Interim response assessment for Hodgkin lymphoma: size matters

Treatment of early stage Hodgkin lymphoma (HL) has been a notable success story in the field of oncology, with more than 90% of patients with stage 1 and 2 HL cured using standard therapy. With this therapeutic success, increasing attention is being paid to the toxicity of therapy, with the goal of minimizing long-term toxicity of both chemotherapy and radiation. Complicating the pursuit of this goal is the fact that, although a significant majority of patients with early stage HL are cured with less intensive chemotherapy and/or elimination of radiation therapy, a minority of patients who will be cured by standard combined ABVD (doxorubicin, bleomycin, vinblastine and dacarbazine) and radiation therapy would not be cured if either the radiation is eliminated, or the chemotherapy regimen is decreased in intensity. This fact has been demonstrated in multiple studies showing a small but reproducible (though not always statistically significant) improvement in long-term lymphoma control in patients who receive combined modality therapy including radiation [1,2].

One way to optimize the balance of treatment efficacy and toxicity would be to reliably identify the minority of patients who do in fact need more intensive therapy. Clinical factors have been marginally helpful in this distinction, but still leave the majority of patients receiving more therapy than is necessary for cure. A more patient-specific strategy has been explored involving the use of [¹⁸F]fluorodeoxyglucose positron emission tomography (FDG-PET) scanning part way through treatment, usually after two cycles of chemotherapy. Several studies have demonstrated the ability of FDG-PET to predict outcome in advanced stage disease with high accuracy [3,4]. In stage 1 and 2 HL, however, the predictive power is less. While a negative PET scan after two cycles of therapy is highly predictive of a good outcome, a positive PET scan, although associated with inferior outcome overall, is still often seen in patients destined to be cured; that is, interim PET in early stage HL has a poor positive predictive value (PPV) for treatment failure [4,5].

Kostakoglu and colleagues address this poor PPV in a study published in this issue by evaluating whether the combination of metabolic data, as measured by FDG-PET, and anatomic data, as measured by computed tomography (CT) scan, can provide better predictive information than either study alone [6]. This analysis was embedded in a multicenter study designed to test an alternative chemotherapy regimen, doxorubicin, vinblastine and gemcitabine (AVG), in an attempt to decrease toxicity for patients with early stage HL [7]. Two different PET interpretation criteria were studied: the International Harmonization Project (IHP) criteria and the Deauville criteria, which differ in the reference for a negative finding, mediastinal blood pool versus liver, respectively [8,9]. They confirmed previous observations that, while the negative predictive value (NPV) of PET scan was high, the PPV was modest, with approximately half of patients with a positive PET scan maintaining long-term disease-free survival. When the authors incorporated data regarding reduction in size, however, the PPV improved significantly. They retrospectively chose a 65% reduction in the sum product of the perpendicular dimensions (SPPD) as a cut-off for a “negative” scan. The combination of metabolic and size response assessment better predicted outcome than either alone, particularly when responses were concordant: patients with both negative PET and CT had a 2-year PFS of 93%, whereas those with positive results on both studies showed a PFS of only 43%. Discordant studies were associated with an intermediate probability of PFS: 70% and 76%, depending on whether the PET or the CT was positive, respectively.

Several caveats to this study must be mentioned. First, as the authors point out, the backbone for this study was a chemotherapy regimen which did not appear to be as effective as standard therapy. Although not performed in a randomized controlled trial, the chemotherapy appeared to be inferior in terms of complete response (CR) rate and PFS as compared to ABVD. This may, in part, explain the lower NPV of PET in this study as compared to others evaluating standard ABVD. Another issue is the choice of interpretation criteria. The IHP response criteria were developed specifically for response assessment after completion of treatment, making this gauge marginally relevant to the current study, whereas the Deauville criteria were developed specifically for interim response assessment. As the authors point out, the study was designed prior to publication of the Deauville criteria, and therefore the IHP scale was chosen. Therefore, to maintain consistency with the original protocol, they included IHP measures in this report, though this likely should not be incorporated into future interim imaging studies.

Elimination of radiation therapy is another relevant issue in this study, but the conclusions to draw from this are less clear. On the one hand, incorporating radiation improves cure rates in early stage HL over standard ABVD chemotherapy alone. Furthermore, the AVG regimen explored here is likely inferior to standard chemotherapy. Basing this investigation on an inferior regimen calls into question whether this study is relevant to standard practice. On the other hand, a major direction of research in HL is dedicated to attempting to identify those patients in whom treatment can safely be reduced, principally with the elimination of radiotherapy. The authors were able to identify a group of patients with excellent outcomes even with inferior chemotherapy when radiation was eliminated through combining interim PET and CT data. This finding suggests that, with the more effective standard ABVD chemotherapy, patients with complete metabolic and anatomic response likely have an extremely high cure rate. We eagerly await completion and publication of randomized studies rigorously evaluating elimination of radiation therapy in patients with complete interim response.

Potential conflict of interest

A disclosure form provided by the author is available with the full text of this article at www.informahealthcare.com/lal.