New clues to the prognostic challenge of Hodgkin lymphoma

One of the greatest success stories in modern anti-neoplastic therapy has been the steady improvement in survival among patients with Hodgkin lymphoma (HL). Although our current treatment strategies manage to cure the majority of these patients, a substantial proportion will still experience relapse or refractory disease and eventually die from this disease. Also, on the other hand, cured patients are at risk of serious treatment-related late morbidity and mortality. A large effort has been made to identify new prognostic factors enabling us to distinguish those patients who might benefit from a reduction in treatment intensity in order to minimize treatment-related toxicity, from those patients in need of a more aggressive treatment strategy in order to avoid treatment failure.

The value of early interim positron emission tomography/computed tomography (PET/CT) scans has been identified as the single most important predictor of treatment outcome [1]. Based on this finding, a large number of trials are currently addressing the question of whether early interim PET/CT-response adapted therapy can be used to select the optimal treatment strategy, thereby improving treatment outcome. Although interim scans undoubtedly provide us with a very important measurement of the initial response to therapy, and even though outcome can be accurately predicted already within the first month of therapy [2], the time-dependent nature of this variable also implies the major drawback of this predictor, as it does not allow a pretreatment risk stratification of patients. Consequently, the initial choice of treatment strategy has to be determined by other clinico-pathological parameters, and for this reason there is an ongoing effort to identify new pre-therapeutic markers that can provide us with reliable tools for selection of the initial treatment, including the choice between the ABVD regimen (doxorubicin, bleomycin, vinblastine and dacarbazine) and the more effective but also more toxic BEACOPPesc regimen (escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone).

The classical HL (cHL) tumor lesion is characterized by a minority of neoplastic cells surrounded by a heterogeneous background population of non-neoplastic bystander cells, including tumor-associated macrophages. A possible association between poor prognostic factors in cHL and tumor-associated macrophages was suggested several years ago [3], and in recent years several reports have supported an association between high numbers of macrophages and a poor outcome in cHL [4–6]. Other studies have failed to report such an association, potentially due to a lack of standardized scoring methods.

In the present issue of Leukemia and Lymphoma, Touati et al. report the results of a retrospective study including a total of 158 patients with cHL and their findings regarding the presence of intratumoral macrophages and their association with interim PET status and prognosis. The authors quantified the degree of macrophage infiltration in a semiquantitative manner in pre-therapeutic biopsies using an anti-CD68 based immunohistochemical strategy. In concordance with the majority of publications regarding this subject, an association between high macrophage counts and an adverse impact on prognosis was demonstrated.

Among the 158 patients in the current study, 68 patients had an interim PET scan available for analysis. Based on this material a correlation between a high number of macrophages and a positive interim PET scan could be demonstrated. In this perspective, a study previously published by Khandani et al. indicated a possible association between the expression of glucose transporters (Gluts) and the intensity of the PET signal in HL [8]. In that study, Gluts were found to be expressed on both the malignant and non-malignant cell populations, and given this observation it was suggested that the inflammatory infiltrate (including macrophages) in the HL lesion could possible play an important role in the tracer distribution. With this in mind, the correlation demonstrated by Touati et al. may reflect more than just an (expected) association between two already known prognostic factors, but could also suggest important biological clues to a better understanding of the biology behind the PET signal in this disease. Is the very high tracer uptake produced by the Hodgkin Reed–Sternberg cell population (less than 1% of all cells in the tumor) or by the inflammatory component? Does the interim PET signal simply tell us whether the tumor-promoting microenvironment is put to silence or not?

Even though the number of patients available for PET2 (PET post-cycle 2) evaluation is limited in the present study, an attempt is carried out to combine the prognostic statement of the intratumoral macrophages and the PET2 scan. Even though no significant impact on overall survival (OS) can be demonstrated, it is noted that four separate groups can be distinguished in relation to progression-free survival (PFS) and that the most favorable impact on survival is found among patients with low CD68 counts and a negative interim PET scan.

At the present time, the number of intratumoral macrophages has convincingly been shown to be associated with an adverse impact on prognosis in HL, but the question remains whether this marker can mature to become a reliable future predictor allowing upfront risk stratification. In order to determine the optimal method for quantification and cut-offs, further studies are needed. Due to the limited number of patients included in the study by Touati et al., this publication calls for a confirmatory study regarding the prognostic role of interim PET and tumor-infiltrating macrophages in cHL.

If an association and a possible additive impact of these prognostic factors are indeed confirmed, we shall still be left with the most important question: how can we integrate these findings into the clinical setting in order to provide our patients with a better and more tailored treatment?

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.