Abstract P012

KEYNOTE-204 final analysis: An open-label, phase 3 study of pembrolizumab (pembro) versus brentuximab vedotin (BV) in relapsed or refractory (R/R) classic Hodgkin lymphoma (cHL)

Background: At the second interim analysis of KEYNOTE-204 (NCT02684292), PFS was significantly improved with pembro vs BV in participants with R/R cHL (median PFS, 13.2 vs 8.3 months; HR, 0.65 [95% CI, 0.48-0.88]; P=0.0027 [P value threshold for significance, 0.0043]). We report results from the final analysis.

Methods: Eligible participants aged ≥18 years with R/R cHL who were post−autologous stem cell transplant (auto-SCT) or ineligible for auto-SCT, had measurable disease, and an ECOG PS of 0 or 1, were randomly assigned 1:1 to pembro 200 mg or BV 1.8 mg/kg IV Q3W for 35 cycles. Participants were stratified by prior auto-SCT (yes vs no) and status after 1L therapy (primary refractory vs relapsed <12 months vs relapsed ≥12 months after 1L). Dual primary end points were PFS per IWG 2007 criteria by BICR including clinical and imaging data after auto-SCT or allogeneic SCT (allo-SCT) and OS. ORR per IWG by BICR and safety were secondary. DOR per IWG by BICR was exploratory.

Results: 304 participants were randomized and 300 participants were treated (pembro, n=148; BV, n=152). At data cutoff (Jan 14, 2026), median (range) follow-up was 97.6 months (90.6-113.9) for pembro and 97.7 months (90.1-114.3) for BV. Median PFS was 13.6 months for pembro and 8.2 months for BV (HR [95% CI], 0.73 [0.55-0.99]); median OS was 106.6 months and not reached (NR; HR [95% CI], 0.93 [0.66-1.33]), respectively (Figure). ORR (95% CI) was 65.6% (57.4-73.1; 40 complete responses, 59 partial responses) for pembro and 54.2% (46.0-62.3; 37 complete responses, 46 partial responses) for BV. Median (range) DOR was 24.9 months (0.0+ to 102.6) for pembro and 13.8 months (0.0+ to 96.2+) for BV. Treatment-related AEs occurred in 75.0% of participants for pembro and 77.0% for BV (grade 3-5, 19.6% and 25.0%). One participant in the pembro arm died due to treatment-related pneumonia. Subsequent therapies included auto-SCT (23.0% pembro, 24.3% BV), allo-SCT (14.2%, 12.5%), BV (47.0%, 13.1%), anti–PD-1 (21.9%, 49.0%), and other (62.9%, 65.4%).

Conclusions: PFS benefit with pembro at the interim analysis was maintained at the final analysis. OS was not statistically significantly improved with pembro vs BV. The converging OS Kaplan-Meier curves are likely due to receipt of effective subsequent therapies, including auto-SCT, allo-SCT, or anti–PD-1 agents. Results illustrate the limitations of using OS as a primary end point in cancers with rapidly changing therapeutic landscapes.