The long-term follow-up analysis of the SWOG0816 trial, presented by Stephens et al1 in this issue of Blood, reminds us that long-term retrospect is needed to understand studies of a highly curable disease of the young.
Advanced-stage Hodgkin lymphoma (aHL) is most commonly treated with adriamycin, bleomycin, vinblastine, and doxorubicin (ABVD) or less commonly with a more intensive regimen containing bleomycin, etoposide, adriamycin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPPesc). ABVD is well tolerated, but results in ∼30% treatment failure. BEACOPPesc is more effective but too toxic to be given to all patients. Many observational studies have shown that early interim positron emission tomography (PET) is highly prognostic in aHL.2 This led to several trials investigating early PET response–adapted therapy, aiming to improve outcomes and reduce the overall toxicity burden.
The United Kingdom–led RATHL study included 1214 aHL patients.3 The 84% of patients with a negative PET after 2 ABVD cycles (PET2) were randomized to ABVD or adriamycin, vinblastine, and doxorubicin (AVD) for the remaining 4 cycles. There was no efficacy difference between the ABVD and AVD arms (so no need to give bleomycin beyond 2 cycles of ABVD in PET2-negative patients), and 3-year progression-free survival (PFS) for all PET2-negative patients was 86%. In the Italian GITIL0607 study, 81% of 782 enrolled patients with aHL were PET2 negative, and they reached a similar 3-year PFS of 87% with continued ABVD treatment.4 With 2 larger and relatively recent studies for comparison, the outcomes for PET2-negative patients in the SWOG0816 study are disappointing. Of 336 patients, 82% were PET negative after 2 cycles of ABVD. Like in the RATHL and GITIL0607 studies, they continued to a total of 6 cycles of ABVD, but the article by Stephens et al reports a 5-year PFS of only 74% for those PET2-negative patients. There is a long way from 86% and 87% to 74%, so what is the reason for this difference? Somehow, the negative predictive value of the central PET review of the SWOG study was lower than in the other studies. However, the longer follow-up in the SWOG study is also an important factor, because the article reports 5-year PFS with a median follow-up of 5.9 years compared with the 2 previous studies, which reported 3-year PFS, both with median follow-up of just over 3 years. As witnessed by the PFS curves presented by Stephens et al, relapses and other PFS events do occur beyond 2 to 3 years, apparently relatively more frequently than in the patients with insufficient initial response to chemotherapy.
Like in the RATHL and GITIL0607 studies, PET2-positive patients in the SWOG0816 study were offered treatment intensification with BEACOPPesc. Among the 18% of PET2-positive patients in the SWOG0816 study, 5-year PFS was 66%. This is in line with the 68% 3-year PFS and the 60% 3-year PFS observed in the RATHL and GITIL0607 studies and compares very favorably with the observational studies in which PET2-positive patients continued treatment with ABVD and reached long-term PFS rates of 10% to 35%. The catch is an alarmingly high rate of secondary cancers among the patients treated with BEACOPPesc. Seven of those patients (14%) experienced second malignancies within the 5.9 years of median follow-up; even though the numbers are small, this is even higher than in the long-term analyses of the German studies in which patients received 6 or 8 cycles of BEACOPPesc.5 Interestingly, in the RATHL study in which PET2-positive patients received only 4 cycles of BEACOPPesc,3 the cumulative rate of second malignancies at 3.4 years of follow-up was 1.7%. In the GITIL0607 study, in which PET2-positive patients received 4 cycles of BEACOPPesc, followed by 4 cycles of BEACOPPbaseline (same drugs and schedule as BEACOPPesc, but lower dose-intensity), no second cancers had been reported at 3.6 years median follow-up.4
What lessons can we learn from the 5-year update of the SWOG0816 study?
It is an important reminder that relapses and second cancers occur late in aHL patients; therefore, long-term follow-up reports like the one presented by Stephens et al are crucial to fully interpret study results.
Early PET response–adapted ABVD-based therapy for aHL results in improved outcomes for early interim PET-positive patients and in excellent overall outcomes (94% 5-year overall survival in SWOG0816; 97% 3-year overall survival in RATHL and GITIL0607).
Interim PET is far from perfect, and the study by Stephens et al suggests an even higher rate of false-negative results than other studies; therefore, more sensitive tools to detect residual unresponsive disease are needed.
Six cycles of BEACOPPesc following 2 cycles of ABVD lead to an unacceptable risk for second malignancies and do not offer better disease control than 4 cycles of BEACOPPesc, as given in the RATHL study.
The introduction of PET and PET response–adapted treatment were the most important new developments in the management of aHL 10 to 15 years ago. Since then, effective targeted agents (brentuximab vedotin, checkpoint inhibitors) have been introduced, and the Echelon-1 study showed improved efficacy in first-line treatment of aHL when bleomycin is replaced by brentuximab vedotin (ABVD vs AVD-A [adriamycin, vinblastine, dacarbazine, brentuximab vedotin]).6 Interestingly, 3-year PFS for ABVD-treated patients in the standard arm of the Echelon-1 study was 76%, regardless of PET2 results, which is better than the 5-year PFS for PET2-negative patients in the SWOG0816 study.7 Recently, refined and very sensitive tools to assess treatment response and minimal residual disease using serum markers and circulating tumor DNA analyses have been reported.8,9
The ongoing COBRA study from the European Organisation for Research and Treatment of Cancer is an attempt to combine novel regimens and refined response adaptation in aHL10 : all patients receive 1 cycle of AVD-A, after which PET/computed tomography is performed, along with analyses of circulating tumor DNA and disease-specific serum chemokines. Responding patients continue treatment with AVD-A, whereas nonresponders are offered intensification using the experimental brentuximab vedotin, etoposide, cyclophosphamide, adriamycin, dacarbazine, and dexamethasone regimen, which is undergoing randomized comparison with BEACOPPesc in the ongoing German HD21 trial.
It is hoped that more effective less toxic regimens, as well as better tools for response-adapted treatment, will further improve the efficacy and safety of first-line treatment of aHL.
Conflict-of-interest disclosure: M.H. has received research funding and honoraria for acting as an advisor/consultant from Takeda, Roche, and Celgene.