Individualized tumor growth mathematical modeling allows for more accurate predictions, optimization of therapy, and low-cost, safe evaluation for different strategies. In this research effort, we used skin cancer data from mice and developed efficient Nonlinear System Identification methods to estimate future tumor volume from noisy data. We used several techniques involving maximum likelihood and maximum a-posteriori estimators. A novelty of this work is that we used numerical integration to compute likelihood functions from noisy measurements. We published two papers ('Tumor growth modeling: Parameter estimation with Maximum Likelihood methods', and 'Individualized growth prediction of mice skin tumors with maximum likelihood estimators', both published in Computer Methods and Programs in Biomedicine).
For this research, I cooperate with Dr. S. Patmanidis and professors A. Charalampidis from CentraleSupélec, G.P. Papavassilopoulos from NTUA and G. Mitsis from McGill University.