Evaluation of three prognostic indexes in follicular lymphoma

Abstract

6595 Background: Three validated prognostic models have been proposed for use in patients with follicular lymphoma (FL), the International Non-Hodgkin’s Lymphoma Prognostic Index (IPI), the ILI model from the Italian intergroup, and recently the FLIPI model from the French International Lymphoma Prognostic Index. The aim of this study was to apply these three models to a group of FL patients to determine the concordance between them and their usefulness in predicting clinical outcome, including rate of transformation. Methods: The IPI index (age >60 years, extranodal involvement ≥2 sites, elevated lactate dehydrogenase (LDH), ECOG performance status ≥2, stage ≥3), the ILI index (age >60, extranodal involvement ≥ 2 sites, elevated LDH, male sex, B symptoms and erythrocyte sedimentation rate ≥ 30) and the FLIPI index (age > 60, stage ≥ 3, elevated LDH, number of nodal sites > 4, hemoglobin < 120 g/L), respectively were calculated and applied to 100 patients diagnosed with FL (grade I-III) between 1994 and 2001. Overall survival (OS) and progression free survival (PFS) were calculated for each prognostic group (low, intermediate and high risk) by the Kaplan-Meier method. Results: The median follow-up of our series was 31 months (mo) (range 1–340). Median OS was 24 mo (range 1–226). The median PFS was 41 mo (range 4–145). Concordance between the three indexes was 40%. The index that best predicted mortality was the ILI (p=0.027). The distribution of patients and mean overall survival according to risk group were similar for each index. Patients who went on to transform to aggressive NHL were rarely classified as high risk at diagnosis of FL by any index. Conclusions: Although the concordance between the three models was low, each is useful for classifying FL patients into risk groups and in predicting outcome in this group of patients. Transformation was not well predicted by any of the models. This information will be combined with cytogenetic, immunohistochemical, and DNA microarray data from frozen tumor tissue (currently proceeding) to develop a combined molecular and clinical prognostic index. A combined index may help select appropriate treatment for individual patients. No significant financial relationships to disclose.