CCG 1953 (n=115 infants): the first trial to genotype every infant found the partner does not predict outcome
also called Hilden 2006 Blood · Children's Cancer Group 1953 infant ALL
This was the first cooperative-group trial to test every enrolled baby for which partner gene was involved. Despite doing the test on everyone, the partner made no measurable difference to how the babies did.
The same thing, for a scientist
CCG 1953 enrolled 115 infants with ALL and is described by its authors as the first cooperative group trial to characterise all infants at the molecular level for MLL/11q23 rearrangement. Every infant was tested by Southern blot, and the 11q23 partner was assigned by RT-PCR into t(4;11), t(9;11), t(11;19) or other. Overall 5-year EFS was 41.7% (SD 9.2) and OS 44.8%; 5-year EFS was 33.6% for MLL-rearranged versus 60.3% for MLL-non-rearranged cases. The difference in EFS between the three major MLL rearrangements did not reach statistical significance, and multivariable Cox regression ranked CD10 negativity, age under 6 months and MLL rearrangement as the prognostic factors - not the partner. The authors state the conclusion directly: molecular 11q23 translocation partners do not predict outcome.
Why this is the strongest pro-class result in the graph
Most negative partner findings in this literature are absences of reporting. This one is a tested and stated negative 1: everyone was genotyped, the partners were assigned molecularly rather than by karyotype, and the comparison was made and reported as non-significant. That is materially different from a trial that simply never broke its results out by partner.
What it cannot establish
115 infants split three ways leaves arms of a few dozen at most; the paper does not report per-arm n or per-arm EFS in its abstract, so the negative is a negative in a small study and not evidence of equivalence. The brief's standing rule applies with full force here - an absence of a detected difference is not a demonstrated absence of difference, and the power to detect a 15-point EFS gap in arms of that size is low. Toxicity, not relapse, was the most frequent cause of death 1, which compresses any biological difference between partners into a common treatment-related floor.
| metric | 5-year event-free survival by molecular 11q23 partner in infants |
|---|---|
| value | 41.7 |
| unit | % (whole cohort, SD 9.2) |
| comparator value | 33.6 |
| comparator unit | % (MLL-rearranged subset); 60.3% MLL non-rearranged |
| population | infants with ALL enrolled on CCG 1953 |
| n total | 115 |
| partner effect | not significant; authors state partners do not predict outcome |
| region | United States (Children's Cancer Group) |
| year | 2006 |
| prespecified | partner genotyping was prospective and universal; the between-partner outcome comparison is a subgroup analysis |
| source | PMID 16556894 |
partner arms
[ "t(4;11)", "t(9;11)", "t(11;19)", "other" ]
independent prognostic factors
[ "CD10 negativity", "age <6 months", "MLL rearrangement" ]
- A stated partner null in infants is the single best argument for offering a menin inhibitor to infants as a class, and it rests on 115 patients in a trial where toxicity killed more children than relapse - the safest-looking foundation in this beat is also one of the thinnest.
- What were the per-arm n and per-arm EFS values? Without them the null cannot be converted into a bound on how large a partner effect was excluded.
- Would the same null hold in a modern MRD-guided infant protocol where treatment-related mortality no longer dominates?
- Did the RT-PCR panel used have a partner-detection blind spot that pushed rare partners into the 'other' arm and flattened the comparison?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In CCG 1953, which molecularly genotyped all 115 enrolled infants with ALL and assigned the 11q23 partner by RT-PCR, the difference in EFS between the three major MLL rearrangements did not reach statistical significance, and the authors concluded that molecular 11q23 translocation partners do not predict outcome; overall 5-year EFS was 41.7% (SD 9.2) with 33.6% for MLL-rearranged versus 60.3% for non-rearranged.
Analysis of prognostic factors of acute lymphoblastic leukemia in infants: report on CCG 1953 from the Children's Oncology Group. · Hilden JM, Dinndorf PA, et al. · Blood 2006What this evidence cannot show
115 infants total, split across t(4;11), t(9;11), t(11;19) and other, so individual arms are small and per-arm n and EFS are not given in the abstract; the full text is not open access and the per-arm table was not retrieved. This is a null result in a small cohort, not a demonstration of equivalence. Toxicity was the commonest cause of death, which can mask biological differences between arms. Treatment era is late 1990s, and the partner assignment used a targeted RT-PCR panel that returns negative for any partner it was not designed around.
Evidence that argues against this (2)
- In children over 1 year with the same fusion class, a 629-patient international cohort reported 5-year EFS ranging from 41.7% to 91.2% across partner-by-lineage cells, so the null seen in infants does not generalise to older children. Attarbaschi A et al. 2023
- In childhood AML, 1,130 patients grouped by partner separated by 24 EFS points with the partner group independent of MRD in multivariable analysis, so the same partners that look inert in infant ALL are strongly prognostic in myeloid disease. van Weelderen RE et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A pooled analysis of 497 children and young adults with 11q23-rearranged ALL treated 1983-1995 found age the most important prognostic factor, with any category of 11q23 abnormality conferring a dismal outcome in infants while t(4;11) and t(9;11) were worse than other 11q23 changes only in older patients.
Outcome of treatment in childhood acute lymphoblastic leukaemia with rearrangements of the 11q23 chromosomal region. · Pui CH, Gaynon PS, et al. · Lancet 2002What this evidence cannot show
Pooled retrospective data from 11 study groups and single institutions over 12 years of older protocols; partner assignment largely cytogenetic. The infant-versus-older-patient contrast is a subgroup comparison within an already heterogeneous cohort.
Evidence that argues against this (1)
- A study of 465 KMT2A-rearranged B-ALL cases spanning 1 month to 89 years found that fusion partner remained an independent determinant of MRD clearance in multivariable analysis including age group and developmental maturity (partner P=0.03), so age does not absorb the whole partner effect. Hartmann AM et al. 2026
universal molecular partner genotyping; no significant EFS difference between the three major partners
partner null in infant ALL against a 24-point partner effect in childhood AML
- ← supportsInfant BCP-ALL (n=175): KMT2A::AFF1 2-year EFS 36% versus 37% for every other partner (P=0.9)
second, RNA-seq-era infant cohort reaching the same null
a second infant protocol in which partner carries no stratifying role
an infant-only cohort cannot detect a partner effect that operates through age