Validation of the PRIMA-Prognostic Index for Patients Treated with Rituximab Plus Chemotherapy and Refinement of Prognostic Parameters for Patients on Rituximab Plus Lenalidomide in the Phase III Relevance Trial

Background: Patients with follicular lymphoma (FL) have a heterogenous prognosis. Recently a simple score, the PRIMA-PI, was developed based on the PRIMA clinical trial (Bachy et al Blood. 2018). With only two factors (beta-2-microglobulin level > 3 mg/L and bone marrow involvement), this index was at least as discriminatory as FLIPI on the training and validation cohorts. The validity of the PRIMA-PI was confirmed on Czech and German FL cohorts and a patient group from the Nordic Lymphoma Group. However, further validation is needed to confirm the use of PRIMA-PI in place of FLIPI for prognostic assessment. Indeed, in the era of new chemo-free treatments, it seems important to challenge the potency of traditional prognostic factors and scores. Recently, rituximab combined with lenalidomide (R2) was compared to conventional immunochemotherapy (R-chemo) in the phase III RELEVANCE trial. The aim of our study was to validate PRIMA-PI in the RELEVANCE trial cohort and compare its performance with FLIPI (Solal-Celigny et al. Blood. 2004) and FLIPI2 (Federico et al. JCO. 2009). A secondary objective was to evaluate potential differences in terms of prognostic bio-clinical parameters between the R2 and R-chemo arms. Methods: All patients with available data for FLIPI, FLIPI2, and PRIMA-PI from the intention to treat population of the RELEVANCE study were included in the analysis. PFS according to each prognostic score were assessed in the total population and by treatment arms. Data were not mature enough to compare OS distributions. Performance metrics (log-rank p value and Net Reclassification Improvement [NRI]) were calculated for each group to assess concordance and discriminating ability of each score. Results: Median follow-up time for the study was 38 months. Overall, 846 RELEVANCE patients were included in the analysis. Data were available for 845 patients for FLIPI score assessment, 832 for FLIPI2 and 807 for PRIMA-PI. Group repartition according to the FLIPI and the FLIPI2 were largely imbalanced compared with PRIMA-PI. FLIPI classified very few patients in the low risk group (15% LR) while 49% of the patients were at high risk (HR), and 36% were at intermediate risk (IR). Similarly, FLIPI2 risk categories were as follow: 8% LR, 50% IR, and 42% HR. On the contrary, PRIMA-PI divided the study population into three equal groups (33%, 33% and 34%). In the total population, FLIPI and PRIMA-PI were predictive of PFS (p=0.029 and p=0.004, respectively); FLIPI