Concurrent chemoradiotherapy for malignant glioma: dead end or highway?

powerful impact on survival than most treatments [7]. On the other hand, the study presented here has important limitations and several of the investigator’s statements are quite disputable. To start with, and strictly speaking, the authors are not in a position to conclude that ‘there was no increase in overall survival with weekly paclitaxel’. The study was not powered sufficiently to demonstrate a difference, and even less powered to demonstrate the absence of difference between the two treatment arms due to the very small number of patients in each subset. Nevertheless, and assuming that there was indeed no effect of paclitaxel with RT, one can speculate at length about the reasons why the combination did not work. For example the blood-brain-barrier issue, including the authors’ proposal to modulate it, may have little relevance in HGG, especially with RT which can in itself alter the bloodbrain-barrier. There are most likely several other ‘culprits’ for the lack of efficacy of the drug or its combination with RT, such as ‘classical’ resistance factors [8] (i.e. tumour proliferation, pH, hypoxia etc.) or more recently described ones [9, 10] (i.e. RT-induced invasion via integrins, MGMT expression, EGFR, VEGF, PDGF expression etc.), all of which can play a critical role in HGG resistance. In order to better identify the various causes of failure of chemotherapy and combined chemoradiotherapy, and in the hope of overcoming these, it is now essential to systematically perform translational studies in parallel with any new clinical trial in GBM. In this regard the demonstration by the EORTC/NCI-C TMZ trial that the MGMT (methylguanine-DNA-methyltransferase) status has a prognostic and predictive value [11] will enable us to define new treatment strategies depending on this particular tumour state. Another explanation of the discrepancy between the in vitro and in vivo data may be the lack of daily RT exposure to paclitaxel simply because the drug had to be given weekly. Thus, one cannot really speak here of a true concurrent chemoradiotherapy scheme, contrary to the EORTC/NCI-C Malignant or high-grade glioma (HGG) are the most common and devastating of all adult brain tumours. Their incidence has steadily increased over the years, especially in patients in their 5th decade and over [1]. Surgery and radiotherapy (RT) have been the mainstay of treatment for more than 40 years [2], while before the landmark EORTC/NCI-C trial on temozolomide (TMZ) [3], chemotherapy – which was mainly based on nitrosoureas – yielded only a marginal improvement in survival, if any [4]. Amongst many drugs, paclitaxel was also tested in conjunction with RT in HGG, based on in vitro data demonstrating that the drug caused a block in G2/M, a more radiosensitive phase of the cell cycle [5]. In this issue of ONKOLOGIE, Montemor and colleagues present a randomized study conducted between 1998 and 2002, in which 61 patients with either glioblastoma (GBM) or anaplastic astrocytoma (AA) were randomized to receive radiotherapy with or without weekly paclitaxel [6]. The results indicate that the combined regimen was well tolerated, but independently of tumour grade (GBM or AA), there was apparently no statistical difference between the two treatment groups. The authors conclude that radiochemotherapy with paclitaxel is safe but that there is no increase in overall survival. They formulate the hypothesis that the difference between the in vitro and in vivo (or clinical) radiosensitising effect of paclitaxel in malignant glioma is likely to be due to the lack of permeability of the blood-brain barrier to this drug. Later in their discussion, and alluding to the results of the EORTC/NCI-C TMZ trial, they state that the survival benefit with TMZ being 2.5 months, other treatments should be investigated. On one hand, Montemor and colleagues should be commended for having tested paclitaxel and RT in a randomized trial, given the fact that the literature on HGG is replete with phase II trials testing various drugs and RT with no control arm. These studies are of course hard to interpret because especially in GBM, pre-therapeutic prognostic factors have a more