Background
International, multicenter, open-label, randomized, noninferiority phase III study of 417 randomized pediatric patients (≤18 years at diagnosis, 4–21 years at HSCT) with acute lymphoblastic leukemia in complete remission undergoing allogeneic HSCT, testing whether chemotherapy-based conditioning could replace total body irradiation (TBI).
Interventions and follow up
Arm A: Fractionated 12 Gy total body irradiation (TBI) + etoposide (n=212)
Arm B: Chemoconditioning with fludarabine, thiotepa, and either busulfan or treosulfan (n=201)
Primary endpoint: 2-year overall survival (noninferiority margin 8%)
mFollow up: 2.1 years
Arm B: Chemoconditioning with fludarabine, thiotepa, and either busulfan or treosulfan (n=201)
Primary endpoint: 2-year overall survival (noninferiority margin 8%)
mFollow up: 2.1 years
Results
2-yr OS: 91% (TBI; 95% CI 0.86–0.95) vs 75% (chemoconditioning; 95% CI 0.67–0.81), P<.0001 (TBI superior)
2-yr cumulative relapse: 12% vs 33%, P<.0001
2-yr treatment-related mortality: 2% vs 9%, P=.0269
2-yr cumulative relapse: 12% vs 33%, P<.0001
2-yr treatment-related mortality: 2% vs 9%, P=.0269
Adverse events
Common grade 3/4 (day 100, both arms): Cytopenia, mucositis, nausea, and infection
Mortality: Treatment-related mortality higher with chemoconditioning (9% vs 2%); a futility stopping rule was applied March 31, 2019 after chemoconditioning proved inferior, halting random assignment
Mortality: Treatment-related mortality higher with chemoconditioning (9% vs 2%); a futility stopping rule was applied March 31, 2019 after chemoconditioning proved inferior, halting random assignment
Conclusions
TBI plus etoposide produced significantly higher overall survival and lower relapse than chemoconditioning, establishing TBI-based conditioning as the standard for patients older than 4 years with high-risk ALL undergoing allogeneic HSCT.
Key Limitations
Median follow-up was short (2.1 years), limiting assessment of late TBI toxicities (growth, endocrine, second malignancy) that are the principal motivation for chemoconditioning. The chemoconditioning arm pooled two distinct backbones (busulfan or treosulfan), reducing interpretability. Restricted to children ≥4 years; results do not apply to infants. Early termination for futility limited the final sample size.
Clinical Context
FORUM established TBI/etoposide as the preferred conditioning for children >4 years undergoing allogeneic HSCT for ALL, and is reflected in international pediatric transplant practice. Chemoconditioning remains reserved for children <4 years or those with contraindications to TBI, where late radiation toxicity outweighs the relapse benefit. The trial reinforces the durable antileukemic potency of TBI despite ongoing efforts to develop radiation-free alternatives.