Clinical Impact of Ibrutinib with R-CHOP in Untreated Non-GCB DLBCL Co-Expressing BCL2 and MYC Genes in the Phase 3 Phoenix Trial

Introduction

In the phase 3, double-blind, placebo (pbo)-controlled PHOENIX trial (NCT01855750), 838 patients (pts) with non-germinal center B-cell-like (non-GCB) diffuse large B-cell lymphoma (DLBCL) were randomized 1:1 to ibrutinib (IBR; 560 mg/day orally) + rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) or pbo + R-CHOP. IBR + R-CHOP did not improve event-free survival (EFS) in the intent-to-treat (ITT) non-GCB population (hazard ratio [HR] 0.934; 95% confidence interval [CI], 0.726-1.200). However, in an exploratory analysis, pts < 60 years benefited from the addition of IBR (HR 0.579; 95% CI, 0.380-0.881 for EFS), whereas pts ≥ 60 years did not (HR 1.228; 95% CI, 0.887-1.699 for EFS) due to increased toxicity and reduced R-CHOP exposure. Based on evidence that co-expression of BCL2 and MYC by immunohistochemistry (IHC) have a worse outcome with R-CHOP, we examined their clinical prognostic effect in the 2 arms of the PHOENIX trial.

Methods

Pretreatment formalin-fixed paraffin-embedded biopsy samples were collected. RNA was extracted and profiled with Illumina RNAseq. Raw sequence reads were aligned to the hs37d5 genome build with STAR v2.5.1b, and gene-level quantification was conducted using RSEM v1.2.23. The median transcript per million mapped reads (TPM) values for BCL2 and MYC gene expression across all pts with RNAseq data (n = 766) were used as the cutoffs between high and low expression for each gene. BCL2 IHC data were available from 184 pts, and based on these, the threshold expression value from RNAseq data that could best approximate positive calls from IHC (≥ 50% lymphoma cells positive) was determined. This threshold value was very close to the median level calculated above and therefore supported the use of the median expression levels as the cutoffs to define high and low expressors in the subsequent analyses. The relationship between expression by RNAseq and survival (EFS, overall survival [OS]) in the 2 study arms was analyzed by Kaplan-Meier estimate with Cox regression to determine HRs and log-rank testing to assess significance.

Results

Based on a cutoff at the median TPM value, the percentage of non-GCB pts with BCL2-high + MYC-high (n = 234, 30.5%) was consistent with previous literature. In the IBR + R-CHOP (n = 386) and pbo + R-CHOP (n = 380) arms, respectively, 123 (31.9%) and 111 (29.2%) pts were MYC-high + BCL2-high by RNAseq. In the pbo + R-CHOP arm, pts with MYC-high + BCL-2-high had worse EFS (HR 1.820; 95% CI, 1.264-2.620; p = 0.0011) and OS (HR 1.662; 95% CI, 1.014-2.724; p = 0.0415) versus those with low expression of 1 or both markers in the ITT population (Figure), consistent with known poor outcomes associated with these genes. However, there was no difference in outcome for high versus low expression of these genes in the IBR + R-CHOP arm in the ITT population. In the ITT population, pts with MYC-high + BCL2-high had better EFS (HR 0.648; 95% CI, 0.423-0.993; p = 0.045) with IBR + R-CHOP versus pbo + R-CHOP, but there was no significant difference in OS (HR 0.783; 95% CI, 0.446-1.372; p = 0.39; Figure). We also examined the outcome of the 97 (30.6%) pts < 60 years of age (n = 317) with MYC-high + BCL2-high and observed an improved EFS and OS with IBR + R-CHOP versus pbo + R-CHOP (HR 0.393; 95% CI, 0.198-0.780; p = 0.0056 for EFS; HR 0.191; 95% CI, 0.055-0.666; p = 0.0037 for OS; Figure). There was no significant difference in EFS or OS in pts < 60 years with MYC-low + BCL2-low or in pts ≥ 60 years with MYC-high + BCL2-high in the 2 arms.

Conclusions

In this exploratory analysis, IBR was associated with improved EFS in combination with R-CHOP compared with pbo + R-CHOP in pts with MYC-high + BCL2-high expression in the ITT (non-GCB) population, without a significant improvement in OS. In pts aged < 60 years, both EFS and OS were significantly better with IBR, while there was no significant difference in older pts. These data suggest that IBR + R-CHOP may particularly benefit pts with MYC-high + BCL2-high-expressing lymphomas, a hypothesis warranting further testing in other DLBCL cohorts.

View largeDownload slide

View largeDownload slide

Close modal