It is rare in the treatment of cancer that a novel regimen offers both superior efficacy and reduced toxicity compared to the prevailing standard treatment. It is also surprising that this might be accomplished in Hodgkin lymphoma, a malignancy with exceptional long-term outcomes. Two recent trials have achieved just this: the US Southwest Oncology Group S1826 trial1 and the German Hodgkin Study Group (GHSG) HD21 trial2 each represent a new and distinct standard of care for patients with advanced-stage Hodgkin lymphoma. We compare the design, population, efficacy and toxicity of these practice-changing trials and raise questions for future clinical trials. Since the development of mechlorethamine, vincristine, procarbazine, prednisone (MOPP) combination chemotherapy at the National Cancer Institute in the 1960s, Hodgkin lymphoma has seen serially improved outcomes from sequential iterations of chemotherapy combinations. ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine), originally designed by Gianni Bonadonna for MOPP-refractory cases, quickly became the standard for both efficacy and reduced toxicity.3 More recently, improvements to Hodgkin lymphoma have evolved in parallel on either side of the Atlantic. Intensification of upfront therapy with BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisone) in standard or escalated (eBEACOPP) dosing is commonly employed in Europe. While HD9 demonstrated an overall survival (OS) advantage of eBEACOPP over ABVD/COPP,4 four cooperative group trials (HD2000, FM/IIL, EORTC 20012, LYSA H34) showed a progression-free survival (PFS) but not OS benefit for (predominantly standard dose) BEACOPP over ABVD, ranging from 12% to 18% absolute PFS benefit.5 Notably, a meta-analysis of five trials reported a potential OS benefit for BEACOPP.6 The HD18 trial led to positron emission tomography (PET) adaptation of eBEACOPP, showing that patients with a complete metabolic response (CMR) after two cycles had similar outcomes when therapy was abbreviated to four cycles total, improving its risk:benefit ratio.7 For many clinicians, especially in the United States, the improved PFS of eBEACOPP is perceived as insufficient to outweigh its increased toxicity—in particular, myelotoxicity and febrile neutropenia in the short term and infertility, early menopause and second malignancies in the long term. Building on ABVD, the ECHELON-1 trial showed that replacing bleomycin with the CD30 antibody-drug conjugate brentuximab vedotin (BV) improves PFS and, subsequently, OS.8 Similarly, the RATHL trial showed that, in patients with PET CMR after two cycles of ABVD, omitting bleomycin for cycles 3–6 did not lead to a decrement in PFS. Long-term follow-up of RATHL, ECHELON-1 and HD18 demonstrated a 5-year PFS of 79.8%,9 82.2%10 and 89.4%7 for these approaches respectively. The major shortcomings of these current standards include bleomycin lung toxicity (eBEACOPP and RATHL with lower frequency), high rates of peripheral neuropathy Brentuximab vedotin plus doxorubicin, vincristine and dacarbazine (BV-AVD), haematological, metabolic and gonadal toxicity (eBEACOPP). The GHSG HD21 trial compares the novel BV-containing Brentuximab vedotin, etoposide, cyclophosphamide, doxorubicin, dacarbazine, and dexamethasone (BrECADD) regimen with eBEACOPP. The intentions of the GHSG are made clear in their selection of co-primary end-points: a reduction in treatment-related morbidity (TRM) and non-inferior efficacy. In addition to tweaking the dosage of etoposide (reduced from 200 mg/m2 day (D) 1–3 to 150 mg/m2 D1–3) and doxorubicin (increased from 35 to 40 mg/m2 D1), BrECADD substitutes vincristine with BV, procarbazine with the less mutagenic and gonadotoxic dacarbazine,11 prednisone with dexamethasone (40 mg/m2 D1–4) and omits bleomycin altogether. HD21 included 1500 patients, aged 18–60, with advanced-stage disease (including stage IIB with bulk or extra-nodal involvement). Both treatment regimens were dose-adjusted based on toxicity and PET adapted after two cycles (four cycles for patients in CMR, six cycles otherwise). Median age was 31, 84% of patients had stage III/IV disease and 23% had an international prognostic score (IPS) ≥4. Notably, BrECADD was able to simultaneously substantially reduce toxicity and improve efficacy. Overall TRM was reduced, as were key adverse events of interest: need for red cell transfusion (52% with eBEACOPP vs. 24% with BrECADD), platelet transfusion (34% vs. 17%) and sensory peripheral neuropathy (49% vs. 39%). In contrast, rates of infection were similar (46% with eBEACOPP 19% grade ≥3 vs. 49% with BrECADD 20% grade ≥3), and rates of febrile neutropenia were increased with BrECADD (21% vs. 28%), of which most occurred during cycle 1. Two-thirds of patients in both arms achieved PET negativity after two cycles and received four cycles of treatment; the remaining one-third received six. The 4-year PFS results are impressive—94.3% of patients who received BrECADD are alive and free from lymphoma after 4 years, compared with 90.9% for eBEACOPP—as the study's authors rightly highlight, the ‘highest reported PFS in any randomised controlled trial in advanced stage classical Hodgkin lymphoma’. The CMR rate at end of therapy was 82% with BrECADD, and 14% of patients received consolidative radiation. Additional follow-up is required to assess OS and late toxicities; however, early OS data are promising (98.6% vs. 98.2% at 4 years), as are fertility data, with substantially higher rates of gonadal function and childbearing compared to eBEACOPP,12 and potentially reduced rates of second malignancies, although longer follow-up is required (2% vs. 3% second malignancies, 2 vs. 6 cases of treatment-related myeloid neoplasms). The results from the S1826 trial are similarly impressive. SWOG and Children's Oncology Group investigators randomised patients to receive either six cycles of BV-AVD or nivolumab-AVD (N-AVD). Programmed death-ligand 1 (PD-L1) is ubiquitously expressed on Hodgkin Reed–Sternberg cells, and checkpoint inhibitors have powerful monotherapy activity in relapsed disease.13 Chemotherapy combinations may synergise checkpoint inhibitor efficacy by increasing tumour antigen expression and altering the immune-suppressive microenvironment, with promising front-line phase 2 data.14 The S1826 trial enrolled 994 patients aged 12 or older with exclusively advanced-stage disease. One quarter of patients were aged 12–17, two-thirds aged 18–60 and 10% over 60. Patients had high-risk disease, with 63% stage 4, 29% bulky disease (>10 cm) and 32% IPS ≥4. Although more patients who received N-AVD became neutropenic (55% vs. 32%), this was in the setting of reduced granulocyte colony-stimulating factor use (54% vs. 95%) and similar rates of febrile neutropenia (5% vs. 7%), sepsis (2% vs. 3%) and infections (5% vs. 8%). Peripheral neuropathy was reduced with N-AVD (29% vs. 55%; grade ≥3 1% vs. 8%), and immune-related adverse events were perhaps less frequent than expected: increased alanine aminotransferase (32% N-AVD vs. 41% BV-AVD), increased aspartate aminotransferase (25% vs. 32%), rash (15% in both), pneumonitis (2% vs. 3%), colitis (1% in both), with more frequent thyroid toxicity (hypothyroidism 7% vs. 1%, hyperthyroidism 3% vs. 0%). Indeed, tolerability is emphasised by only 9.4% of patients discontinuing nivolumab compared to 22.2% discontinuing brentuximab. After median 2.1 years of follow-up, the PFS with N-AVD was 92% compared to 83% with BV-AVD in the intention-to-treat population (92% vs. 86% among 18- to 60-year-olds). The complete response rate at end of therapy was 84% with N-AVD and only 0.4% of patients received consolidative radiation. OS was similar between two arms (99% vs. 98%), noting that it took 6 years of follow-up for the OS advantage to become evident in ECHELON-1. The benefit of N-AVD seems particularly pronounced in the >60 age group (2-year PFS 65% vs. 88%, HR 0.3 95% confidence interval 0.12–0.72), probably due to improved tolerability.15 Major advantages of BrECADD include a total duration of chemotherapy of 12 weeks in two-thirds of patients, low cumulative anthracycline dose (160 mg/m2 doxorubicin in 2/3 patients) and substantially reduced toxicity over eBEACOPP. Disadvantages include logistical complexity (12–24 chemotherapy treatment days in total), high rates of febrile neutropenia and infection (grade ≥3 19%), high steroid exposure (dexamethasone 40 mg ×4 per cycle), use of radiation in 14% of patients and uncertainty about how biochemical measures of fertility will translate to clinical outcomes. Although the PFS may seem better than in S1826, especially given HD21's longer follow-up, cross trial comparison is hampered by HD21's inclusion of stage IIB patients (16%) and exclusion of patients aged 60 years. Major advantages of N-AVD include simplicity (all comers, non-PET adapted, non-dose adjusted), tolerability (5% rate of febrile neutropenia, minimal myelotoxicity), safety and efficacy in the elderly population and minimal use of consolidative radiation (Table 1). Disadvantages include a longer duration of treatment (24 weeks, although many patients may continue to work), higher cumulative anthracycline dose (doxorubicin 300 mg/m2), uncertainty regarding autoimmune toxicity (e.g. thyroid dysfunction), shorter follow-up (70% information fraction) and theoretical concerns about checkpoint inhibitor sensitivity in subsequent lines of treatment. Although autoimmune colitis and pneumonitis are often seen after single-agent checkpoint inhibitor use in solid tumours, these were seen rarely and with similar frequency in the N-AVD and BV-AVD arms (table S9 of original manuscript), and longer follow-up is required to comprehensively characterise potential immune-related adverse events. Included: 18–60 years Median: 31 years 15% 50–60 years Included: >12 years Median: 27 years 10% >60 years III, IV IIB with bulk or EN (16%) III, IV (zero stage II) 12 weeks (2/3 patients) 18 weeks (1/3 patients) 160 mg/m2 (2/3 patients) 240 mg/m2 (1/3 patients) 4-year PFS 94.3% (95% CI 92.6–96.1) 2-year PFS 92%b (95% CI 89–94) Ultimately, both regimens represent a major improvement on two good standard approaches, and both remove the need for bleomycin. Once these regimens become more widely available, in the absence of head-to-head trials, physicians will need to personalise therapy based on age, fitness, comorbidities (e.g. autoimmune disease, cardiac disease), patient preference/philosophy and institutional familiarity. Of course, we can always do better: novel combinations are needed in early-stage disease and more precise ways of tailoring therapy type and duration (e.g. circulating tumour DNA, molecular profiling) to minimise toxicity are desirable. In the new era of Hodgkin lymphoma, global collaboration can ensure that we retain the excellent efficacy demonstrated in these two trials, with shorter and less toxic treatments. All three authors conceived of, wrote and edited the manuscript, and all authors approved the final version for submission. Open access publishing facilitated by The University of Melbourne, as part of the Wiley -The University of Melbourne agreement via the Council of Australasian University Librarians. No specific funding for this manuscript. ERSC has received research funding from Arnold Ventures outside this work. MJD has received research funding (to organisation), honoraria and speaker fees from Roche, Novartis, Bristol-Myers Squibb, AbbVie, GenMab and Kite/Gilead. As this manuscript did not generate any new data, this was not considered human subjects' research, and therefore, ethics approval was not sought. No new data were generated in this manuscript—all data are taken from published trial manuscripts.
Cliff et al. (Mon,) studied this question.