Background
Phase III RCT. 396 patients with relapsed or refractory DLBCL (including PMBCL and transformed FL) after ≥1 prior rituximab-containing regimen. Conducted by GELA/HOVON. Designed to identify the optimal salvage regimen prior to autologous stem cell transplantation (ASCT) and to evaluate the prognostic impact of prior rituximab exposure.
Interventions and follow up
Arm A: R-ICE (rituximab 375 mg/m² D1 + ifosfamide 5,000 mg/m² CI D2 + carboplatin AUC 5 D2 + etoposide 100 mg/m²/d D1–3) × 3 cycles, then ASCT in responders
Arm B: R-DHAP (rituximab 375 mg/m² D1 + dexamethasone 40 mg/d D1–4 + cisplatin 100 mg/m² CI D1 + cytarabine 2×2,000 mg/m² D2) × 3 cycles, then ASCT in responders
Primary endpoint: Response rate allowing ASCT after 3 salvage cycles
Median follow up: 36 months
Arm B: R-DHAP (rituximab 375 mg/m² D1 + dexamethasone 40 mg/d D1–4 + cisplatin 100 mg/m² CI D1 + cytarabine 2×2,000 mg/m² D2) × 3 cycles, then ASCT in responders
Primary endpoint: Response rate allowing ASCT after 3 salvage cycles
Median follow up: 36 months
Results
Transplantation rate: 57.8% (R-ICE) vs 59.7% (R-DHAP) — not significant
ORR: 63.5% vs 62.8% — not significant
3-year PFS (transplanted): 26% vs 26% — not significant
3-year PFS by prior rituximab (all patients): 21% (prior-R) vs 47% (rituximab-naive) — HR 2.1, P<.001
ORR: 63.5% vs 62.8% — not significant
3-year PFS (transplanted): 26% vs 26% — not significant
3-year PFS by prior rituximab (all patients): 21% (prior-R) vs 47% (rituximab-naive) — HR 2.1, P<.001
Adverse events
R-ICE (grade ≥3): Neutropenia 73%, thrombocytopenia 58%, febrile neutropenia 27%
R-DHAP (grade ≥3): Renal toxicity 7% (cisplatin), GI toxicity 18%, thrombocytopenia 56%
Both arms: Mobilization failure ~15%; treatment-related mortality ~1% each
R-DHAP (grade ≥3): Renal toxicity 7% (cisplatin), GI toxicity 18%, thrombocytopenia 56%
Both arms: Mobilization failure ~15%; treatment-related mortality ~1% each
Conclusions
R-ICE and R-DHAP produced equivalent salvage outcomes in R/R DLBCL prior to ASCT. Prior rituximab exposure was the dominant prognostic factor, with a 2-fold increase in hazard among rituximab-exposed patients, fundamentally changing understanding of salvage outcomes in the modern era.
Key Limitations
All enrolled patients had received prior rituximab, so a contemporaneous rituximab-naive control was absent. About 40% of patients did not proceed to ASCT, limiting interpretation of the primary endpoint. The era (2000s) predates CAR-T, now preferred for eligible relapsed DLBCL. Randomization at relapse diagnosis — not at confirmed transplant eligibility — introduces selection bias.
Clinical Context
CORAL established R-ICE and R-DHAP as equivalent salvage platforms prior to ASCT, allowing either as standard. In current practice, randomized trials (ZUMA-7, TRANSFORM) have shown superiority of axi-cel and liso-cel over salvage chemo-ASCT in second-line chemosensitive DLBCL, displacing the CORAL paradigm (per ESMO) for eligible patients. R-ICE/R-DHAP remain options for patients ineligible for or lacking access to CAR-T.
References