Background: head and neck cancers are a heterogeneous group of malignancies, affecting various sites and subsites, with differing prognoses. The aim of this study was to analyse survival for European head and neck cancer patients in populations covered by population-based cancer registries (CRs), in relation to tumour subsite as prognostic factor.
Patients and methods: we analysed 51 912 adult head and neck cancer cases (36 322 mouth-pharynx and 15 590 larynx) diagnosed from 1995 to 1999 and archived by 45 CRs in 20 countries participating in EUROCARE-4. Five-year age-standardised relative survival was estimated for mouth-pharynx and larynx sites by sex and country. Relative survival was modelled to provide estimates of relative excess risks (RERs) of death by country, adjusted for confounding factors.
Results: a large but site-variable proportion of tumours were incompletely specified. Five-year age-standardised relative survival was low in Slovakia and high in The Netherlands. Adjustment for subsite reduced RERs of death for most countries; 5-year relative survival increased from 1990-1994 to 1995-1999 for all subsites, while between-country differences in survival narrowed.
Conclusion: differences in subsite distribution explain a considerable part of the survival differences for head and neck cancers, however, incomplete/inaccurate subsite reporting complicate interpretation.