[Clinical trials in oncology: specific methodology problems].

Clinical benefit in oncology patients can only be derived from coordinated and successive experimental trials. The following consensus thus emerged from the working party: The assessment of tolerance during clinical trials requires the actuarial rate of acute adverse events (AE) to be registered and the development of epidemiological structures involved in long-term assessment of AE. Experimental trials have to be assessed according to descriptive epidemiological data (generated in France and Europe) and completed with observational studies on post-marketing medical practice. The principle of an a priori authorization given by the Afssaps for gene and cell therapy trials was legalised in October 2001. The working group spoke strongly in favour of authorizing early trials only validating the process, and for reasonable objectives in terms of viral security assessment of annex therapeutic products. With regard to the development of new antiproliferative agents, the active dose (or optimal biological dose) replaces the maximal tolerable dose. Early efficacy criteria could be biological and specific. With cumulative subacute rather than acute toxicity, these agents can not be properly evaluated with a short term benefit/risk ratio as is done for cytotoxic drugs. Elderly people constitute a fragile, heterogeneous and increasing population, in whom adequate assessment of anticancer agents is mandatory. It is important to promote controlled clinical trials in children and to systematically monitor them over a very long period of time. Only rare tumours may be considered as orphan situations. Finally, a list of all clinical trials conducted in oncology should be made by the Afssaps, with the help of the FNLCC (Fédération Nationale des Centres de Lutte Contre le Cancer) who have already begun this work.