Application of relevant data provided by molecular biology to the treatment of cancers is a major challenge in medical oncology. In this context, it is a widely held belief that, in the years to come, these data will allow treatment tailored to each individual patient's tumour. However, the results of the present study, which, to our knowledge, is the largest prospective study ever performed to evaluate the prognostic value of biological markers in adult gliomas, tend to suggest that these markers will be of limited value for individual treatment decisions, particularly for the most heterogeneous, that is, high-grade, tumours. If tumour markers were to have a role in treatment decisions in adult gliomas, it would probably be limited to less heterogeneous, that is, low-grade, tumours, as illustrated by the study of certain p73 isoforms. The absence of correlation between LOH and OS of low-grade tumours observed in this study could be due to lower genetic instability, whereas the absence of correlation with gene expressions could be due to the fact that the genes studied here are involved in later stages of gliomagenesis.
MGMT has a special place in this context, as, although it has a poor specific prognostic value, it can be used to predict the chemosensitivity of these tumours. Although it is recommended to include MGMT in clinical trials of chemotherapy, it cannot be recommended as part of everyday clinical practice. For all of these reasons, it would, therefore, be unrealistic to expect any contribution of these markers to treatment decisions for adult patients with glioma.
When this study was initiated, tumour stem cell and pluripotent stem cell biology was only in the early stages. A growing body of literature has demonstrated the importance of these stem cells to propose an explanatory model for gliomagenesis that could possibly open the way to major therapeutic progress, as these cells constitute the genuine target of treatments. Preclinical animal models reproducing all phases of gliomas could also be used to develop a similar approach to that conducted in this study, based on whole tumours and not stem cells, which can now be specifically isolated and studied. Progress in tumour stem cell (TSC) biology has led to identification of cells possessing the specific characteristics of true glioma stem cells, specific markers (e.g. CD133), capacity for self-renewal and capacity for differentiation into the various cell lines of brain tissue (neuronal, glial) (Gilbertson and Rich, 2007
). These results have led to the development of a gliomagenesis model directly resulting in GBM, which postulates the malignant transformation of these cells at the time of ‘insults', which selectively affect the genes involved in the initial steps of carcinogenesis. Access to these tumour stem cells has now been defined by tumour culture protocols, which will allow molecular studies on much more homogeneous populations. This type of approach applied to individual tumors should result in much more effective use of markers currently identified in heterogeneous cohorts. At the same time, this type of approach would also allow a more accurate definition of the limits of treatments (e.g. resistance of TSC to radiotherapy and acquisition of resistance to chemotherapy) (Rich, 2007