibrated to replicate the behaviour of fibroblasts, mesothelial cells and each HGSOC cell line. The resulting parameter configurations were then used in a preliminary analysis aimed at modelling the behaviour of each primary sample. A procedure to select which parameter set best approximates specific biological features (e.g., proliferation, invasion) was devised and used to identify which combination of simulated HGSOC cell lines was expected to better replicate each primary sample. Computational simulations of these parameter mixtures were then conducted, and the results compared with the treatment responses measured in the laboratory models.

Our analysis confirmed that the computational model was effective in recapitulating treatment response in the individual patients and provides a tool to study the interaction between cancer and healthy cells at the metastasis site.

While further analysis on a larger sample pool is required to confirm the effectiveness and accuracy of this method, the proposed approach holds great potential, both to study inter-patient variability in HGSOC progression and metastasis and as a personalised medicine tool for the treatment of this disease.

  1. [1] Algethami M, Kulkarni S, Sadiq MT, Tang HKC, Brownlie J, Jeyapalan JN, Mongan NP, Rakha EA, Madhusudan S. Towards Personalized Management of Ovarian Cancer. Cancer Manag Res. 2022;14:3469-3483 https://doi.org/10.2147/CMAR.S366681
  2. [2] Kenny, H., Lal-Nag, M., White, E. et al. Quantitative high throughput screening using a primary human three-dimensional organotypic culture predicts in vivo efficacy. Nat Commun 6, 6220 (2015). https://doi.org/10.1038/ncomms7220