Prognostic markers in diffuse large B-cell lymphoma

Diffuse large B-cell lymphoma (DLBCL) is the most common non-Hodgkin lymphoma and comprises distinct subtypes [1]. Currently, patients with DLBCL are treated with immunochemotherapy, usually R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone). Despite the improved outcome of patients receiving this therapy, there are still a considerable number of patients who fail therapy. Especially for the latter patient group, there is a need for reliable prognostic markers that may guide alternative treatment options. While studying prognostic markers, it must be realized that these will change constantly as therapy changes, which some compare to trying to hit a moving target [2]. Currently, three types of prognostic factors are recognized: clinical, radiologic, and molecular. The last includes markers at the DNA, RNA, and protein levels.

Clinical prognostic factors, especially the international prognostic index (IPI) or modifications of this index, have hitherto proved to be very robust [3]. The IPI was developed before the introduction of immunochemotherapy, but can also be applied to patients treated with R-CHOP [4]. This index is composed of five factors, age >60 years, Eastern Cooperative Oncology Group (ECOG) performance ≥ 2, elevated lactate dehydrogenase (LDH) levels, >1 extranodal disease site, and stage III/IV disease. The IPI is still the gold standard to which novel prognostic markers are compared, and novel markers are most valuable when proved to be independent of the IPI. Positron emission tomography (PET) performed early during treatment has also proved to be an early predictor of treatment response and prognosis in DLBCL, although the results are variable according to the study [5]. Molecular prognostic markers for DLBCL have been studied extensively. Genetic, gene expression, and protein expression studies have revealed several markers that can be used for prognosis prediction in DLBCL. Gene expression analysis enabled recognition of two distinct subtypes of DLBCL, one of germinal center B-cell origin (GCB) and one of activated B-cell (ABC) origin [6]. Genetic analysis confirmed the existence of these two subtypes. GCB DLBCL has a better prognosis than ABC DLBCL, also when treated with R-CHOP [7]. In addition, gene expression analysis permits determination of other gene signatures derived from stromal cells that further define prognosis for DLBCL, independently from the IPI. Genetic markers as well as gene expression profiles are good prognostic markers, although suffer from the drawback that they cannot easily be introduced into today's routine practice. Their application awaits the development of less complicated technology to be used on readily available materials such as formalin-fixed paraffin-embedded tissues. Markers that are analyzed by immunohistochemical methods on paraffin-embedded tissues, on the other hand, can more easily be introduced into routine practice. The expression of markers such as Bcl-2, Bcl-6, CD10, CD5, survivin and derived major histocompatibility complex (MHC) class II antigens, to name some, has been shown to correlate with disease outcome. Nonetheless, these methods also need proper validation and standardization before most of these markers can be used reliably. The lack of standardization is likely responsible for the lack of reproducible results regarding the prognostic value of some of these markers [8].

In this issue of Leukemia and Lymphoma, Rydström et al. [9] report on the use of CD40 expression as analyzed by immunohistochemistry in paraffin-embedded tissues as a prognostic marker. The expression of CD40 is reported as a favorable marker for prognosis for patients with DLBCL who have been treated with immunochemotherapy. The results are interesting, but as the authors indicate, the prognostic value of CD40 should be determined in other patient series for validation. The lack of independent prognostic value with multivariate analysis that includes the IPI index, as shown by the authors, may diminish the interest of the findings if also proved by other studies. Although CD40 negativity as defined by the authors identifies a group of patients with DLBCL with a worse prognosis, this group still shows an overall survival and progression-free survival of 54% and 49%, respectively. This is no real improvement over the IPI. Finding a prognostic marker that defines a population with an even worse outcome would allow better identification of a group of patients in need of novel treatment. Irrespective of this, the use of CD40 staining in DLBCL may be useful not only for prognosis prediction, but perhaps also for defining better those patients who may benefit from the use of anti-CD40 antibodies for the therapy of DLBCL [10].