Epidemiological and laboratory evidence indicate that, in addition to tobacco and alcohol, human papillomaviruses (HPV) play an important aetiological role in a subset of head and neck squamous cell carcinoma (HNSCC). To evaluate the molecular pathogenesis of HPV-infected HNSCC, we compared gene expression patterns between HPV-positive and -negative HNSCC tumours using cDNA microarrays. Tumour tissue was collected from 42 histologically confirmed HNSCC patients from an inner-city area of New York. Total DNA and RNA were extracted and purified from frozen tumour samples and gene expression levels were compared to a universal human reference RNA standard using a 27 323 cDNA microarray chip. HPV detection and genotyping were performed using an MY09/11-PCR system and RT-PCR. HPV was detected in 29% of HNSCC tumours. Most harboured only HPV16 and expressed the HPV16-E6 oncogene. HPV prevalence was highest in pharyngeal tumours (45%). Gene expression patterns that differentiated HPV-positive from negative tumours were compared by supervised classification analysis, and a multiple-gene signature was found to predict HPV16 prevalence in primary HNSCC with a false discovery rate < 0.2. Focusing on never-smokers, we further identified a distinct subset of 123 genes that were specifically dysregulated in HPV16-positive HNSCC. Overexpressed genes in HPV-positive HNSCC tumours included the retinoblastoma-binding protein (p18), replication factor-C gene, and an E2F-dimerization partner transcription factor (TFDP2) that have also been found to be overexpressed in cervical cancer. An additional subset of genes involved in viral defence and immune response, including interleukins and interferon-induced proteins, was found to be down-regulated in HPV-positive tumours, supporting a characteristic and unique transcriptional profile in HPV-induced HNSCC. Copyright © 2007 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.