18-Gy craniospinal irradiation remains unsupported for unselected standard-risk medulloblastoma

Pooled five-year disease-control survival was 81.0% after 23.4-Gy CSI versus 71.3% after 18 Gy, supporting caution with further dose reduction.

KEY POINTS

  • This systematic review and meta-analysis included 15 studies evaluating craniospinal irradiation (CSI) combined with chemotherapy for pediatric standard-risk or average-risk medulloblastoma. The main contemporary analysis focused on 18-Gy versus 23.4-Gy CSI, excluding historical high-dose and intermediate-dose protocols from the primary synthesis. Eight independent treatment arms from seven studies, representing 1,180 patients, contributed to the principal five-year disease-control survival analysis.
  • The pooled five-year disease-control survival estimate was 79.6% (95% CI 75.8–82.8%) across contemporary 18-Gy and 23.4-Gy cohorts. In dose-specific analyses, 23.4 Gy yielded 81.0% (95% CI 78.7–83.1%; n=1,036), compared with 71.3% (95% CI 62.8–78.5%; n=144) after 18 Gy. Disease-control survival combined event-free, progression-free, or relapse-free survival according to the definitions used in the original studies. These pooled estimates were not derived from a randomized between-dose comparison.
  • Overall survival estimates also numerically favored 23.4-Gy treatment. Across six independent contemporary cohorts comprising 1,029 patients, pooled five-year overall survival was 85.3% (95% CI 81.2–88.6%). Dose-specific survival was 86.7% after 23.4 Gy (95% CI 84.9–88.3%; n=885) versus 79.1% after 18 Gy (95% CI 71.3–85.3%; n=144). Although the direction was consistent with disease-control outcomes, the two dose groups differed in patient selection, chemotherapy, and molecular characterization.
  • The randomized ACNS0331 trial provided the strongest direct evidence against routine 18-Gy CSI in younger children. Among children aged 3–7 years randomized to 18 versus 23.4 Gy, reduced-dose CSI was associated with inferior event-free survival (HR 1.67; p=0.028), with a reported one-sided 80% upper confidence limit of 2.10. The systematic review retained this randomized comparison separately rather than merging its hazard ratio with single-arm survival proportions. This distinction is important because the pooled 18-Gy and 23.4-Gy estimates cannot independently establish a causal dose effect.
  • Evidence supporting 23.4 Gy was substantially broader and more consistent. Six treatment arms involving 1,036 patients contributed to the 23.4-Gy disease-control estimate, with I²=0%. In contrast, only two 18-Gy cohorts involving 144 patients were available, including a small 28-patient pilot study. The review also cited SJMB03, where 227 average-risk patients treated with 23.4-Gy CSI and intensive chemotherapy achieved 83.2% five-year progression-free survival and 88.0% overall survival.
  • Reducing the boost volume may offer a different risk-benefit balance from reducing the entire craniospinal dose. Evidence from ACNS0331 suggested that restricting the posterior fossa boost to the tumor bed could preserve disease control, whereas reducing CSI to 18 Gy in younger children produced less favorable outcomes. Long-term neurocognitive, endocrine, auditory, vascular, and secondary malignancy outcomes were inconsistently reported across studies and could not be reliably pooled. Consequently, the analysis cannot quantify how much late toxicity would be prevented by reducing CSI from 23.4 to 18 Gy.
  • Evidence certainty differed substantially between analyses. Using GRADE, the authors rated pooled contemporary disease-control and overall survival estimates as low certainty, with the absolute 18-Gy estimates rated very low certainty. The direct randomized ACNS0331 dose comparison was considered moderate-certainty evidence. Most historical cohorts lacked comprehensive molecular stratification, while contemporary data suggest that WNT, SHH, Group 3, and Group 4 tumors require more refined clinicomolecular risk assessment before considering further CSI de-escalation.

CLINICAL TAKEAWAY

Craniospinal irradiation to 23.4 Gy combined with chemotherapy remains the best-supported reduced-dose approach for children with conventionally defined standard-risk medulloblastoma. Available evidence, including randomized ACNS0331 results, does not support routine reduction to 18 Gy in unselected patients, particularly younger children. Further de-escalation should be investigated within prospective, molecularly stratified protocols incorporating long-term neurocognitive and quality-of-survival outcomes.

SOURCE

Pediatric Blood & Cancer

Browse more research Suggest a correction