AUTHORS
Justin Ferdinandus, MD, Alex Francisco Herrera, MD2, Janina Jablonski, MSc3, Hongli Li, MS4, Michael Leblanc, MD5, Helen Kaul6, Sonali Smith, MD, FASCO7, Michael Fuchs8, Peter Borchmann, MD9, Jonathan W. Friedberg, MD10
AFFILIATIONS
1 German Hodgkin Study Group, Cologne, Germany, University Hospital Cologne, Cologne, Germany
2 City of Hope Comprehensive Cancer Center, Duarte, California, United States
3 German Hodgkin Study Group, Cologne, Germany, University Hospital Cologne, Cologne, Germany
4 Fred Hutchinson Cancer Center, Seattle, Washington, United States, SWOG Statistics and Data Management Center, Seattle, Washington, United States
5 Fred Hutchinson Cancer Center, Seattle, Washington, United States, SWOG Statistics and Data Management Center, Seattle, Washington, United States
6 German Hodgkin Study Group, Cologne, Germany, University Hospital Cologne, Cologne, Germany
7 University of Chicago, Chicago, Illinois, United States
8 German Hodgkin Study Group, Cologne, Germany, University Hospital Cologne, Cologne, Germany
9 German Hodgkin Study Group, Cologne, Germany, University Hospital Cologne, Cologne, Germany
10 Wilmot Cancer Institute, University of Rochester Medical Center, Rochester, New York, United States
ABSTRACT
Introduction: PET-guided BrECADD (GHSG HD21) and nivolumab-AVD (N-AVD; SWOG S1826) are both standards of care for newly-diagnosed advanced-stage classic Hodgkin lymphoma (cHL), but never been compared head-to-head. GHSG and SWOG therefore performed a joint analysis of individual patient data (IPD) from both trials, to compare indirectly efficacy and tolerability and to explore consistency of effects across relevant subgroups.
Methods: Patients randomized to BrECADD in HD21 and to N-AVD in S1826 were analyzed were analyzed according to their randomized treatment assignment within each trial (centrally confirmed cHL histology, no prior malignancy). To harmonize eligibility, analyses were restricted to HIV-negative patients aged 18-60 years with stage III or IV, leaving 620 and 310 patients from HD21 and S1826, respectively. Baseline imbalances were addressed by inverse probability of treatment weighting (IPTW) with stabilized weights to estimate average treatment effects; the propensity scores were estimated by logistic regression using age, sex, clinical stage and the four laboratory International Prognostic Score (IPS) categories (albumin, hemoglobin, leukocytes, lymphocytes). Progression-free (PFS), event-free (EFS) and overall survival (OS) were estimated by weighted Kaplan-Meier method and hazard ratios (HR) by weighted Cox regression with robust 95% confidence intervals (CI). Toxicity was assessed by weighted rates and compared by weighted logistic regression. Analyses were repeated in subgroups (female vs male, age <40 vs >=40 years, IPS 0-2 vs 3-7, stage III vs IV) and possible treatment/subgroup interaction assessed. Follow-up of HD21 was censored to align median observation time between trials; unweighted and uncensored analyses served as sensitivity analyses.
Results: All propensity score variables were closely balanced after weighting (standardized mean differences (SMD) <0.02); effective sample sizes were 606 (HD21) and 285 (S1826). The A-HIPI, though not part of the propensity score, was balanced after weighting as well (mean 0.749 with BrECADD (n=613) vs 0.753 with N-AVD (n=309); SMD 0.059). Treatment-related morbidity was more frequent after BrECADD (44% vs 11%; OR 6.6, 4.5-9.8), comprising grade 4 anemia, thrombocytopenia or infection (33% vs 2%) and grade 3/4 organ toxicity (19% vs 10%). Febrile neutropenia was 29% vs 6% (OR 6.2, 3.8-10.1) and grade 3/4 infection 22% vs 8% (OR 3.1, 2.0-4.8). Sensory (39% vs 38%; OR 1.03, 0.78-1.36) and motor neuropathy (4% vs 3%; OR 1.21, 0.58-2.51) were comparable, as was early treatment discontinuation (4% vs 5%; OR 0.65, 0.34-1.25). Fatal events with indicated treatment relation (TRM 3 pts vs 1 pt) occurred rarely. Second primary malignancies (2% vs 1%) were rare in both. Use of consolidative radiotherapy (RT) was 14% vs 1%. At 3 years, PFS was 94.9% (95% CI 93.1-96.7) after BrECADD and 90.6% (87.2-94.1) after N-AVD (difference 4.3%, 95% CI 0.4-8.2; HR 0.58, 0.35-0.97); EFS was 93.2% (91.2-95.3) vs 87.7% (83.8-91.6); HR 0.59 (0.38-0.92). OS was similar and high in both cohorts (98.8%,97.9-99.7 vs 98.9%, 97.8-100; HR 1.12, 0.30-4.19). Twelve deaths occurred in total, one was HL-related in each group. Absolute 3-year PFS differences between BrECADD and N-AVD varied with baseline risk: 1.2 percentage points (-2.7 to 5.1) for IPS 0-2 vs 7.6 (1.1 to 14.0 ) for IPS 3-7, respectively. IPS 3-7 was a prognostic factor for PFS events in both cohorts. Risk subgroups based on A-HIPI will be presented at the conference. Absolute PFS differences between BrECADD and N-AVD were similar for age <40 with 4.0 percentage points vs 4.3 for >40 years; 4.8 for stage III vs 4.0 for stage IV.
Conclusions: In this joint IPD IPTW-based comparison of adults aged 18-60 years with stage III/IV cHL, 3-year PFS and EFS were higher after PET-guided BrECADD. Larger absolute PFS differences were observed in patients with IPS 3-7 while patients with IPS 0-2 had similar PFS outcomes. Acute tolerability profiles differ with more severe acute hematologic and infectious episodes with BrECADD. Although residual confounding and impact of consolidative RT on differences in survival outcomes cannot be excluded and interpretation of long-term toxicity is limited by different adverse event reporting frameworks, this data may support shared decision-making for patients with newly diagnosed AS-cHL in the context of risk features and balancing efficacy and toxicity.
DISCLOSURES
Justin Ferdinandus, MD: Takeda - Honoraria; Pfizer - Consultancy (Includes expert testimony)