Sub-classification of low-grade cerebellar astrocytoma: is it clinically meaningful?

The objectives were to identify prognostic factors for the survival of children with cerebellar astrocytoma, and to evaluate the reproducibility and prognostic value of histological sub-classification and grading.

Children aged 0-14 years treated in Denmark for a cerebellar astrocytoma in the period 1960-1984 were included and followed until January 2001 or until their death. The histological specimens from each patient were reviewed for revised grading and classification according to three different classification schemes: the WHO, the Kernohan and the Daumas-Duport grading systems.

The overall survival rate was 81% after a follow-up time of 15-40 years. The significant positive prognostic factors for survival were "surgically gross-total removal" of the tumour at surgery and location of the tumour in the cerebellum proper as opposed to location in the fourth ventricle. No difference in survival time was demonstrated when we compared pilocytic astrocytoma and fibrillary astrocytoma. Moreover, we found that the Kernohan and the WHO classification systems had no predictive value and that the Daumas-Duport system is unsuitable as a prognostic tool for low-grade posterior fossa astrocytomas.

Discordant observations due to interobserver variability make histological sub-classification of low-grade cerebellar astrocytomas in children insufficient for predicting prognosis and biological behaviour. Similar survival rates in a population of paediatric low-grade cerebellar astrocytomas of grades I and II indicate that tumour grade has no prognostic significance within this group of patients. "Surgically gross-total removal", especially if the tumour is located in the fourth ventricle is of the highest importance for long-term survival. Histological sub-classification of the tumours has no predictive value.