Efficacy of Nivolumab and AVD in Early-Stage Unfavorable Classic Hodgkin Lymphoma: The Randomized Phase 2 German Hodgkin Study Group NIVAHL Trial

Key Points

Question  What is the efficacy of concomitant or sequential nivolumab and doxorubicin, vinblastine, and dacarbazine (N-AVD) as first-line treatment for early-stage unfavorable classic Hodgkin lymphoma?

Findings  In this investigator-sponsored phase 2 randomized clinical trial including 109 adult patients, very high interim complete remission rates were observed after treatment with 2 cycles of N-AVD (87%) or 4 doses of nivolumab (51%). After end of treatment with 4 cycles of N-AVD and 30-Gy involved-site radiotherapy, efficacy measures, such as complete remission rates, 1-year progression-free survival, and 1-year overall survival were excellent in both groups.

Meaning  Nivolumab-based first-line treatment is highly effective in patients with early-stage unfavorable classic Hodgkin lymphoma and warrants further investigation.

Importance  In early-stage unfavorable classic Hodgkin lymphoma (cHL), conventional therapy induces high cure rates but also relevant acute and long-term toxic effects. Nivolumab is well tolerated and highly effective in relapsed/refractory cHL but has not been adequately studied in first-line treatment of early-stage cHL. The NIVAHL trial evaluated nivolumab in this setting with the aim to develop a highly effective yet tolerable systemic therapy to ultimately mitigate morbidity in patients who survive cHL.

Objective  To evaluate efficacy of 2 experimental nivolumab-based first-line treatment strategies in patients with early-stage unfavorable cHL.

Design, Setting, and Participants  This was an open-label, multicenter, phase 2 randomized clinical trial, open between April 2017 and October 2018. The trial took place at 35 trial centers across Germany, ranging from academic centers to private offices. Eligibility was defined by age 18 to 60 years, cHL confirmed by expert pathology review, early-stage unfavorable disease by German Hodgkin Study Group criteria (stage I to II with risk factor[s]), and absence of serious concomitant disease or organ dysfunction. Among 110 enrolled patients, 109 were eligible.

Interventions  Systemic therapy, per random assignment (1:1) to either concomitant treatment with 4 cycles of nivolumab and doxorubicin, vinblastine, and dacarbazine (N-AVD) or sequential treatment with 4 doses of nivolumab, 2 cycles of N-AVD, and 2 cycles of AVD at standard doses, followed by 30-Gy involved-site radiotherapy.

Main Outcomes and Measures  Complete remission (CR) rate after study treatment, aiming at excluding a CR rate of 80% or lower via a 2-sided 95% CI for each treatment group.

Results  Of 109 patients included in this study, 65 (59.6%) were women, and the median (range) age was 27 (18-60) years. At interim staging after 2 cycles of N-AVD or 4 doses of nivolumab monotherapy, 54 of 54 (100%) and 49 of 51 (96%) response-eligible patients, respectively, achieved an objective response, with CR in 47 (87%) and 26 (51%) patients, respectively. Among 101 patients eligible for primary end point analysis, 46 of 51 (90%; 95% CI, 79%-97%) patients receiving concomitant therapy and 47 of 50 (94%; 95% CI, 84%-99%) patients receiving sequential therapy achieved CR after study treatment. With a median follow-up of 13 months, 12-month progression-free survival was 100% for patients receiving concomitant treatment and 98% (95% CI, 95%-100%) for patients receiving sequential therapy.

Conclusions and Relevance  Both strategies combining nivolumab and AVD are feasible and resulted in high remission rates. Despite narrowly missing the efficacy benchmark in the concomitant group, the excellent 12-month progression-free survival and the unexpectedly high CR rate after 4 doses of nivolumab monotherapy warrant further evaluation of this approach in the first-line treatment of patients with early-stage cHL.

Trial Registration  ClinicalTrials.gov Identifier: NCT03004833

Risk-adapted first-line treatment of classic Hodgkin lymphoma (cHL) has markedly improved cure rates and long-term survival over the past several decades. For patients with early-stage unfavorable cHL, combined-modality treatment with 4 cycles of chemotherapy and consolidating 30-Gy involved-site radiotherapy (IS-RT) is the standard of care.¹ The chemotherapy regimen ABVD (doxorubicin, bleomycin, vinblastine, and dacarbazine) is a well-established regimen with good tolerability yet insufficient long-term progression-free survival (PFS).² More intensive approaches including 2 cycles of eBEACOPP (dose-escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisolone) improve 5-year PFS^3,4 but also increase short-term and potentially long-term toxic effects on patients.^5-8 Thus, reducing treatment-related adverse effects while maintaining the very high primary cure rate of intensified chemotherapy is an ongoing quest.

A crucial role of programmed cell death protein 1 (PD-1)/programmed cell death ligand interaction was described in cHL,⁹ and Hodgkin and Reed-Sternberg cells frequently express programmed cell death 1 ligand 1 either by chromosome 9p24.1 amplification or by Epstein-Barr virus infection.¹⁰ The PD-1 axis contributes to the composition of the protective niche provided by the cHL tumor microenvironment.^11,12 The anti–PD-1 antibodies nivolumab and pembrolizumab have been approved for relapsed cHL because of a favorable safety profile and high objective response rates.^13,14 Long-term follow-up revealed an unprecedented duration of response and sustained PFS in patients achieving complete remission (CR).^14,15

Immunogenic effects involved in sustained response to conventional chemotherapy and radiotherapy are increasingly understood.¹⁶ We thus hypothesized that combining the well-established chemotherapeutic backbone of AVD (doxorubicin, vinblastine, and dacarbazine; the omission of bleomycin mitigates the risk of pneumonitis) with nivolumab would yield additive effects at least. The Nivolumab and AVD in Early-Stage Unfavorable Classic Hodgkin Lymphoma (NIVAHL) trial was designed to test this. Because the optimal therapeutic sequence is unknown, we conducted this multicenter phase 2 randomized clinical trial investigating either a fully concomitant or a sequential immunochemotherapy, both followed by consolidative 30-Gy IS-RT. We aimed to demonstrate efficacy of the 2 regimens and to describe their safety and feasibility in patients with early-stage unfavorable cHL.

This open-label, multicenter phase 2 randomized clinical trial was open for recruitment at 35 hospitals and specialist practices in Germany; 28 trial centers enrolled patients (eMethods 2 in Supplement 1). Patients were randomized between 2 treatment groups to be analyzed separately. This investigator-initiated trial was designed by the German Hodgkin Study Group (GHSG) steering committee, sponsored by the University of Cologne, and supervised and approved by the Ethics Committee of the Faculty of Medicine of the University of Cologne. We recruited patients aged 18 to 60 years with newly diagnosed, histology-proven cHL of early-stage unfavorable risk by GHSG criteria (ie, clinical stage I-II with 1 or more of the following risk factors: large mediastinal mass [one-third or more of the maximal thoracic diameter], extranodal disease, elevated erythrocyte sedimentation rate [50 mm/h or more without B symptoms, 30 mm/h or more with B symptoms], or 3 or more involved nodal areas, excluding patients with stage IIB who had large mediastinal mass and/or extranodal disease; eMethods 1 in Supplement 1). Primary biopsies were centrally reassessed. All patients provided written informed consent before study entry according to the Good Clinical Practice guidelines of the International Conference on Harmonisation and national regulations. The trial protocol can be found in Supplement 2.

Before starting treatment, patients were centrally randomized (1:1) between 2 parallel groups: those receiving concomitant therapy with nivolumab and AVD (N-AVD) on day 1 and 15 of each 28-day cycle for 4 cycles, or sequential therapy starting with 4 infusions of nivolumab in 14-day intervals followed by two 28-day cycles N-AVD and two 28-day cycles AVD. In both groups, consolidating 30-Gy IS-RT was scheduled after systemic treatment (eFigure 1 in Supplement 1). Randomization was performed centrally before treatment was begun using the minimization method including a random component, stratified according to age (<45 years vs ≥45 years) and sex.

Procedures are described in detail in eMethods 1 in Supplement 1: AVD was administered at standard doses, as previously described,¹⁷ and nivolumab was administered at a flat dose of 240 mg in 14-day intervals, on the days of AVD chemotherapy during concomitant applications. Systemic therapy was delayed or discontinued in case of prespecified higher-grade adverse events (AEs). Both nivolumab and N-AVD were initially administered without antiphlogistic or steroid-containing antiemetic prophylaxis. To improve tolerability and potentially mitigate high-grade early-onset immune-related AEs, mandatory dexamethasone-based prophylaxis was put into effect from December 2017 onward.

Positron emission tomography (PET) and computed tomography (CT)-based response assessments were conducted after the fourth nivolumab dose (ie, after the second N-AVD cycle in the concomitant group) and after the end of systemic treatment. Another CT-based restaging was scheduled for 4 to 6 weeks after completion of IS-RT. In case of residual lesions that had not been considered negative on PET results in an earlier restaging, another PET/CT was scheduled for 3 months after the end of IS-RT. Response assessment was completed centrally by a multidisciplinary panel of experts from medical oncology, nuclear medicine, and radiation oncology. Response criteria were adapted from 2007 International Working Group criteria (eMethods 1 in Supplement 1). An independent data-monitoring board reviewed data on a regular basis.

The primary end point of this study was the CR rate after end of study treatment, including systemic treatment with nivolumab and AVD and IS-RT, if performed. Treatment-related morbidity was analyzed as a key secondary end point; this was defined as any organ-related AE in the categories cardiac, gastrointestinal, hepatobiliary, nervous system, kidney and urinary, respiratory, thoracic and mediastinal disorders of Common Terminology Criteria for Adverse Events (CTCAE) grade 3 or 4, or anemia, thrombocytopenia, or infection of CTCAE grade 4. Further secondary end points included PFS (defined as the time from randomization until progression, relapse, or death from any cause, or censored at the date of last information on the disease status), overall survival (OS; defined as time from randomization until death from any cause, or censored at the date of last information on the patient being alive), early response to systemic therapy, and treatment-related AEs of CTCAE grades 3 and 4.

The primary objective of this study was to demonstrate the efficacy of the 2 treatment strategies. The Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines were followed. The 2 independent treatment groups were analyzed separately. The trial was designed to exclude a CR rate of 80% or less via a 2-sided 95% CI with a power of 90% for each treatment group, assuming the true CR rates were around 95%. Patients with severe protocol deviations (ie, inclusion criteria violation or fewer than 3 full cycles of AVD or fewer than 4 doses of nivolumab, unless owing to progressive disease) were not eligible for the primary end point analysis. We recruited 110 patients to perform the primary end point analysis with at least 48 eligible patients per treatment group. Except for the primary end point, all analyses were completed according to the intention-to-treat principle. We analyzed secondary end points descriptively, using the Kaplan-Meier method for time-to-event end points. We preplanned exploratory subgroup analyses for the primary end point in the pooled treatment groups and for early response to nivolumab in the sequential treatment group using univariate logistic regression. Subgroups were defined by potentially relevant clinical characteristics, such as sex, body mass index, a white-blood cell/eosinophil risk score,¹⁸ vitamin D deficiency,¹⁹ administration of corticosteroid prophylaxis, from start of study therapy, and highest CTCAE grade. We used SAS, version 9.4 (SAS Institute) for all analyses.

We enrolled 110 patients with cHL between April 2017 and October 2018. Expert pathology review confirmed the cHL diagnosis in all but 1 patient, resulting in an intention-to-treat population of 109 patients (Figure 1). Of 109 patients included in this study, 65 (59.6%) were women, with a median (range) age of 27 (18-60) years. Baseline characteristics are displayed in Table 1.

Systemic therapy with N-AVD was feasible: 44 (80%) patients in the concomitant treatment group and 49 (91%) patients in the sequential treatment group received the preplanned 8 infusions of nivolumab, and 50 (91%) patients in the concomitant treatment group and 48 (89%) patients in the sequential treatment group received 4 full cycles of N-AVD. Single agents including nivolumab were rarely discontinued from study treatment (eFigure 2 in Supplement 1). The mean relative total doses of all agents were greater than 90% in both groups. Involved-site radiotherapy was administered to 51 patients (93%) and 48 patients (89%), respectively.

Severe protocol deviations as defined above occurred in 4 patients in each treatment group (Figure 1); reasons were treatment-related AEs (n = 5), patients’ wishes (n = 2), and incorrect allocation to the early-stage unfavorable risk group (n = 1). Adverse events leading to early discontinuation were 1 case each of encephalitis, nivolumab-associated infusion-related reaction, proctosigmoiditis, pneumonitis, and polyneuritis. Of affected patients, 4 discontinued therapy before any response assessment and are not included in efficacy analyses. Except for the 2 patients who discontinued therapy at their own wishes, all patients with these protocol deviations received further conventional treatment and achieved a sustained CR thereafter. Adverse events were fully reversible after discontinuation of nivolumab, except for the grade 4 proctosigmoiditis, which only resolved to intermittent colitis grade 1 at last follow-up.

At first interim restaging after 2 cycles N-AVD or 4 doses nivolumab, 54 of 54 patients (100%) in the concomitant group and 49 of 51 patients (96%) in the sequential group achieved objective response, defined as either partial remission or CR, with CR observed in 47 patients (87%) and 26 patients (51%), respectively. At the end of systemic treatment, objective response was observed in 54 of 54 patients (100%) and 50 of 51 patients (98%), with CR in 45 of 54 patients (83%) and 43 of 51 patients (84%), respectively (Figure 2A). Three patients with CR at first restaging were rated as having partial remission after completion of systemic treatment due to a flare in Deauville score (DS) from DS3 to DS4, which was still present in 2 of these patients in the final restaging. One patient developed histologically proven primary progression during 4 doses nivolumab in the sequential treatment group and achieved an ongoing CR after administration of 2 cycles eBEACOPP, 2 cycles ABVD, and 30-Gy IS-RT.

Prespecified exclusion of patients with severe protocol deviation resulted in 101 patients being evaluable for the primary end point. Complete remission after the end of study treatment was achieved in 46 of 51 patients (90%; 95% CI, 79%-97%) in the concomitant treatment group and 47 of 50 patients (94%; 95% CI, 84%-99%) in the sequential treatment group, respectively. Regarding the 95% CIs, the 80% benchmark for efficacy was narrowly missed in the concomitant treatment group and met in the sequential treatment group (eTable 1 in Supplement 1). Objective response after study treatment was observed in 51 of 51 patients (100%) and 49 of 50 patients (98%), respectively.

In addition to 1 case of early disease progression described above, 7 patients did not achieve centrally measured CR by the predefined PET-based remission criteria. Because all 7 patients were considered in remission by the treating physicians, none received further therapy. With continued follow-up and repeated imaging at investigators’ discretion, 6 of 7 patients have yet been judged as having CR at a later time point, and none experienced progressive disease or relapse (eFigure 3 in Supplement 1).

Prespecified exploratory analysis of potential risk factors identified deficient baseline vitamin D and a combination of high white blood cell count and low relative eosinophil count to be associated with failure to achieve CR after study treatment. There was also a trend toward impaired response to nivolumab monotherapy in patients with vitamin D deficiency (eTables 2 and 3 in Supplement 1).

With a median (range) follow-up of 14 (6-27) months in the concomitant group and 13 (8-28) months in the sequential treatment group, 12-month PFS was 100% for concomitant, and 98% (95% CI, 95%-100%) for sequential (Figure 2B). Overall survival at 12 months was 100% in both treatment groups (eFigure 4 in Supplement 1).

Any treatment-related AE was documented in 54 patients (98%) in the contomitant therapy group and 53 patients (98%) in the sequential therapy treatment group, with treatment-related AEs grade 3 or greater reported in 42 patients (76%) and 43 patients (80%) in the concomitant and sequential groups, respectively. Hematologic AEs grade 3 or greater, accounting for the vast majority of higher-grade AEs, were observed in 39 patients (71%) and 35 patients (65%). Infections of grade 3 or greater were rare (3 [5%] and 1 [2%], respectively), and febrile neutropenia was observed in 4 of 55 (7%) patients in the concomitant treatment group and 4 of 54 (7%) patients in the sequential treatment group. Organ-related AEs of grade 3 or greater occurred in 13 patients (24%) and 16 patients (30%), respectively, with hepatobiliary/pancreas, gastrointestinal, and skin AEs accounting for most cases. Infusion-related reactions were rare (1 [2%] and 2 [4%]). Further toxic effects reported frequently as text entries without grading or causality included fatigue, muscle/bone/joint pain, abnormal laboratory test findings, malaise, thyroid disorders, and headache/migraine. Treatment-related morbidity was reported in 10 patients (18%) and 12 patients (22%), respectively, with organ-related AEs accounting for all cases of treatment-related morbidity. During nivolumab monotherapy in the sequential treatment group, 47 patients (87%) had any AE; the highest grades were 1 and 2 in 36 patients (67%), 3 in 10 patients (19%), and 4 in 1 (2%) patient (eFigure 6 in Supplement 1). Serious AEs were reported in 21 patients (38%) and 15 patients (28%) in the concomitant and sequential group, respectively, with all but 1 AE graded serious owing to hospitalization. Details on AEs, serious AEs, and treatment-related morbidity are summarized in Table 2. Details on prophylactic measures are summarized in Table 3. Data on AEs are also presented in eTable 4 and eFigures 5 and 6 in Supplement 1.

With ongoing follow-up, data on potential late AEs are only emerging. Regarding potential immune-related late AEs, 7 cases of hypothyroidism of grade 2 or lower and 1 case each of intermittent pneumonitis of grade 2 and intermittent colitis of grade 1 have been reported thus far.

The investigator-initiated NIVAHL phase 2 randomized clinical trial is the first trial to our knowledge evaluating a first-line treatment of early-stage unfavorable cHL based on anti–PD-1. Patients received either fully concomitant or sequential therapy with nivolumab and AVD followed by consolidative 30-Gy IS-RT. We found both treatment strategies to be feasible, associated with noteworthy yet manageable toxic effects, and highly effective. An unexpectedly high proportion of patients achieved CR with nivolumab monotherapy, and 12-month PFS was 100% for concomitant, and 98% (95% CI, 95%-100%) for sequential treatment, respectively. Overall survival at 12 months was 100% in both treatment groups.

The add-on cohort D of the pivotal phase 2 randomized clinical trial of nivolumab in patients with cHL (CheckMate 205 study)²⁰ reported anti-PD-1-based first-line treatment in advanced-stage cHL. Among 51 patients who received 4 doses of nivolumab followed by 6 cycles of N-AVD, 1 patient from each arm discontinued study treatment during nivolumab monotherapy due to AEs and progressive disease. An additional 5 patients discontinued treatment during N-AVD administration, mostly owing to AEs, and 1 treatment-related death occurred. Compared with the present study, grade 3 or 4 AEs were also mainly hematological, but overall observed less frequently, and organ-related AEs were mostly of grades 1 or 2.

Incidence of treatment-related AEs with nivolumab alone in the NIVAHL study (grade 3 toxic effects or greater, 21%) was similar to data from the pivotal nivolumab trial by Younes et al¹³ in patients with relapsed cHL (grade 3 toxic effects or greater, 25%). Type and frequency of AEs are largely within the previously described safety profile of AVD or nivolumab. However, we observed relevant immune-related organ-related AEs earlier than previously described, potentially due to preserved immunocompetence in treatment-naive patients. Most serious AEs were attributed to the combination of N-AVD, and fewer than 10% were caused merely by either AVD or nivolumab. Of note, we did not observe any measurable difference regarding occurrence of AEs or efficacy after establishing the antiphlogistic prophylaxis during trial conduct. In summary, treatment with N-AVD was associated with noteworthy yet manageable immune-related toxic effects beyond AVD chemotherapy alone.

Disease control with N-AVD in terms of response rates and early PFS was higher in patients with early-stage unfavorable cHL than in those with advanced-stage cHL, especially considering the unexpectedly high CR rate of 51% after 4 doses nivolumab in the present trial compared with only 18% in cohort D of the CheckMate 205 study.²⁰

Similar to the present trial, most patients in cohort D from the CheckMate 205 study²⁰ not in CR at the end of treatment by central review were deemed to be in CR by the treating physician and did not receive additional therapy (7 of 12 patients in the CheckMate 205 study; 7 of 7 patients in NIVAHL). Importantly, 6 of 7 nonresponders in the present study were deemed to be in CR at last follow-up. These observations highlight the emerging issues regarding PET-based response assessment in the context of checkpoint inhibition. Immune activation by checkpoint inhibition may result in PET findings mimicking residual disease due to inflammatory persistent metabolic activity of DS4. Similar observations with sustained DS4 despite marked reductions in tumor size were made in an interim analysis of an ongoing phase 2 trial (NCT03226249) investigating sequential first-line therapy with the anti–PD-1 antibody pembrolizumab and AVD chemotherapy in patients with cHL.²¹ Modifications of response criteria have been proposed (eg, introducing the concept of an indeterminate response).²² Unfortunately, these updates and other immune-related response criteria²³ are tailored to an indefinite immunotherapy and thus are not helpful in determining CR after fixed-duration first-line treatment. These issues critically influence the validity of PET-measured CR as a primary end point for investigating anti–PD-1 cHL therapy, as it likely causes underestimation of its true efficacy.

From a patient’s perspective, achieving a sustained response and ideally a cure with first-line therapy is of highest priority, and PFS is therefore considered a more robust end point than CR rate.²⁴ Definition of CR and its limited association with long-term PFS is also an issue regarding conventional chemotherapy. In the GHSG HD14 trial,³ 5-year PFS was considerably better with 2 cycles eBEACOPP and 2 cycles ABVD than 4 doses ABVD despite identical CT-based CR rates of 95.4%. In 2016 a trial²⁵ in 30 patients with early-stage unfavorable cHL reported a PET-based CR rate of 100% after 4 doses BV-AVD and 30-Gy IS-RT, only after 2 cases that were initially judged to be in PR due to DS4 were retrospectively rated as being in CR months later during follow-up. With 2 cases of primary progressive disease, 12-month PFS was 93%, with a median follow-up of 19 months. Follow-up within the present trial is still limited, but the 12-month PFS of 100% for the concomitant group and 98% for the sequential group is outstanding. Taking into account the discussed pitfalls concerning PET after anti–PD-1 therapy and the favorable clinical course of the 7 patients formally failing to achieve a CR, we conclude very good efficacy of both treatment strategies despite marginally missing the primary end point in the concomitant therapy group.

The high but stagnating CR rates from first to final restaging in the concomitant treatment group and the high early CR rate after nivolumab monotherapy suggest that a more individualized treatment strategy might be promising. Because all patients in the sequential treatment group received further chemotherapy and radiotherapy, we cannot estimate the long-term efficacy of anti–PD-1 therapy alone in first-line treatment of patients with cHL. However, a PET-guided approach might allow for a chemotherapy-free and radiotherapy-free fixed-duration anti–PD-1 treatment in selected patients with early-stage cHL who have early and sustained negative disease activity seen on results of PET scan. This should be evaluated in a phase 2 trial, ideally also including patients with comorbidities or of older age who are in need of innovative treatment. Older patients who might disproportionally benefit from the effective yet tolerable anti–PD-1-based therapy were not included in the present trial, and beyond the 6 patients older than 60 years enrolled in the CheckMate 205 study,²⁰ feasibility and efficacy in older patients remain largely unknown.

This study had several limitations. In addition to the discussed issues concerning CR rate as primary end point, the predefined exclusion of patients with severe protocol deviations from the primary end point analysis might introduce some uncertainties to the estimation of final response rates. Importantly, all other analyses were completed according to the intention-to-treat principle. Another potential limitation of this analysis is the limited follow-up so far. While we conclude excellent short-term disease control, the long-term efficacy and curative potential as well as potential late effects of both treatment strategies are to be determined with the designated longer follow-up of 3 years within the NIVAHL trial.

In conclusion, both strategies combining nivolumab and AVD are feasible, highly effective, and result in excellent 12-month PFS. We observed early and sustained responses in both treatment groups and an unexpectedly high interim CR rate even after 4 doses of nivolumab alone. Anti–PD-1-based first-line treatment of early-stage cHL warrants further evaluation within future trials, comparing a fully concomitant treatment with standard of care or an individualized approach with sequential anti–PD-1 treatment guided, eg, by interim PET.

Accepted for Publication: February 2, 2020.

Corresponding Author: Andreas Engert, MD, Faculty of Medicine and University Hospital of Cologne, Department I of Internal Medicine, Center for Integrated Oncology Aachen Bonn Cologne Düsseldorf, University of Cologne, Kerpener St 62, D-50924 Cologne, Germany (a.engert@uni-koeln.de).

Published Online: April 30, 2020. doi:10.1001/jamaoncol.2020.0750

Author Contributions: Ms Goergen and Dr Engert had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr Engert was the principal investigator.

Concept and design: Bröckelmann, Goergen, Fuchs, von Tresckow, Borchmann, Engert.

Acquisition, analysis, or interpretation of data: Bröckelmann, Goergen, Keller, Meissner, Ordemann, Halbsguth, Sasse, Sökler, Kerkhoff, Mathas, Hüttmann, Bormann, Zimmermann, Mettler, Fuchs, von Tresckow, Baues, Rosenwald, Klapper, Kobe, Borchmann, Engert.

Drafting of the manuscript: Bröckelmann, Goergen, Engert.

Critical revision of the manuscript for important intellectual content: Bröckelmann, Goergen, Keller, Meissner, Ordemann, Halbsguth, Sasse, Sökler, Kerkhoff, Mathas, Hüttmann, Bormann, Zimmermann, Mettler, Fuchs, von Tresckow, Baues, Rosenwald, Klapper, Kobe, Borchmann, Engert.

Statistical analysis: Goergen.

Obtained funding: Bröckelmann, Engert.

Administrative, technical, or material support: Bröckelmann, Meissner, Ordemann, Sasse, Sökler, Kerkhoff, Fuchs, von Tresckow, Baues, Rosenwald, Klapper, Kobe.

Supervision: Bröckelmann, von Tresckow, Borchmann, Engert.

Conflict of Interest Disclosures: Dr Bröckelmann reports grants from Bristol-Myers Squibb (BMS) during the conduct of the study; grants from Merck Sharp & Dohme (MSD) and Affimed Therapeutics; and grants, personal fees, and nonfinancial support from BMS and Takeda outside the submitted work. Dr Keller reports personal fees and travel support, support in an advisory role, and speaker’s honorary from BMS during the conduct of the study, as well as personal fees from BMS outside the submitted work. Dr Meissner reports travel grants from MSD, BMS, Takeda, Hexal, and Celgene outside the submitted work. Dr Hüttmann reports personal fees from Celgene, Takeda, Roche, Gilead Sciences, and Pfizer, and travel support from Celgene, Takeda, and Roche outside the submitted work. Dr Zimmerman reports personal fees and nonfinancial support from Takeda and nonfinancial support from Pfizer, Novartis, Gilead Sciences, Roche, BMS, Astellas, and MSD outside the submitted work. Dr Fuchs reports grants from BMS during the conduct of the study and personal fees from Takeda, Amgen, Celgene, and BMS outside the submitted work. Dr von Tresckow reports grants and nonfinancial support from BMS during the conduct of the study; personal fees from Amgen, Pfizer, Gilead Sciences, and Roche; grants, personal fees, and nonfinancial support from MSD and Takeda; and grants and nonfinancial support from Novartis outside the submitted work. Dr Klapper reports grants from Roche, Amgen, Takeda, and Regeneron paid to his institution outside the submitted work. Dr Borchmann reports grants from BMS during the conduct of the study. Dr Engert reports grants and nonfinancial support from BMS during the conduct of the study, and personal fees from Takeda, BMS, and MSD outside the submitted work. No other disclosures were reported.

Funding/Support: Bristol-Myers Squibb provided the study drug nivolumab as well as financial support.

Role of the Funder/Sponsor: Bristol-Myers Squibb had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Meeting Presentation: Data were presented in part as an oral abstract at the 61st Annual Meeting of the American Society of Hematology; December 7, 2019; Orlando, Florida.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We thank all participating patients, their families, and their treating physicians for accepting the theoretically increased risk of treatment failure in this trial to help improve treatment for future patients with classic Hodgkin lymphoma. We thank all participating investigators and their teams at the German Hodgkin Study Group NIVAHL trial centers. We thank the data monitoring board, which includes Alexander Fosså, MD (Department of Oncology, Oslo University Hospital, Oslo, Norway), Dominik Wolf, MD (Internal Medicine V, Medical University of Innsbruck, Innsbruck, Austria), Josée M. Zijlstra, MD (Department of Hematology, Cancer Centre Amsterdam, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands), and Heinz Haverkamp, Dipl.-Math. (Institute of Medical Statistics and Computational Biology, University of Cologne, Cologne, Germany). We thank Daria Siury, MSc, and Daniel Armbrust, emplo, for their excellent administrative support and data management, respectively, as employees of the GHSG central office.