Molecular Characterisation of Diffuse Large B Cell Lymphoma in Patients of 80 Years Old or More: Clinical Relevance in a Multicentric Randomized Phase III Study of the Lysa (SENIOR Study)

Introduction

The outcome of very elderly patients with newly diagnosed diffuse large B cell lymphoma (DLBCL) remains poorer than that of younger patients, often because of co-morbidities and physiological organ function impairment. However, the biological specificities of DLBCL in very elderly patients are unknown. The LYSA conducted a multicentric, phase III, open-label, randomized trial in newly diagnosed DLBCL > 80y evaluating the efficacy of the combination of lenalidomide with R-miniCHOP (R2-miniCHOP) in comparison to the standard R-miniCHOP (SENIOR trial, NCT02128061). Based on targeted NGS sequencing and gene expression profile (GEP) analysis, the BIOSENIOR project aimed to provide a molecular characterisation of the cohort and detect clinically relevant biomarkers in according to treatment arm.

Methods

The SENIOR study enrolled patients aged of 80 years or more with non-previously treated CD20+ DLBCL, age-adjusted IPI= 0 to 3, and Ann Arbor stage II to IV. 249 patients were randomized (127 in R-miniCHOP arm and 122 in R2-miniCHOP arm). Median age was 83y (range 80-96. Patients with available tumor DNA /RNA from FFPE biopsies were included in the BIOSENIOR substudy. GEP was performed using Lymph2CX assay (Nanostring®) and a new RTMLPA assay (Bobée et al. J Mol Diagn 2017) able to quantify the expression of 137 relevant genes involved in B-cell biology, microenvironment or therapeutic response (cereblon). Genotyping (QiaSeq ®) was performed using a dedicated 36 genes panel. GEP features were compared to DLBCL patients < 60y enrolled in the LYSA GAINED trial (NCT01659099).

Results

Tumor DNA/RNA was available for 164/249 patients (66%). GEP was performed in 154 cases by RTMLPA and in 103 cases by Lymph2CX. The concordance to classify GCB/ABC was of 94% between the two technologies. DLBCL were mainly classified in the ABC subtype [77 = ABC (50%); 53 = GCB (34.4%); 17 unclassified (11%)]. Four cases (2.6%) were classified as EBV+ DLBCL and 1 as PMBL. 81 cases (52.5%) were considered as BCL2+, 23 (15%) as MYC+ and 16 (10.4%) as double-expressors (DE) by GEP. Comparison of GEP with younger patients highlighted distinct features and geriatric specificities (figure 1A). Genotypes were obtained in 159 patients. MYD88, PIM1, CREBBP, CD79B and TP53 were the five most frequently mutated genes (Figure 1B). Correlations between biomarkers and outcomes (OS and PFS) were performed in the overall population and according to treatment arms.

The GCB/ABC subtyping, as determined by Lymph2CX (n = 97) was related to OS [HR = 0.38 (0.18 - 0.81) Log-rank: 0.009] and PFS in the overall population. This result was not confirmed in the extended cohort (n = 154) by RTMLPA. The MYC / BCL2 DE status assessed by GEP was a strong predictor of outcome (OS, HR = 2.89 (1.39 - 5.97) Log-rank: 0.003] in both arms. Importantly MYC/BCL2 DE was independent of the IPI score and albuminemia in multivariate analysis. In an exploratory analysis, cereblon (HR = 2.360; CI-95% 1.157-4.812, Log-rank: 0.015) and PD1 [HR = 0.33 (0.16 - 0.66) Log-rank: 0.001] expressions correlate exclusively with PFS in the R2-miniCHOP arm.

MYD88 mutations (L265P and others variants) were associated with a worse outcome in the overall population (OS, HR= 2.05, CI-95% 1.190-3.554, Log-rank: 0.008; PFS, HR =1.731, CI-95% 1.059-2.830 Log-rank: 0.027). FOXO1 mutations were detected in 15 patients (9%). Its impact was only observed in the R-miniCHOP arm (OS, HR = 4.07 (1.52 - 10.87) Log-rank: 0.0024; PFS HR = 4.139 (1.59 - 10.79) Log-rank: 0.0016)] and was erased in the R2-miniCHOP arm. TP53 mutations, reported in 24% of cases were predictive of the PFS in the overall population [HR = 1.811 (1.09 - 3.01) Log-rank: 0.0199](prognostic factors are summarized in Table 1).

Conclusion

BIOSENIOR is an ancillary study of the first prospective phase III trial in ³80 years old patients with newly diagnosed DLBCL. Some biomarkers, known to be predictive in younger patients, are relevant in this cohort, including MYC expression, BCL2/MYC DE, mutations of MYD88, TP53 or FOXO1. However, GEP characteristics of DLBCL occurring in this very old population are highly different of younger patients, suggesting distinct underlying oncogenic mechanisms. BIOSENIOR indicates that, beyond geriatric frailty, molecular features of DLBCL patients > 80y play also a crucial role that should be considered in the design of future trials. Markers that may predict R2-miniCHOP efficacy remain to be confirmed.

View largeDownload slide

View largeDownload slide

Close modal