Background
Phase III RCT. 615 children (3–21 years) with average-risk medulloblastoma (M0, <1.5 cm² residual). Three-arm randomization: CSI dose (18 Gy vs 23.4 Gy) and posterior fossa boost vs tumor bed boost. Conducted 2003–2013 through COG. Primary question: can CSI dose be safely reduced to 18 Gy with maintained local control?
Interventions and follow up
Arm A: CSI 18 Gy + tumor bed boost to 54 Gy (reduced dose)
Arm B: CSI 23.4 Gy + posterior fossa or tumor bed boost to 54 Gy (standard dose)(2×2 factorial for CSI dose and boost volume)
Primary endpoint: EFS
mFollow up: 7.6 year
Arm B: CSI 23.4 Gy + posterior fossa or tumor bed boost to 54 Gy (standard dose)(2×2 factorial for CSI dose and boost volume)
Primary endpoint: EFS
mFollow up: 7.6 year
Results
5-yr EFS (18 Gy CSI): 82.5% vs 23.4 Gy 85.8%, HR 1.28, P=.047 — reduced dose inferior
5-yr OS: 85.9% (18 Gy) vs 90.5% (23.4 Gy), HR 1.37, P=.045
Tumor bed vs PF boost: No difference in EFS (HR 1.07, P=.58)
Spinal failure (18 Gy): Significantly higher than 23.4 Gy
5-yr OS: 85.9% (18 Gy) vs 90.5% (23.4 Gy), HR 1.37, P=.045
Tumor bed vs PF boost: No difference in EFS (HR 1.07, P=.58)
Spinal failure (18 Gy): Significantly higher than 23.4 Gy
Adverse events
Main adverse events: Neurocognitive: significantly better IQ preservation in 18 Gy arm at 3 years (mean FSIQ 97 vs 89, P<.001). Ototoxicity: grade ≥2 hearing loss similar between arms. Neuroendocrine: growth hormone deficiency in ~50% by 5 years regardless of arm.
Conclusions
Reducing CSI from 23.4 to 18 Gy resulted in inferior EFS and OS in average-risk medulloblastoma, despite better neurocognitive outcomes. Tumor bed boost is equivalent to posterior fossa boost and should be the standard. The trial confirmed 23.4 Gy CSI as the standard dose for average-risk disease.
Key Limitations
Key Limitations: Molecular subgrouping (WNT, SHH, Group 3, Group 4) was not incorporated prospectively — the 18 Gy arm inferiority may not apply equally to all molecular subgroups. WNT-activated tumors have excellent prognosis and may tolerate dose de-escalation. Photon-based therapy only; proton RT was not tested.
Clinical Context
ACNS0331 definitively closed the question of 18 Gy de-escalation in unselected average-risk medulloblastoma. Current trials (ACNS1422, SIOPE) are evaluating 18 Gy CSI specifically in WNT-activated medulloblastoma, where de-escalation may be safe. Tumor bed boost is now the standard at most centers.
References
References: Michalski JM et al, NEJM 2021 (ACNS0331)