OBJECTIVES: Modern targeted anticancer therapy introduces powerful treatment modality in cancer care. However, due to its cost the overall number of administrations has to be monitored and prospectively planned. The objective of this work is to design, apply and verify a multi-component statistical model that employs data from population records, clinical registries and data provided by an expert panel to predict the number of cancer patients who would be probably treated with targeted therapy in the following years. METHODS: Computations are primarily based on data from the Czech National Cancer Registry, which covers the entire population since 1977. The modelling process has two steps. In the first step, number of potentially treated cancer patients irrespective of drug is identified. The estimate combines the prediction of number of newly diagnosed and treated patients with the prediction of number of patients who were diagnosed previously and would probably be treated in a given year due to a relapse or progression of the primary malignant disease. In the second step patients are localised to individual drugs according to treatment scheme provided by the panel of experts, adjusted using information from clinical registries. RESULTS: The model currently provides predictions of potentially treated patients with quantified uncertainties. Regarding the most common targeted anticancer therapies, following numbers of treated patients are estimated for 2009 (90% CI):  836 (725; 948) patients receiving trastuzumab for breast cancer, 1283 (859; 1705) patients receiving bevacizumab for colorectal cancer, 300 (151; 449) patients receiving cetuximab for colorectal cancer and 514 (469; 557) patients receiving erlotinib for non-small cell lung carcinoma. CONCLUSIONS: Presented model allows us to predict the number of patients requiring targeted anticancer therapy in the years to come. Final outcome is used by decision makers for planning of health care expenses in the Czech Republic.