Background and objective: The subset of non-follicular non-Hodgkin's lymphoma (NHL) includes patients with varied prognoses, thus suitable for different therapeutic approaches. The International Prognostic Index (IPI), originally proposed for aggressive NHL, has been demonstrated to be of prognostic relevance also in follicular NHL. The main aim of the study was to validate the IPI in this histologic category; in addition, the specific prognostic classification, currently employed in the Gruppo Italiano per lo Studio dei Linfomi (GISL) prospective therapeutic trials and based on different features, more similar to those applied to chronic lymphocytic leukemia, was analyzed.
Design and methods: The present series consists of 137 evaluable patients affected by Working Formulation group A NHL out of 256 cases referred to the GISL Registry. The retrospective prognostic study included the evaluation by both univariate and multivariate analyses of overall survival, response to therapy and response duration. The IPI was applied as originally proposed. The GISL definition of indolent and aggressive disease at diagnosis was based on the presence of B symptoms, bulky disease, anemia and thrombocytopenia.
Results: The distribution of patients in IPI risk groups was rather unbalanced with 18%, 47%, 28% and 7% of cases classified as low (L), intermediate-low (IL), intermediate-high (IH) and high (H) risk, respectively. The median overall survival was not reached in either L or IL risk groups, and was 84.1 and 7.4 months for IH and H risk groups, respectively (p=0. 0005). A simplified IPI model was designed merging patients in both intermediate risk groups and the statistical difference of survival retained its significance. GISL prognostic stratification was demonstrated to have a significant association with survival, with a median survival of 71.3 months in aggressive disease and a median survival not reached at 152 months in indolent disease. Both the simplified IPI model and the GISL risk definition retained their significance in multivariate analysis for overall survival, while for response to therapy only the simplified IPI model resulted to be of statistical significance. In addition, the GISL prognostic stratification identified patients with different outcomes within the IPI intermediate risk group, with a median survival of 70.2 months for patients with aggressive disease wheras the median survival for those with indolent disease was not reached. Finally, a prognostic score resulting from the integration of the simplified IPI and the GISL system was statistically validated.
Interpretation and conclusions: The retrospective analysis of this series demonstrates the validity of the IPI in non-follicular indolent NHL and the usefulness of integrating the IPI parameters with disease specific prognostic variables.