In recent years, the development of xenograft models using patient-derived human xenotransplants (PDX) has become increasingly important in addressing previous inconsistencies between human tumours and traditional xenotranplant models of human cell lines. PDXs have been shown to more faithfully reflect clinical responses achieved in patients. For this reason, the NCI has recently moved to use PDXs as a platform technology for their cancer drug screens. Despite reports on the genomic, proteomic and transcriptomic similarities between PDX models and their tumours of origin there is a paucity of data characterising the pressures and events that occur in the immediate period following implantation. We now report our findings on the establishment of a head and neck cancer (HNC) xenotransplant model in NSG mice. In the first instance we established a robust technique for reducing the microbial burden of the resected tumour prior to implantation. Following this we used a combination of histopathology and immunohistochemistry to follow the fate of tumour tissue following implantation. To date we have implanted over 45 HNC specimens. For 9 of these specimens we quantitated i) proliferation (BrdU), ii) apoptosis (cleaved caspase 3), iii) tumour volume, iv) presence of human cytokeratin positive tissue, v) histopathological evidence of tumour tissue and vi) histopathological evidence of tumour necrosis at 4,7 and 14 days post implantation. We report attrition of the implanted tumour with little evidence of detectable tumour tissue at 14 days post-implantation. Despite these acute events tumours start to develop within 3-4 months and can be used to successfully generate up to 20 2nd generation PDX mice. Second generation tumours form with 3-6 weeks of implantation and are currently being used in a clinical trial by proxy. In 80% of cases the histopathological characteristics of the parent, 1st generation and 2nd generation were histopathologically identical. Thus, we show that establishment of PDX tumours is subject to considerable attrition and selection pressure. Despite this selection pressure 1st and 2nd generation PDXs retain the histoptahological features of the tumour of origin.