Biomarker Impact on Survival in Diffuse Large B-cell Lymphoma

March 12, 2020
Nichole Tucker

The largest prospective dataset in previously untreated diffuse large B-cell lymphoma revealed the molecular heterogeneity of diffuse large B-cell lymphoma with potential treatment targets harbored by the cell-of-origin subtypes, based on data from a phase III GOYA study.&nbsp;<br /> &nbsp;

The largest prospective dataset in previously untreated diffuse large B-cell lymphoma (DLBCL) revealed the molecular heterogeneity of DLBCL with potential treatment targets harbored by the cell-of-origin (CCO) subtypes, based on data from a phase III study (GOYA, NCT01287741).1

Results from GOYA were presented at the 2019 American Society of Hematology Annual Meeting and showed that the combination of obinutuzumab (Gazyva) or rituximab (Rituxan) plus cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) did not improve investigator-assessed progression-free survival (PFS). There was also no significant difference seen in the 5-year overall survival rate in the experimental arm with obinutuzumab versus the control arm of rituximab plus CHOP.2

The beneficial information from GOYA, however, was in the gene profiles of the patients involved. Although multiple studies have confirmed the prognostic impact of COO subtypes, there has not been much research on how key genomic alterations impact clinical outcomes in patients with DLBCL. For a clearer understanding of this and whether some somatic mutations in DLBCL could provide prognostic value outside of established clinical and biological risk factors, like COO and International Prognostic Index (IPI), an integrated analysis of the GOYA data was conducted. The results from the analysis were published inHaematologica.1

At baseline, the disease characteristics between patients who had next-generation sequencing (NGS) available and those in the overall GOYA intention-to-treat population were similar. A total of 465 genes were sequenced, and of those, 13% were identified that are known or likely to alter the function of a gene (n = 59) in at least 10 of 499 patient samples, and 334 additional genes with alterations were identified in 1 or more patients. Three percent of patients had no identified mutation. A median of 6 gene alterations per patient were observed (range, 0-17). Additionally, there was a median of 4 single nucleotide variants (SNVs) per patient (range, 0-16), and a median of 0 copy number abnormalities ([CNAs], range, 0-10).

At least one alteration was found in 97% of cases, and 2 or more alterations were found in 93% of cases. Of the gene alterations observed, SNVs were the most common, while CNAs were only found in a few gene types, includingCDKN2A/BandREL.

A total of 31 gene rearrangements were analyzed and of those,BCL2,MYCandBCL6were the

most frequently rearranged with immunoglobulin heavy chain locus being the most frequently observed translocation partner. Immunoglobulin heavy chain locus was found in 100% of theBCL2profiles (n = 92), 90.6% of theMYCprofiles (n = 29), and 57% of theBCL6profiles (n = 57).

In terms of the frequency of genomic alterations among COO subsets, patients who had both COO and NGS data available (n = 482) were categorized as either GCB (n = 272), unclassified (n = 78), or activated B-cell ([ABC] n = 132). The most prevalent gene mutations in the GCB category wereBCL2(32%),MLL2(KMT2D) (30%), andCREBBP(22%). In the ABC category, loss ofCDKN2A(49%) andCDKN2B(30%) and mutations ofMYD88(34%) were frequently observed.

There were 15 genes that appeared to be significantly differentially mutated between the GCB and ABC subtypes with false discovery rates (FDR) below 0.05.

In cases of GCB DLBCL, alterations includingBCL2,CREBBP,TNFRSF14,EZH2,REL,BCL7AandSGK1were more frequently observed, whereas the ABC DLBCL category showed alterations inBCOR,ETV6,PRDM1,PIM1,CD79b,CDKN2B,MYD88andCDKN2Amore frequently. Specific types of alterations inBCL2andCDKN2Adisplayed different frequencies between the GCB and ABC subtypes. Specifically,BCL2translocations and SNVs were more frequently found in the GCB subtype while high levels ofBCL2amplifications were found in the ABC subtype (6.8% vs 1.5%; Fisher’s exact testP= .012).

Alterations in 23 gene were evaluated for correlation with PFS. Bolen et al observed prognostic trends withBCL2,CREBBP,REL,TP53,andCDKN2A. YetBCL2was most strongly associated with PFS (HR, 2.6; 95% CI, 1.6-4.2; FDR, .0037) by both univariate and multivariate analysis. No biomarker, however, demonstrated a significant difference in prognostic impact between the treatment arms of the GOYA study.

The authors found that the negative effect ofBCL2alterations was seen with both SNVs and translocations.BCL2translocations specifically were significantly enriched in the GCB subgroup and associated with shorter PFS for patients with this subtype. The translocations were found to be associated with high mRNA and protein expression levels, which have been associated with a poor prognosis in DLBCL. According to Bolen et al, these findings indicate that pharmacological inhibition of theBCL2protein could be a promising treatment strategy in select patients with DLBCL.

No genetic alterations were found to be significantly associated with outcomes for patients with the ABC subtype of DLBCL, which typically has a worse prognosis than the GCB subtype.

To carry out the integrated analysis, investigators performed targeted DNA NGS in 499 formalin-fixed paraffin-embedded tissue biopsies from patients who had no prior treatment. Investigators also examined the prevalence of genetic alterations and mutations and used multivariate Cox regression models to assess the prognostic value of each genomic alteration.

“Using the largest prospective dataset in previously untreated DLBCL to date, we demonstrated the molecular heterogeneity of DLBCL, with potential treatment targets harbored by the distinct COO subtypes. Only alterations inBCL2were significantly associated with clinical outcome independently of COO and clinical factors, thereby demonstrating the strong prognostic value of COO for clinical outcome in DLBCL,” Bolen et al wrote.

The authors also noted that mutations in many potentially targetable pathways were found in the DLBCL samples, includingBCL2alterations that could be potentially treated with venetoclax (Venclexta). “These mutations, along with COO subtype information, would be useful for the design of clinical trials involving combinations of novel targeted therapies,” the study authors wrote.


  1. Bolen CR, Klanova M, Trneny M, et al. Prognostic impact of somatic mutations in diffuse large B-cell lymphoma and relationship to cell-of-origin: data from the phase III GOYA study [Published online November 14, 2019]. Haematologica. doi: 10.3324/haematol.2018.227892.
  2. Sehn LH, Martelli M, Teny M, et al. Final analysis of GOYA: a randomized, open-label, phase iii study of obinutuzumab or rituximab plus CHOP in patients with previously untreated diffuse large b-cell lymphoma. Blood. 2019:134(suppl 1);4088. doi: 10.1182/blood-2019-123433.