DIFFERENTIAL EFFICACY OF BORTEZOMIB IN SUBTYPES OF DIFFUSE LARGE B‐CELL LYMPHOMA (DLBL): a PROSPECTIVE RANDOMISED STUDY STRATIFIED BY TRANSCRIPTOME PROFILING: REMODL‐B

Abstract

Background: DLBL subtypes identified by patterns of gene expression correspond to germinal center (GCB) or activated (ABC) B‐cells like. The latter demonstrate dysregulation of the NF‐KB pathway. Outcomes of treatment with R‐CHOP are inferior for ABC DLBL in retrospective series. This study investigated whether adding bortezomib (B), a putative NF‐KB inhibitor, can reverse this phenotype. Methods: The REMoDL‐B study was a collaboration between the UK NCRI group and the SAKK. Patients (pts) newly diagnosed with DLBL commenced conventional R‐CHOP. During the first cycle, whole transcriptome expression profiling (GEP) was performed on formalin‐fixed paraffin‐embedded tissue by Illumina DASL array. Pts with successful GEP were randomised (1:1) to continue R‐CHOP +/‐ bortezomib (1.3 mg/m2 IV or 1.6 mg/m2 SC) days 1 + 8 for cycles 2‐6. The primary endpoint was progression‐free survival for the GCB + ABC population; the study was powered for a 10% benefit in PFS from bortezomib. Results: Between 6/2011 and 5/2015, 1076 eligible pts were registered. Median age was 64 yrs (20‐86); Stage I 2.9%; II 27.5%; III 30.8% and IV 38.8%. The distribution of IPI scores: Low, 25.7%; low int 25.9%; high int. 30.6%, high 17.8%. There was no difference in baseline demographics between arms. Cell of origin results were: GCB n = 475 (44.1%); ABC n = 244 (22.7%); unclassified (U) n = 199 (18.5%); no profile n = 158 (14.6%). Mutational frequency of MYD88, PRDM1, CD79B was higher in ABC, whilst mutations in CREBBP, EZH2, DDX3X, FAS and KMT2D were more frequent in GCB. ABC pts were older (median age ABC 67 yrs; GCB 63 yrs; U 63 yrs; P < 0.005). Bulk was more common in the GCB (GCB 33.8%; ABC 20.7%; U 27.8%; P < 0.001). RB‐CHOP was not associated with increased haematological toxicity; Grade ≥ 2 neuropathy occurred in 20.7% RB‐CHOP vs 12.5% R‐CHOP pts. There was no difference in PFS in the combined GCB + ABC population between RB‐CHOP and R‐CHOP; HR = 0.84; P = 0.225. PFS at 30 months (PFS30) was 74.3% and 70.1% respectively. Bortezomib did not significantly affect PFS in either the GCB pts HR = 0.87; P = 0.458 (PFS30 75.8% vs 72.9%) or ABC pts HR = 0.79; P = 0.309 (PFS30 71.5% vs 64.7%). However, pts with low IPI had a significantly better PFS when bortezomib was added to R‐CHOP, HR = 0.37; P = 0.012. This benefit was seen only in the ABC group. There was no difference in overall survival between arms HR = 0.85 (0.59‐1.23); P = 0.397. Retrospective application of a Burkitt‐like (BL) molecular classifier identified a group of GCB pts (17%) with particularly poor (Figure Presented) prognosis. BL pts had a higher mutational burden than other GCB with an excess of c‐MYC rearrangements or extra copies (44/61 available). There was a trend towards improved PFS in BL pts treated with bortezomib (HR = 0.56; P = 0.069). Conclusion: The addition of bortezomib to R‐CHOP chemotherapy in DLBL may result in PFS benefit in sub‐groups of patients defined by molecular phenotyping. Cancer Research UK E/10/024.