Grzegorz S. Nowakowski, MD, responds to questions raised regarding the use of immunohistochemistry in his phase II study, as well as future plans for the investigation of R2-CHOP in non-GCB DLBCL.
Patients with non-germinal center B-cell-like (GCB) diffuse large B-cell lymphoma (DLBCL) have historically experienced poorer outcomes with traditional R-CHOP regimens compared to patients with GCB DLBCL. However, the addition of lenalidomide (Revlimid) to R-CHOP (R2-CHOP) showed promise in non-GCB DLBCL in a phase II study published in the Journal of Clinical Oncology (JCO).1
In the study, efficacy with R2-CHOP was similar between DLBCL subtypes. At a 24-month analysis, progression-free survival rates in patients treated with R2-CHOP were 60% and 59% in non-GCB and GCB subtypes, respectively (P = .83). Overall survival rates were 83% and 75%, respectively (P = .61).
However, a recent letter to the editor of JCO2 questioned the effectiveness of immunohistochemistry, which was the method used to identify subtypes in the phase II study. It also questioned the potential impact that possible errors in subtyping may have had on results.
The authors of the editorial suggested gene expression profiling might have produced more accurate results. Since the study was first initiated, advances in technology have made gene expression profiling more commonplace, and it is being used in the ongoing phase III ROBUST trial, an efficacy and safety study of lenalidomide plus R-CHOP versus placebo plus R-CHOP in untreated activated B-cell (ABC) type DLBCL.
Grzegorz S. Nowakowski, MD, assistant professor of Medicine, Mayo Clinic, and an author of both the phase II and ROBUST trials, recently spoke with OncLive. In his interview, he responds to questions raised regarding the use of immunohistochemistry in his phase II study, as well as future plans for the investigation of R2-CHOP in non-GCB DLBCL.
OncLive: What are the biggest challenges in the treatment of DLBCL?
Dr Nowakowski: There are two major subtypes of DLBCL: ABC subtype and GCB. What is interesting about DLBCL is that it is very heterogeneous in its outcome and we can, on average, cure about 60% of the patients with R-CHOP chemotherapy. However, if you look at the subtypes more closely, we can actually cure many more GCBs than ABCs.
We have known about this for a long time. The problem is, we cannot typically determine ABC versus GCB by gene expression profiling in real time. It takes a significant amount of time to generate the data regarding if a tumor is ABC or GCB, so most studies have been retrospective in the past. With immunohistochemistry, which we used in our phase II study looking at lenalidomide and R-CHOP in DLBCL, we can actually look at a set of markers and then make a real-time identification of the subtype at the time the patient is seen. This can identify GCB and non-GCB subtypes, which include ABC and possibly other patients who normally would not be classified.
What were the most significant findings from the phase II study?
In the phase II study, regardless of subtype, we added lenalidomide to R-CHOP in 64 patients. We looked at the outcome of those patients based on their initial risk factors, and they appeared to be slightly better than we would expect based on historical control. Among non-GCB patients, who typically have worse outcomes, those who received R-CHOP plus lenalidomide did much better than those who just received R-CHOP historically.
A recent letter to the editor in the Journal of Clinical Oncology referred to the immunohistochemistry methods used in your phase II as “a poor prognostic tool, which is not sufficiently accurate for the selection of patients for targeted therapies.” What is your response to this criticism?
The authors were concerned that immunohistochemistry was not as reliable as gene expression profiling to identify patients’ subtypes. In my mind, there is no question that gene expression profiling is better; however, this technology was not available at the time we did this study.
However, immunohistochemistry does provide us with a pretty good approximation when performed in a central laboratory and interpreted by an experienced hematopathologist. When it is done in this method, a significant level of concordance can be achieved. In the real world, there is much more variability in interpretation. More importantly, we used the same immunohistochemistry methods in the control cohort and in the lenalidomide arm. Therefore, the findings we determined with this method appear to be very real.
What are the next steps in this research?
The data from the phase II trial has led to the development of two studies. ECOG 1412 compared R2-CHOP versus R-CHOP for the initial treatment of DLBCL. The study is almost completed, and is currently being amended to allow a few more patients.
The second, which is more important, is the phase III ROBUST trial, which took into account our data that shows the benefit to be mainly in the non-GCB patients. Now, we can identify them much more reliably by gene expression profiling using a new technique called nanostrings. Nanostrings are a huge breakthrough in DLBCL subtyping because it can be done in real time without doing a biopsy.
To my knowledge, this is the only study that is using the biomarker of gene expression profiling in real time. We can identify non-GCB patients, which we suspect from our phase II study will benefit from lenalidomide, and just take those patients into the clinical trial and randomize to R-CHOP alone and lenalidomide plus R-CHOP. It is truly a personalized medicine approach. This is the first time we have been able to identify patients who have a higher risk of progression and relapse, and intervene with those patients with the addition of lenalidomide.
Note: Nowakowski published a column in JCO responding to Gleeson et al’s editorial: Nowakowski GS, et al. Reply to M. Gleeson et al [published online August 10, 2015]. J Clin Oncol. pii:JCO.2015.61.8280.