The process of metastasis formation is still subject of intense discussion and even established models differ considerably in several basic details and conclusions drawn from them. In addition, it is difficult to evaluate these different models for their clinical relevance. A computer model was developed which permits comparison of such models quantitatively with clinical and experimental data including effects of treatment interventions.

The computer model is based on a discrete event simulation procedure. The growth of the primary tumour and the metastases is described via analytical functions, while a rate function models the intravasation events of the primary tumour and its metastases. Events describe the behaviour of the emitted malignant cells until the formation of new metastases.

The comparison of the simulated data with the clinical data of a hepatocellular carcinoma revealed that metastasis is an early event and that late metastases are of no clinical relevance for the patient’s outcome if he is left untreated. Only the first metastases seeded from the primary tumour contribute significantly to the tumour burden and thus cause the patient’s death [1]. The analysis of experimental data of a HT29 human colon cancer xenograft mouse model showed that natural killer cells decelerate the growth of the primary tumour, kill 80% of those disseminated tumour cells which could have otherwise established a new metastasis and hamper the proliferation of the malignant cells in distant tissue.

Furthermore, the computer model allows simulating and analysing the effects of different treatments, such as the resection of the primary tumour, radionuclide therapy, radioembolization and chemotherapy.