Should poor risk patients with Hodgkin's disease be sorted out for intensive treatments?

Death remains a common event in patients with advanced stage Hodgkin's disease (HD) and also occurs in early stages. Identification of the population of HD patients who would best benefit from an intensification of the treatment is therefore necessary in both cases. A comparative review of prognostic criteria must be attempted through recent randomized trials using conventional treatment, independently of criteria used by the autologous bone-marrow transplant (ABMT) teams which rather aim to select relapsing patients. When freedom from progression (FFP) is considered in recent large series, the factors which most often predict for a high risk of failure are bulky mediastinum, B symptoms, stage in advanced HD and the number of nodal areas in early HD. However, prognostic data available to date cannot sort out any group of patients that would clearly require a high dose/potentially risky therapy. Because of the failure to identify consensual prognostic factors, the population considered till now for upfront intensification is restricted to a very small subgroup of patients with advanced disease (< to 5% of them), which makes the conduct of randomized trials very difficult within reasonable delays. The potential to identify and to clear part of the difficulties encountered with the prognostic analysis depends on the population studied: size of the data base, follow-up, disease stages, treatments given, outcome criteria selected. Finally, "sorted out" new patients with advanced and perhaps early stage HD may benefit from treatment intensification. As the initial selection characteristics are still poorly defined, a proposition is made to collect, eventually through the International Data Base for Hodgkin's Disease (IDHD) pertinent recent trials and to launch a survey. The results of the survey will be discussed at the Third International Symposium on Hodgkin's lymphoma in Köln (21-23 September 1995). Hazards of intensified regimens may be better quantified by then, so that newer more efficient strategies may be designed.