A comparison of treatment results for recurrent malignant gliomas.

Retreatment of malignant gliomas may be performed with palliative intent after careful consideration of the risks and benefits, and with special regards to iatrogenic neurotoxicity and quality of life (QOL). This review compares studies of several retreatment strategies (published between 1987 and 2000) based on the quality of their evidence. Depending on both established prognostic factors and previous treatment, individually tailored retreatment strategies are possible. In all studies that included a multivariate analysis of prognostic factors, performance status was the most important. So far, predictive factors for response, which might facilitate patient selection, have not been unequivocally defined. In terms of QOL, single-agent chemotherapy (temozolomide, nitrosoureas, platinum and taxane derivatives) may offer a better therapeutic ratio than polychemotherapy. For glioblastoma multiforme, progression-free survival and QOL were more favourable after temozolomide than procarbazine (level 1 evidence). The survival of patients after various radiotherapy techniques is broadly similar. However, considerable toxicity is associated with radiosurgery or brachytherapy. Fractionated stereotactic radiotherapy plus radio-sensitizing cytostatic agents has shown promising initial results in small groups of selected patients and awaits further evaluation. Level 2 evidence derived from non-randomized studies does not suggest a substantial prolongation of survival by re-resection as compared with chemotherapy or radiotherapy alone. Level 1 evidence derived from a randomized trial suggests that application of BCNU polymers significantly improves the outcome after re-resection. However, most studies reported median survival in the range of only 25-35 weeks, thereby emphasizing the need for the development and clinical evaluation of new innovative treatment approaches. Copyright 2000 Harcourt Publishers Ltd.