The efficacy-toxicity conundrum: breaking the mold

Although the cure of advanced-stage Hodgkin lymphoma (HL) is a tremendous success, each new improvement in outcome has come with a significant price. Mechlorethamine, vincristine, procarbazine, and prednisone resulted in leukemia; doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) added possible fatal pulmonary toxicity; and bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP) improved the cure but at a price of increased toxicity and increased leukemia. The ECHELON-1 study, reported by Straus et al in this issue of Blood, may be the first advance to break this pattern.¹ 

Even more so today, physicians who are recommending treatment of patients with advanced-stage, classic HL (cHL) are faced with the dilemma of maximizing efficacy and limiting toxicity. This dilemma results from the fact that the 2 major chemotherapy regimens are at opposite poles in terms of toxicity and efficacy. Data from the German Hodgkin Study Group HD9 study support that, for cyclophosphamide, vincristine, procarbazine, and prednisone (COPP)/ABVD (a surrogate for ABVD), freedom from treatment failure (FFTR) was 64% and overall survival (OS) was 75%, whereas escalated BEACOPP (eBEACOPP) FFTR was 82% and OS was 86%.²  On the other hand, acute and chronic toxicity were just the reverse. For COPP/ABVD vs eBEACOPP, grade 4 leukopenia was 19% vs 90%, grade 4 thrombocytopenia was 2% vs 47%, grade 4 infection was 1% vs 8%, grade 3 or 4 mucositis was 1% vs 8%, and grade 3 or 4 pain was 2% vs 9%.³  The long-term toxicity of acute myeloid leukemia was 0.4% vs 3.0%, respectively.² 

Contemporary studies have gone in 2 directions to address this efficacy/toxicity conundrum. One direction is to use the more toxic eBEACOPP regimen for the poor-prognosis patients whose positron emission tomography (PET) scan is positive after 2 cycles (PET2⁺). Based on historically controlled data, advanced HL PET2⁺ patients have progression-free survival (PFS) between 30% and 44%.⁴  There have been 4 large studies that used dose escalation to BEACOPP to address this PET2⁺ problem^5-8  (see table for the detailed results). Several conclusions emerge from examination of these data. When treatment is escalated to a form of BEACOPP in PET2⁺ patients, the PFS is in the 60% to 70% range, which is significantly above the historical PFS of 30% to 40%. One of the critical data points in this update of the ECHELON-1 trial is that brentuximab vedotin (A) plus doxorubicin, vinblastine, and dacarbazine (AVD) (A+AVD) continued without change after PET2⁺; the results for continuing A+AVD are in the same range of PFS as changing to BEACOPP in PET2⁺. The initial comparator for these 4 studies was historically controlled data sets. Using the Deauville criteria, 3-year PFS for PET2⁺ advanced-stage International Prognostic Index (IPI) ≤ 2 was 30%; and, for IPI ≥ 3, the PFS was 44%.⁴  When measured against these historical controls of 30% and 44%, the PFS of PET2⁺for the 4 escalated BEACOPP studies (68%, 91%, 60%, and 66%) appears to show an improvement in disease control.

The update of the ECHELON-1 study, reported in this issue, provides important data on the efficacy side of the question. Advanced-stage, cHL patients were randomized to A+AVD or ABVD. An interim PET after 2 cycles was required. The median follow-up is now 37 months. Although the 3-year PFS rates were 83.1% with A+AVD and 76% with ABVD, the critical point for PET2⁺ patients was 69.2% with A+AVD and 54.7% with ABVD. Although this is not a randomized comparison between PET2⁺ patients receiving eBEACOPP and A+AVD, it does provide strong evidence that the PET2⁻ patients have better PFS with A+AVD than with ABVD and the PFS of the PET2⁺ patients with A+AVD is better than with ABVD and in the same range as the eBEACOPP patients.

The second part of the conundrum is toxicity. To get the higher PFS of A+AVD, what is the cost in terms of toxicity? The major toxicities grade ≥3 were neutropenia (54% vs 39%), peripheral neuropathy (PN; 4% vs <1%), abdominal pain (3% vs <1%), infectious death (1.1% vs 0%), and pulmonary deaths (0% vs 1.7%) for A+AVD and ABVD, respectively. All of the infectious deaths in the A+AVD arm occurred in patients who had not received growth factors before neutropenia. At the 3-year follow-up, the residual toxicities for any residual PN, grade 1, grade 2, and grade ≥3 were: 25.6%, 14.2%, 7.8%, and 2.7%, respectively, for A+AVD and 11.5%, 7.3%, 3.6%, and <1%, respectively, for ABVD. What these data reveal is that if the infection complications are controlled with growth factor, the toxicity of A+AVD is minimally increased but the efficacy is a marked step forward. The efficacy-toxicity conundrum is very possibly changed.

The decision regarding how to treat advanced-stage cHL is still a difficult one, and individual patient characteristics still need to be considered; however, the treating physician now has sufficient information that A+AVD: has a higher PFS than ABVD, gets similar results to dose-escalated eBEACOPP studies, eliminates or reduces the bleomycin pulmonary toxicity, results in rare (3%) residual grade ≥3 PN, and may significantly reduce or eliminate the risk of infectious death with the upfront use of growth factors.

The authors’ conclusions are fairly conservative in stating that A+AVD compares favorably to PET-adapted strategies and has a manageable and predictable safety profile. Historically, efficacy and toxicity had increased together, creating the efficacy/toxicity conundrum, but with A+AVD, many physicians may well decide that they can now offer their patients greater efficacy and decreased toxicity.