Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin lymphoma. Genetic analyses have revealed molecular heterogeneity of DLBCL tumors, which has led to the cell-of-origin (COO) classification of DLBCL into two subtypes: germinal center B-cell–like (GCB) and activated B-cell–like (ABC). An overview of the molecular classification of DLBCL, as well as current and future approaches to treatment, were the focus of the Education Session, “Diffuse Large B-Cell Lymphoma: Are We Ready for Molecular Subtype-Specific Therapy?” held on Saturday, May 30.
Dr. David Scott
Ascertaining Cell-of-Origin Subtypes in DLBCL
The GCB and ABC subgroups have different pathologies and treatment outcomes. For example, patients with the ABC subtype typically have poorer outcomes compared with patients with the GCB subtype. The COO of patients with DLBCL has, therefore, become an important predictive biomarker. David W. Scott, MBChB, PhD, of the BC Agency, led the session by reviewing the progress and remaining challenges of accurately determining the COO of DLBCL.
The initial gene expression profiling in DLBCL was performed using a micro-array that required fresh frozen tissue. Unfortunately, this methodology is not practical for use in clinical practice because patient biopsies are typically formalin-fixed, paraffin-embedded tissue (FFPET). Assay development efforts have, therefore, concentrated on immunohistochemistry (IHC) and gene expression methods that can utilize FFPET specimens. Evidence to date indicates that IHC assays have not progressed to the point that they can be routinely used to inform clinical decisions. This is largely a result of uncontrolled variables related to specimen processing and storage, lack of methodology standardization, antibody performance, and inter-observer variability. However, one IHC assay is being used to determine the COO of patients in the ongoing phase III PHOENIX trial (NCT01855750), which is investigating the addition of ibrutinib to rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) in non-GCB DLBCL. The assay is provided as a kit and is being performed in central laboratories using defined scoring criteria to minimize variability.
Advances in genetic analysis now enable genetic expression profiling of highly degraded RNA extracted from FFPET. A number of gene expression–based assays have been reported that use FFPET to determine COO, with the number of genes tested ranging from eight to approximately 180 with a work flow of 2-4 days. Two of these tests are being used in ongoing phase III randomized trials. Patients enrolled in the REMoDL-B trial (NCT01324596) are being stratified into equal groups of ABC and GCB DLBCL using the Illumina whole-genome complementary DNA-mediated Annealing, Selection, Extension, and Ligation assay.1 A NanoString-based assay2 is being used for patient selection in the ongoing phase III ROBUST trial (NCT02285062) of patients with the ABC DLBCL subtype. Dr. Scott noted that approximately 15% of patients with DLBCL will be considered unclassified, meaning they do not fit the criteria for either ABC or GCB DLBCL subtypes. “I’m a little concerned that clinical trials will exclude these patients, and they’ll be left in a therapeutic wasteland,” Dr. Scott said.
Assays used to determine COO must be externally validated and comprehensively characterized prior to routine clinical use. In addition, adequate amounts of tissue must be available for gene expression–based testing. Excisional biopsies provide enough tissue, but core needle biopsies may not be sufficient, particularly if the material is used for multiple tests during the diagnostic work-up. This may require routine excisional biopsies, or core needle biopsies dedicated for genetic testing in patients with DLBCL. Despite the remaining challenges, Dr. Scott is confident that reliable assays robust enough for clinical use will be available in the near future.
Treatment Based on DLBCL Subtype
Grzegorz S. Nowakowski, MD, of the Mayo Clinic, gave an overview of front-line treatment of DLBCL and molecular subtype–specific therapy. The R-CHOP regimen remains the standard first-line treatment for patients with DLBCL. Given the heterogeneity of DLBCL, investigators are studying more regimens and combinations that may provide benefit to patients based on their molecular subtype. Several ongoing trials are investigating the effects of novel agents when added to R-CHOP in patients with DLBCL.
Phase Ib study data indicate that adding the proteasome inhibitor bortezomib to R-CHOP (Bor-RCHOP) in 16 patients with DLBCL and four patients with mantle cell lymphoma led to an overall survival of 75% at 4 years, with a 58% progression-free survival.3 To follow up these findings, the REMoDL-B study is underway, in which patients with DLBCL undergo one cycle of R-CHOP while gene expression profiling is performed. The trial randomization is stratified by molecular subtype, and patients are assigned either to Bor-RCHOP or R-CHOP for five cycles. Interim analyses will be conducted to determine if the addition of bortezomib provides benefit for the GCB group.
In a phase II trial of 18 patients with DLBCL, the addition of Bruton’s tyrosine-kinase inhibitor ibrutinib to R-CHOP (IR-CHOP) led to a 100% objective response rate (ORR) with 15 complete responses and three partial responses.4 The phase III, randomized, double-blind PHOENIX study is enrolling patients with newly diagnosed non-GCB subtype DLBCL as determined by real-time genetic expression profiling. Four hundred patients will be randomly assigned to R-CHOP every 21 days (six to eight cycles) plus 560 mg/day of ibrutinib, and 400 will receive R-CHOP every 21 days (six to eight cycles) plus placebo.
Two phase II studies investigating the addition of the immunomodulatory agent lenalidomide to R-CHOP (R2-CHOP) as front-line therapy for DLBCL have reported an ORR of 98% in 64 patients5 and 92% in 44 elderly patients.6 Overall, the R2-CHOP regimen was well-tolerated, and grade 3/4 toxicities and events of thrombosis were infrequent. These promising data are being further investigated in two studies. The first is the phase II E1412 trial (NCT01856192), which will randomly assign up to 300 patients to receive R-CHOP or R2-CHOP in equal numbers. Subtype analyses will be performed using IHC and gene expression analysis, and the efficacy analysis will be based on these subtypes. As previously mentioned by Dr. Scott, the randomized, phase III ROBUST trial (NCT02285062) will use real-time genetic analysis to determine patient eligibility. Patients with the DLBCL ABC subtype will receive either RCHOP or R2-CHOP, with 280 patients per treatment group. Dr. Nowakowski concluded by encouraging clinicians to enroll patients in these trials, including those who are elderly.
Treatment of Double-Hit DLBCL
MYC is a transcription factor involved in many important cellular functions, such as growth, proliferation, metabolism, protein synthesis, DNA replication, and angiogenesis. MYC rearrangements (MYC-R) have been identified in 7%-14% of patients with DLBCL and is associated with poorer prognosis and outcomes. A subset of patients with DLBCL have both MYC and BCL2 translocations, known as double-hit lymphoma (DHL). Patients with DHL are often characterized by a median age in the 70s, late-stage disease, high-intermediate to high International Prognostic Index score, elevated levels of lactase dehydrogenase, and extranodal sites of disease, including central nervous system (CNS) involvement. Other patients with DLBCL may have what is called double-expressor lymphoma (DEL), in which there is overexpression of the MYC and BCL2 proteins. CHOP and R-CHOP are not effective treatments in these patients, therefore, their disease prognosis is poor. Myron S. Czuczman, MD, of the Roswell Park Cancer Institute, presented the current treatment approaches in these types of aggressive lymphoma.
A variety of drug combinations have been studied in DHL, but no optimal therapy currently exists. As MYC is responsible for many functions in healthy cells, it is important to select drugs that target pathways activated by MYC only in malignant cells. “If we try to target MYC or the MYC protein, we are going to be in trouble because normal cells depend on MYC for the cell cycle and maintenance,” Dr. Czuczman said.
Although patients with DHL DLBCL treated with rituximab, etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (R-EPOCH) had better event-free survival and overall survival compared with other regimens,7 novel agents that target MYC and BCL2 have the best chance for making a difference in patients with DHL. Examples of potentially effective agents include the BCL2 inhibitor ABT-199, BET bromodomain inhibitors, the aurora-kinase inhibitor alisertib, and the proteasome inhibitor ixazomib, as well as immune checkpoint inhibitors. Ongoing clinical trials enrolling patients with DHL DLBCL will hopefully provide much needed direction regarding appropriate treatments.
Until optimal therapies are identified, Dr. Czuczman recommends that clinicians obtain biopsy specimens from all newly diagnosed patients with DLBCL and that these be tested via fluorescence in situ hybridization and IHC to determine DHL and DEL status, respectively. Patients with DHL or DEL should be encouraged to enroll in clinical trials, if at all possible. If trial participation is not an option, R-EPOCH induction with CNS prophylaxis is a reasonable approach.
Watch the session, visit the ASCO Virtual Meeting website.