Population-based incidence and survival of central nervous system (CNS) malignancies in Girona (Spain) 1994-2005

The purpose of this study was to describe the incidence and survival of primary Central Nervous System (CNS) malignancies using data from the population-based cancer registry for Girona province (north-east Spain). We included all cases of primary CNS malignancies registered between 1994 and 2005. Pathological diagnoses were reviewed and grouped according to the 2007 WHO Classification. Meningeal, soft tissue tumours, spinal cord tumours and primary CNS lymphoma were not included. Cases notified only by death certificate were excluded from the survival analysis. Kaplan and Meier survival curves were calculated from date of diagnosis to death or end of study (31 December 2005), as was relative survival. A total of 493 new CNS cancer patients were registered during the study period: 49.3% astrocytic, 3.4% oligodendroglial and oligoastrocytic tumours, 2.6% ependimal tumours, 3.7% embryonal tumours, 0.2% choroid plexus and 41% without histological confirmation. The mean age (in years) for embryonal tumours was 18.17 years, these being the younger patients in the sample, and 66.34 years for those without histological confirmation, the older patients. Overall, the age standardised incidence rate was 5.88 cases/100,000 people/year (men = 6.81; women = 4.99) with an increasing trend by age until the 70-74 age group. Five-year observed survival rates were: 14.6% for astrocytic tumours, 35.7% for oligodendroglial and oligoastrocytic tumours, 41.0% for ependymal tumours, 32.4% for embryonal tumours and 7.5% for those without histological confirmation (log rank test: P < 0.001). Five-year observed survival rates for astrocytic tumours were analyzed separately by tumour grading, with 37% for diffuse astrocytoma, 7.1% for anaplastic astrocytoma and 4.7% for glioblastoma (log rank test: P < 0.001). Our results show that astrocytic tumours are most frequently diagnosed and glioblastoma patients have the worst survival figures for the area covered by our population cancer registry. The high observed incidence of histologically unverified tumours is most probably due to easy access to state of the art CNS imaging in our area.