Prognostic factors in low grade (WHO grade II) gliomas of the cerebral hemispheres: the role of surgery.

OBJECTIVE To assess the role of surgery on survival of patients with grade II gliomas of the cerebral hemispheres. METHODS One hundred and thirty one low grade hemispheric gliomas surgically treated (biopsied patients excluded) between 1978 and 1989 were retrospectively reviewed. Thalamic, basal ganglia, callosal, or ventricular location were not considered. All tumours were World Health Organisation (WHO) grade II gliomas: 42 fibrillary and 11 gemistocytic astrocytomas, 49 oligodendrogliomas, and 29 oligoastrocytomas. Patients' ages ranged from 14 to 63 (mean 32.9, median 34) years, Karnofsky performance from 0.50 to 0.90 (mean 80.7, median 80), and postsurgical follow up of the living patients from 24 to 190 (mean 97.02, median 93) months. Postoperative external radiotherapy was performed in 49 cases. RESULTS The overall survival probability at five years was 97.1%, at eight years 76.1%, and at 10 years 62.7% (median survival time 144 months). The impact on survival of the following variables was analysed: age (< 20, 21-40, and > 40 years), Karnofsky score (80-100, 70 < or = 70), histology, tumour extension (T1 < 3 cm, T2 3-5 cm, T3 > 5 cm maximum diameter), extent of surgical resection (S1 radical, S2 subtotal < 10% residual tumour, S3 partial-10%-50% residual tumour), and radiotherapy (either performed or not). A significant positive association with survival at univariate analysis was found for the age group < 20 years (P = 0.003), for total and subtotal surgical resections (S1 and S2; P < 0.001) and for the non-irradiated patients (P = 0.0049), whereas a shorter survival probability was noticed for gemistocytic astrocytomas (P < 0.001) and for tumour extension > 5 cm (T3; P = 0.0193). Karnofsky performance did not show any significant association with survival. The most relevant factor affecting survival at the multivariate analysis was the extent of surgical resection, which resulted as the only variable retaining a significant value (P = 0.001, risk factor = 2.20). CONCLUSIONS The data strongly support the role of a surgical removal as extensive as possible in the treatment of these tumours.