Molecular analyses and prognostic relevance of HPV in head and neck tumours

Abstract

HPV involvement in head and neck (HN) cancer is still under active investigation. Fresh frozen and archival clinical samples from 115 patients affected by HN carcinomas were analysed by PCR-based methods and direct sequencing. HPV types, intra-type variants, physical status, viral load and viral transcript presence were determined. HPV positivity was correlated with the main clinical-pathological features, including smoker and drinker status, and the clinical outcome. Twenty-one tumours were HPV positive (18.3%) with HPV16 being the most frequent type (n=14) followed by HPV6 (n=4), HPV33, HPV35, and HPV58 (n=1, each type). Tonsil carcinomas contained more high-risk HPV types (6/8; 75%) than all other sites (p=0.0004). HPV16 genome was integrated in all analysed tumours, as pure integrated form or mixed with concomitant episomal forms (4 cases). The viral load showed a wide variability (range, 0.7-485 copies per cell) with the highest value detected in a larynx tumour and the lowest one in a case of cancer of the oral cavity. In 9 HPV-positive samples where mRNA was available, transcripts of viral early oncogenes originating by integrated, episomal or mixed forms of the viral genome were found. A statistically significant correlation was evidenced between HPV and tumour differentiation, being the virus more associated with tumour grade G3/G4. Multivariate Cox regression analysis revealed that lymph-node and grade status were significant independent factors for a worse disease-free survival and overall survival, whereas the HPV status was associated with a better overall survival (OR, 0.33; 95% CI, 0.13-0.81; p=0.01). Taken together these results indicate that distinct pathological mechanisms for the malignant transformation in each single HN subsite should be taken in account; HPV molecular analyses should be considered a valid tool to distinguish subsets of oropharyngeal tumours and HPV presence could be useful for the prognostic assessment of HNSCC.