Abstract:
Mathematical modelling is used to validate:possible mechanisms of tumour growth, resistance, and treatment outcome. A key task of precision medicine is to predict theevolution in time of a tumour under treatment, as this can beused to optimise individual treatment. There is huge heterogeneity of tumour dynamics among patients. We used multitype measurements acquired routinely on a single breast tumour, including histopathology, magnetic resonance images,and molecular profiling, to personalize a complex multiscale model of breast cancer treated with chemotherapeutic and antiangiogenic agents.The model accounts for drug pharmacokinetics and pharmacodynamics and can be seen asa digital twin of the biological tumour. We show how our modelcan predict the outcome of the treatment, using only thebaseline data at treatment initialisation. One way to describethe time dynamic of the tumour is through the ordinary differential equation (ODE)followed by the tumour density, acting as a biomarker of treatment response. We use a neural network to estimate the ordinary differential equation of the individual tumour in symbolic form. indeed, we will show theODE of a breast cancer patient, discuss possibilities and limitations.The combination ofcomplex mathematical modelling (differential equations, cellular automata, stochastic processes),approximate Bayesian inference and machine learning can open new avenues for precision medicine.
About the Speaker:
Arnoldo Frigessi is an internationally renowned biostatisticianand a member of the Norwegian Academy of Science andLetters. He is currently the Head of the Department ofBiostatistics at the University of Oslo, where his researchfocuses on big data modeling and statistical methods in fieldssuch as cancer, infectious diseases, and genetics.
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
