Clinical prognostication in the SWOG 1826 randomized clinical trial of nivolumab-AVD versus brentuximab-AVD: Performance of the advanced-stage Hodgkin lymphoma international prognostic index (A-HIPI)
Bibliographic record
Abstract
Abstract Background The most widely used prognostication tool in classical Hodgkin lymphoma (cHL) historically has been the International Prognostic Score (IPS), developed to predict survival at 5 years (yrs) in patients (pts) with newly diagnosed advanced-stage (AS) HL. However, performance of the IPS has been suboptimal among contemporarily treated pts. The Hodgkin Lymphoma International Study for Individual Care (HoLISTIC) consortium developed a multivariable prediction model (the A-HIPI) for primarily ABVD-treated HL pts ages 18-65 yrs (Rodday. JCO 2023), utilizing continuous data values for 5 of 7 variables (ie, age, stage, albumin, hemoglobin and lymphocyte count). The A-HIPI demonstrated superior discrimination and calibration for progression-free survival (PFS) and overall survival (OS) versus the historic IPS. The c-statistics for 5-year PFS were 0.60 and 0.59 in the initial development (clinical trials) and validation (real world registries) cohorts, respectively, and 0.60-0.65 in subsequent validation analyses among Nordic and Brazilian external cohorts of AS HL pts treated with conventional chemotherapy regimens (Jørgensen. JCO Informatics 2024; Buccheri. BJHaem 2025). The phase 3 randomized SWOG S1826 study, the largest National Clinical Trials Network study of AS cHL, compared nivolumab with doxorubicin, vinblastine, and dacarbazine (N+AVD) with brentuximab vedotin with doxorubicin, vinblastine, and dacarbazine (BV+AVD) in pts 12 yrs or older with AS newly diagnosed cHL. N+AVD resulted in longer PFS than BV+AVD, resulting in a new standard of care incorporating novel checkpoint inhibitors. With the evolution of novel treatments, it is imperative to validate existing prognostic models against contemporary therapeutic outcomes for more precise clinical risk stratification. Thus, we sought to validate the A-HIPI as a clinical prognostic tool in the S1826 randomized trial, with a focus on adults aged ≥18 yrs. Methods S1826 was a phase 3, multicenter, open-label, randomized trial including pts 12 yrs of age or older with AS newly diagnosed cHL. The primary end point was PFS. Median follow up time is 3.1 yrs (range: 0 to 5.3 yrs). For the validation, all model parameters defining the A-HIPI were locked prior to S1826 application as previously defined. Hazard ratio (HR) estimation was conducted using Cox Regression and Harrell's index (C-index) was calculated for both IPS and A-HIPI models. Results A total of 731 pts 18 yrs or older were included with 367 receiving N+AVD and 364 receiving BV+AVD. Median age was 32 (range: 18-83) yrs; nodular sclerosis was the most common histology. The overall estimated 3-yr PFS is 86% (95% CI: 83%-88%) in this cohort. Using the A-HIPI, the predicted 3yr PFS by quartile (Q) of predicted risk was 92% for Q1 and Q2 (95%CI 86%-95%), 81% for Q3 (95%CI 74%-86%), and 79% for Q4 (73%-85%). Kaplan Meier plots stratified above and below median A-HIPI (n=365 and 366 pts) showed 3 yr PFS of 92% (95% CI 88%-94%) and 80% (95% CI 76%-84%), respectively, P<.001, and 3 yr PFS stratified by IPS 0-3 (n=509) versus 4-7 (n=222) were 88% (95% CI 84%-90%) & 82% (95% CI 76%-87%), respectively, P=.03. The c-statistic for the A-HIPI was .63 (95% CI .58-.68) and .59 for IPS-7 (95% CI .54 -.65); the difference was compared (P=.10, 95% CI .01-.07) using bootstrap resampling. The A-HIPI performed similarly in both treatment arms (c statistic: N+AVD .64 (95% CI .54-.72) and BV+AVD .62 (95% CI .54-.69)). Additionally, a multivariable Cox model with the A-HIPI and IPS showed superiority for the A-HIPI (A-HIPI HR 1.52 (95% CI 1.15-2.0) versus IPS HR 1.0 (95% CI .82-1.23)) with P=.003 and .97, respectively). Full analyses by treatment arm and calibration plots will be presented at the meeting. Conclusions The baseline A-HIPI clinical prediction tool performed well in the S1826 trial incorporating novel agents for adults with AS HL and was comparable to the original ABVD-treated cohort. Furthermore, the A-HIPI had superior prognostic performance compared to the IPS in S1826. Altogether, the A-HIPI is a treatment-agnostic prediction tool with superior capability at discriminating outcomes using both historical and novel targeted treatment strategies. This refined prediction model, which leverages continuous data values, is poised to serve as the new standard for risk stratification for adults with AS cHL. Funding NIH/NCI grant awards U10CA180888, U10CA180819, R01CA262265-04
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".