Abstract P136

Post-ASCT Brentuximab Vedotin (BV) maintenance in high-risk relapsed/refractory Hodgkin lymphoma in the era of novel salvage therapies: a bias-adjusted analysis of 4 LYSA real-world cohorts.

Introduction: AETHERA showed that Brentuximab Vedotin (BV) maintenance after autologous stem-cell transplantation (ASCT) improved PFS in BV-naïve high-risk relapsed/refractory classical Hodgkin lymphoma (R/R cHL). However, BV and PD-1 inhibitors before ASCT may reduce its added value. To address this question, we analyzed 4 French LYSA cohorts using methods accounting for treatment-selection confounding and immortal time bias.

Methods: Adult R/R cHL patients eligible for BV maintenance after ASCT were analyzed. BV maintenance was compared with observation. Propensity scores included demographic and relapse-risk factors; inverse probability weighting was applied. Immortal time bias from delayed BV initiation was addressed using a weighted Cox model with BV as a time-varying exposure and a clone-censor-weight analysis comparing BV initiation within 4 months versus no maintenance. Outcomes were PFS and OS from ASCT

Results: Among 304 patients, 256 with complete propensity-score covariates were analyzed. Median age was 31 years; 60% were male. At relapse, 65% had disseminated disease, 49% primary refractory disease, and 25% early relapse. Most had received BV before ASCT (n=216, 84%) and achieved CR before ASCT (n=217, 85%). Post-ASCT management was BV maintenance (n=165, 64%) or observation (n=91, 35%). After a median follow-up of 39 months, 2-year PFS/OS were 81%/96% with BV and 76%/95% with observation (p=0.086/0.98). After weighting, BV maintenance was not associated with improved outcomes. In the time-varying Cox model, BV was not associated with PFS (HR 0.80, 95% CI 0.46–1.41; p=0.45) or OS (HR 1.71, 95% CI 0.50–6.02; p=0.40). Among BV-exposed patients, PFS did not differ by maintenance use (p=0.34), whereas BV maintenance was associated with improved PFS in the small BV-naïve subgroup (n=38; p<0.001). In patients in CR before ASCT, the apparent unadjusted PFS benefit of BV maintenance (p=0.046) disappeared after adjustment and immortal-time bias correction (HR 0.77, 95% CI 0.42-1.42; p=0.41).

Conclusion: In 4 LYSA cohorts, the added value of post-ASCT BV maintenance appeared limited after prior exposure to novel agents when robust causal methods were applied. A potential benefit was mainly observed in the small BV-naïve subgroup, whereas no clear benefit was evident after BV-based salvage, particularly in patients achieving CR before ASCT. These findings support refinement of post-ASCT consolidation strategies beyond historical AETHERA criteria.