A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group [see comments]

T HE HISTOLOGIC categorization of lymphoma has been a source of frustration for many years for both clinicians and pathologists.In the last 10 years, much new information has become available about the lymphomas, resulting in recognition of new entities and refinement of previously recognized disease categories, raising the question of whether it is time for a new lymphoma classification.In this paper we report the result of an international review of lymphomas, which we hope may clarify some of the confusion surrounding this topic.This review was conducted at a meeting of 19 hematopathologists with particular interest and experience in lymphomas (the International Lymphoma Study Group) in Berlin, Germany, in April 1993.At previous meetings in Europe and the United States, we had come to believe that, despite the variety of classification schemes used, many hematopathologists appeared to agree on a rather large number of distinct lymphoma entities that they recognize and diagnose in daily practice.We believed that we could provide a useful service to both pathologists and clinicians struggling with the classification of lymphomas by attempting to arrive at a consensus regarding the categories of lymphoid neoplasia that can be reliably recognized at present.What emerged from this meeting was, first, that each of us had independently evolved ways of viewing these diseases that were essentially identical.Surprisingly, there was little divergence between European and US participants.Second, it was evident that, while many of these lymphoma entities are recognized in the Kiel Classification,"6 the Lukes-Collins Classification,' and the Working Formulation,* they often go by different names in different publications and may have variable criteria for diagno~is.~Furthermore, we found that many of us had doubts about both the practical feasibility and the scientific validity of distinguishing certain subtypes in these systems.We also found that while some lymphoma categories are easy to recognize, others are disturbingly prone to subjective variability.This feature of lymphoma diagnosis has not been emphasized in previous schemes for classification, which imply that all categories are equally easy for the pathologist to recognize.Ideally lymphomas, like most other tumors, should be classified according to their presumed normal counterpart, to the extent possible.This should provide the best information about disease biology, natural history, and response to treatment.However, despite extensive study, the definition of lymphoid compartments in humans and movement of cells between these compartments still contains many uncertainties.Furthermore, there are difficulties in defining the full extent of the neoplastic clone in individual cases of lymphoma, and some well-defined lymphoma types lack obvious normal counterparts.Consequently, although differentiation schemes provide useful conceptual frameworks for