Animal tumor models are still essential for the development of new medication and therapy concepts. For the field of oropharyngeal cancer only few reliable xenogeneic tumor models are available. In a two-part experiment a new xenogeneic tumor model was established. PART 1 OF THE STUDY: The growth rates of two different tumor cell lines were compared in the RAG-2 mouse, the SCID mouse and the Swiss nude mouse. Cells from a lymph-node metastasis of an oral squamous cell carcinoma (XF-354) showed a better growth rate and clearer histology than cells from a primary squamous cell carcinoma of the floor of the mouth (UM-SCC-14C). The tumor growth rate was highest on the RAG-2 mouse, followed by the SCID mouse. The Swiss nude mouse showed no tumor growth at all. The combination of the XF-354 tumor cell line and the RAG-2 mouse was the most successful, with a tumor growth rate of 95%. The animal model is very reliable and robust and enables manipulations under anaesthesia outside the sterile barrier for up to 30 min. The single steps for cell cultivation and tumor implantation are described and discussed in detail. PART 2 OF THE STUDY: The second part of this study investigated the clinical efficacy of the macromolecular linked photosensitizer--mTHPC-PEG compared with the free photosensitizer mTHPC. Macromolecular linked photosensitizers have theoretical advantages owing to their pharmacokinetics and tumor selectivity. For the animal experiment, the photosensitizer mTHPC showed significantly better results than the macromolecular linked photosensitizer mTHPC-PEG. The reasons for these in vivo results are discussed in detail.