RAPID COMMUNICATION A Clinical Evaluation of the International Lymphoma Study Group Classification of Non-Hodgkin’s Lymphoma By The Non-Hodgkin’s Lymphoma Classification Project

The recognition of several new types of non-Hodgkin’s nosis of NHL was confirmed in 1,378 (98.2%) of the cases. The most common lymphoma types were diffuse large Blymphoma (NHL) in recent years has led to proposals for changing lymphoma classifications, including a new procell lymphoma (31%) and follicular lymphoma (22%), whereas the new entities comprised 21% of the cases. Diagposal put forth by the International Lymphoma Study Group (ILSG). However, the clinical significance of the new entities nostic accuracy was at least 85% for most of the major lymphoma types, and reproducibility of the diagnosis was and the practical utility of this new proposal have not been studied. Therefore, we performed a clinical evaluation of the 85%. Immunophenotyping improved the diagnostic accuracy by 10% to 45% for a number of the major types. The ILSG classification. A cohort of 1,403 cases of NHL was organized at nine study sites around the world and consisted of clinical features of the new entities were distinctive. Both the histologic types and the patient characteristics as deconsecutive patients seen between 1988 and 1990 who were previously untreated. A detailed protocol for histologic and fined by the International Prognostic Index predicted for patient survival. In conclusion we found that the ILSG classificlinical analysis was followed at each site, and immunologic characterization as to Tor B-cell phenotype was required. cation can be readily applied and identifies clinically distinctive types of NHL. However, for clinical application, Five expert hematopathologists visited the sites and each classified each case using the ILSG classification. A consenprognostic factors as defined by the International Prognostic Index must be combined with the histologic diagnosis for sus diagnosis was also reached in each case, and each expert rereviewed a 20% random sample of the cases. Clinical correappropriate clinical decisions. q 1997 by The American Society of Hematology. lations and survival analyses were then performed. A diag-