Can Cell of Origin Inform Management of DLBCL?

The idea of determining the “cell of origin” (COO) in diffuse large B-cell lymphoma (DLBCL), and using it as a prognostic indicator or to guide treatment, remains somewhat controversial. In a talk at the 11^th Annual National Comprehensive Cancer Network (NCCN) Hematologic Malignancies Congress, held September 30–October 1 in New York, an expert broke down the latest research in this field and how COO might be used.

“This is a topic that has been kicking around for 16 years,” said Andrew D. Zelenetz, MD, PhD, of Memorial Sloan Kettering Cancer Center in New York. The COO was first described in 2000, and was important in firmly establishing that DLBCL is in fact three relatively distinct diseases: germinal center (GCB), activated B cell (ABC), and a third category best described as “unclassified” disease.

“Activated B cell and GCB are simply different diseases,” Zelenetz said. “There are real differences in molecular features, and hence, potentially in molecular targets for the treatment of these diseases.” Studies have found that patients with each of these types of disease respond to the same therapies differently, and have differing survival outcomes; thus, determining COO can thus offer prognostic information and may help guide treatment selection.

It is important to remember, though, that COO does not tell the whole story, Zelenetz said. Just as with other malignancies, somatic mutations can complicate the issue, though some specific mutations do occur largely or only in GCBs or ABCs, and some of those mutations offer therapeutic targets themselves. “Some mutations predict for sensitivity to treatment, others actually predict for resistance,” Zelenetz said.

Before COO can actually be put to use, of course, clinicians must have a reliable and easy method for determining it. Gene expression profiling on fresh tissue is considered the “gold standard” for doing so, but this is not practically applicable in clinical practice. Thus, immunohistochemistry approaches are often used.

There are many assays available to use immunohistochemistry to determine COO. One such method involves the Hans algorithm, a simple three-step process: a sample positive for CD10 is immediately classified as GCB; if negative for CD10, and negative for BCL6 as well, the tumor is deemed non-GCB. If negative for CD10 but positive for BCL6, the algorithm turns to MUM1/IRF4, and a negative result is classified as GCB and a positive as non-GCB.

Another option is the Choi algorithm, and one more is the Muris algorithm, which proceeds in a different order than the others. “The accuracy and reproducibility of Hans and Choi is quite questionable,” Zelenetz said, adding that the Muris may more accurately reflect results one would get with the gold standard of gene expression profiling on fresh tissue.

More recently, the Lymph2Cx assay was developed, which allowed the extraction of RNA tissue from paraffin-embedded tissue and assay it. Validation studies showed a strong correlation with the gold standard with this assay. “Here we have a robust assay,” Zelenetz said. “It allows us to then ask the question: Is COO prognostic in large cell lymphoma today?”

There is data suggesting the answer to that is yes. In one population registry-based analysis of 339 patients, LymphC2x allowed assignation of COO to 335 of 339 patients (99%); most of those (56%) were GCB, followed by 32% ABC, and 11% unclassified. There was in fact a difference in both progression-free and overall survival between GCB and ABC tumors, with 5-year outcomes favoring the GCB tumors (P < .001 for both).

Notably, though, the same assay applied to other data did not show as large a difference, Zelenetz said, though there was a trend in favor of GCB. The assays are clearly not perfect, though they do seem to offer useful information.

The next question is whether different chemotherapy regimens can influence outcomes in each COO group. Again, some early results suggest this is the case, though one large trial (CALGB 50303) has specifically asked that question-its results will be presented later this year at the American Society of Hematology (ASH) annual meeting. In another trial that compared R-ACVBP chemotherapy with the R-CHOP regimen, there was no difference at all in GCB tumors, but an “enormous” difference in favor of R-ACVBP in the ABC patients, which Zelenetz said represents “evidence you can overcome the adverse effect of these tumors with chemotherapy.” However, the two regimens are actually quite similar, complicating the result somewhat. Trials of other therapies will also soon shed light on the influence of COO on outcome.

Zelenetz said that all things considered, COO does seem to have some prognostic value, though that remains controversial. “We need to understand and confirm the good outcomes of GCB,” he said, and further study is needed to understand the “wild card” somatic mutations that complicate COO’s place in DLBCL management.