Semi-mechanistic modelling of neutropenia in oncology drug development

Resumen

The work presented in this thesis describes the application of the population modelling approach in oncology drug development, mainly focussing on the haematological adverse events related to cytotoxic drugs. The thesis starts comparing several semi-mechanistic PK/PD models published in the literature to describe drug related neutropenia. Results indicated that the semi-mechanistic model first described by Friberg et al, 2002 provides a good balance of predictive capacity and parsimony of parameters. Then, the neutropenic effects of a drug in clinical development (BI 2536) were described semi-mechanistically, and it was shown how the modelling results could be applied to support further development. Also, the feasibility of predicting the neutropenic effects of a combination treatment between an already approved drug (Pemetrexed) and a drug in clinical development (BI 2536) was investigated. The predictions based on monotherapy data and assuming additivity were adequate, showing another application of the model in the planning of clinical studies. Finally the predictive ability of the semi-mechanistic model for two investigational drugs was evaluated. It was shown that when the model parameters were conditioned on data from trials different from a trial being evaluated, similar predictions of the drug related-neutropenia profiles were obtained as when the parameters were conditioned on the actual data of the evaluated trial.