Diagnostic and prognostic markers in primarily non-smoking related head and neck cancer

Abstract

Head and neck cancer (HNC) includes cancers of the oral cavity, the pharynx (i.e., the nasopharynx, the oropharynx, and the hypopharynx), the larynx, the nasal cavity, the paranasal sinuses, and the salivary glands. Traditional risk factors are smoking, alcohol, opium, betel chewing, and virus infections, such as human papillomaviruses (HPV) and Epstein-Barr Virus (EBV). These viruses play a major role in some cancers, and it is evident that the aetiology of different HNC types differs. In Sweden, HNC risk profiles have changed in the past decades due to a decrease of smoking in both men and women, while an increase in HPV-related cases has been noted. This increase is anticipated to continue for some decades, until the introduction of the HPV vaccine will hopefully prevent most of them. This thesis has therefore focused on HNC primarily not associated with smoking. It includes studies on oropharyngeal squamous cell carcinoma (OPSCC) and HPV-related multiphenotypic sinonasal carcinoma (HMSC) in a broader context. Finally, we also studied adenoid cystic carcinoma (AdCC), where the aetiology is mainly unknown, and the diagnostics are still very challenging.

In Paper I we investigated the relationship between the presence of HPV DNA and p16INK4a (p16) overexpression and prognosis of different OPSCC subsites. We found that the presence of HPV DNA and p16 overexpression were favourable prognostic markers in the tonsillar and base of tongue cancer subsites (TSCC and BOTSCC, respectively) but not in other OPSCC cancer subsites (otherOPSCC). We also showed the importance of testing for both HPV DNA and p16 expression status for better prognostication.

In Paper II we investigated a possible prognostic role of psoriasin expression, examined by immunohistochemistry in base of tongue cancer. In this pilot study we could show that low psoriasin expression was a favourable prognostic marker in HPV-positive (HPV+) BOTSCC.

Paper III includes a systematic literature review of studies on HMSC and the presence of different HPV types in these tumours and their various locations. The data indicated that HPV+ tumours with such characteristics may not only be located within the sinonasal region.

In Paper IV we initially investigated whether the presence of HPV and human polyomaviruses (HPyVs) played a role in the prognosis of adenoid cystic carcinoma (AdCC) where the aetiology is still mainly unknown. In addition, we wanted to examine whether the presence of these viruses could play a diagnostic role. HPyVs did however not have a major role in the aetiology of AdCC, as no case was positive for HPyV. Of 68 patients analysed, there were three HPV+ AdCC cases, but upon re-examination their pathology was more similar to HMSC. This suggested that HMSC may not be limited to the sinonasal region. Furthermore, our findings indicated that the presence of HPV could be used in diagnostics when distinguishing between AdCC and HMSC.

In Paper V we investigated a large cohort of AdCC patients (155 cases) regarding clinical presentation, treatment, and survival. We found that subsite (major salivary glands), early stage (stage I-II), and multimodal treatment were positive prognostic factors, while age, gender, perineural growth, or negative surgical margins did not influence clinical outcome.

In conclusion we were able to show that the presence of HPV DNA and p16 overexpression were favourable prognostic markers in TSCC and BOTSCC but not in otherOPSCC. In addition, low psoriasin expression was found to be a positive prognostic marker for HPV+ BOTSCC. In the systematic literature review of HMSC we disclosed that this tumour entity could also arise at sites outside the sinonasal area. Furthermore, we could show that neither HPyVs nor HPV played a major role in AdCC, but that presence of HPV could be of differential diagnostic value, especially in the sinonasal area. The final paper describing a large AdCC patient cohort confirmed that commonly used prognostic factors, e.g., gender, age, and smoking history did not correlate with survival, notably in this study neither did perineural invasion nor radical surgery of the tumour primary.