Defining ultra‐high‐risk dlbcl patients prior to initial treatment based on an integrative host and disease prognostic score (from remarc study)

Background: The risk stratification of patients (pts) with diffuse large B-cell lymphoma (DLBCL) is based on the widely used IPI. Three of the IPI variables (LDH, stage and number of EN sites) may be refined in two metabolic measures obtained by 18FDG-PET, the total metabolic tumor volume (TMTV) and the distance between 2 lesions (SDmax). We recently demonstrated that the adipose tissue alteration measured by the lumbar subcutaneous adipose density (LSAD) (i.e an increased density) is an independent prognostic parameter in DLBCL. The aim was to analyze if these combined parameters characterizing the host (ECOG PS) and the disease (TMTV, SDmax, LSAD) could help to better identify patients at high risk of relapse, prior to initial treatment. Patients and Methods: Pts were included in the REMARC study (NCT01122472), (DLBCL >60 to 80 years old, responder to R-CHOP randomized between lenalidomide or placebo). TMTV, SDmax, LSAD were measured on the baseline PET/CT performed before treatment. Statistical methods included multiple Cox regression for developing the MVE2D score. Results: 273 pts were analyzed. 195 (71%) pts were scored high-risk (3-5) IPI, and 155 (57%) high-intermediate-risk NCCN IPI and 36 (13%) high-risk NCCN IPI. ECOG PS and TMTV, SDmax, LSAD were identified as the 4 independent prognostic factors for PFS and OS (Table 1). Accordingly, a score Metabolic Volume ECOG Distance Density (MVED2) was calculated as 0.57 (if ECOG PS ≥2) + 0.76 (if SDMAX > 32) + 0.70 (if LSAD > −90) + 0.59 (if TMTV>220). According to the MVED2 score, pts were classified into low risk (65% of pts, median PFS and OS of 68 months and not reached, respectively), intermediate risk (25% of pts, median PFS and OS of 60 months and not reached, respectively) and high risk (10% of pts, median PFS and OS of 16 and 44 months, respectively) (Fig 1). The MVED2 score has a significant impact on PFS (p < 0.001) and OS (p < 0.001) (Figure 1). More extra-nodal sites (89.3% >1 vs 47.3%, p < 0.001), more high-IPI 3-5 (96.4% vs 68.6%, p = 0.001), more high-NCCN IPI (35.7% vs 10.6%, p < 0.001), more ABC profile (46.4% vs 20.8%, p = 0.033), more involved bone marrow biopsy (42.9% vs 14.7%, p < 0.001) and male sex (78.6% vs 58%, p = 0.041) were observed in high risk patients than in the low- and intermediate-risk patients. Conclusions: The MVED2 is the first prognostic index based on new metrics measured on 18FDG-PET and ECOG PS that may help to discriminate ultra high-risk DLBCL pts, even responder to R-CHOP. Progreesion-free survival and overall survival according to the IPI, the NCCN IPI and the MVED2 score Keywords: Diagnostic and Prognostic Biomarkers Conflicts of interests pertinent to the abstract C. Thieblemont Consultant or advisory role: BMS, Novartis, Kyte/Gilead, Roche, Incyte Honoraria: BMS, Novartis, Kyte/Gilead, Roche, Incyte