Oncofusion

start / graph / epi-interfant-protocols-partner-stratified-outcome-not-reported

Epidemiologymoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q19-partner-outcome

Interfant-99 and Interfant-06: what they report about partners, and what they do not

also called Interfant partner-stratified outcome absence · bounded absence: Interfant per-partner EFS

The two worldwide trials for babies with this leukemia both name the KMT2A rearrangement as the strongest predictor of outcome, but neither publishes survival broken down by which partner gene was involved. That missing table is exactly the number this project needed.

The same thing, for a scientist

Interfant-99 (n=482 infants, 17 study groups, 22 countries, 1999-2005) reports 4-year EFS 47.0% (SE 2.6; 95% CI 41.9-52.1) and states that all types of rearrangements in the MLL gene, very high WBC, age under 6 months and poor prednisone response were independently associated with inferior outcome. Interfant-06 (n=651, 18 study groups) reports 6-year EFS 46.1% (SE 2.1) and OS 58.2% (SE 2.0), and names KMT2A rearrangement as the strongest prognostic factor for EFS, followed by age, WBC and prednisone response. Neither primary report gives EFS by individual 3-prime partner in its published abstract, and neither is available as open-access full text through PubMed Central or Europe PMC, so per-partner arms could not be read. The phrase in Interfant-99 - all types of MLL rearrangement conferred inferior outcome - is the closest either protocol comes to a partner statement, and it is a class statement, not a stratified one.

Go deeper
Retrieval record, so the next reader does not repeat it

Both PMIDs were checked through the route the standing brief requires: Europe PMC search?query=EXT_ID:<PMID>&resultType=core&format=json returned pmcid: null for PMID 17658395 (Interfant-99) and PMID 31283407 (Interfant-06), so no fullTextXML fetch was possible for either. Abstracts were read from NCBI efetch XML. The per-partner table, if it exists, is behind the publisher paywall in Lancet 2007 and JCO 2019 respectively and was not read. This entry therefore records an absence in the retrievable record, not an absence in the papers, and it must not be cited as evidence that Interfant found no partner effect.

A class statement is not a stratified result

Interfant-99's multivariable finding is that all types of MLL rearrangement are independently adverse 1. That is compatible with three quite different underlying facts: every partner is equally bad; partners differ but all sit below the germline comparator; or partners differ and the model, which used a single binary rearrangement term, could not have seen it. The standing rule applies - an absence of testing is not a negative result. Nothing in either protocol's abstract establishes which of the three is true.

The brief this beat was dispatched under was wrong on this point

The dispatch stated that Interfant-99 and Interfant-06 both report outcome by partner. From the retrievable record that could not be confirmed: neither abstract contains per-partner EFS, and neither full text is open. The claim may still be true of the papers' tables. It is recorded here as unverified, in line with the standing instruction that the literature, not the brief, decides.

Facts and figures
metricpartner-stratified EFS availability in the Interfant protocols
valuenot reported in either published abstract; full texts not open access
unitavailability
populationinfants under 1 year with ALL
regioninternational
year2019
prespecifiedneither trial pre-specified a partner comparison; both randomised other questions
sourcePMIDs 17658395 and 31283407
trials
[
  {
    "name": "Interfant-99",
    "pmid": "17658395",
    "n": 482,
    "efs_4y_pct": 47,
    "se": 2.6,
    "nct": "NCT00015873",
    "isrctn": "ISRCTN24251487",
    "partner_statement": "all types of MLL rearrangement independently adverse",
    "title": "A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial.",
    "authors": "Pieters R, Schrappe M, De Lorenzo P, et al.",
    "venue": "Lancet",
    "year": 2007,
    "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
  },
  {
    "name": "Interfant-06",
    "pmid": "31283407",
    "n": 651,
    "efs_6y_pct": 46.1,
    "se": 2.1,
    "os_6y_pct": 58.2,
    "partner_statement": "KMT2A rearrangement named as strongest prognostic factor; partner not broken out",
    "title": "Outcome of Infants Younger Than 1 Year With Acute Lymphoblastic Leukemia Treated With the Interfant-06 Protocol: Results From an International Phase III Randomized Study.",
    "authors": "Pieters R, De Lorenzo P, Ancliffe P, et al.",
    "venue": "J Clin Oncol",
    "year": 2019,
    "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
  }
]
retrieval record
{
  "europepmc_core_pmcid_17658395": null,
  "europepmc_core_pmcid_31283407": null,
  "fulltextxml_fetched": false,
  "route": "Europe PMC search?query=EXT_ID:<PMID>&resultType=core&format=json, then <PMCID>/fullTextXML",
  "conclusion": "absence is in the retrievable record, not established in the papers"
}
How this could bite us back
  • Interfant-21 (NCT05327894) is randomising blinatumomab in this exact population now. If it does not pre-specify partner-stratified reporting, the field will spend another protocol cycle without the number, and a menin-inhibitor successor trial will inherit the same blind spot.
  • Because the Interfant models used a binary KMT2A-rearranged term, any future claim that Interfant showed partner does not matter would be an absence-of-testing error, and it is the kind that reads as a finding.
Still open
  1. Do the Interfant-99 and Interfant-06 full texts or supplements contain per-partner EFS? A single library request would settle this and is the cheapest open item in this beat.
  2. Would the Interfant network release per-partner EFS with n for a pooled 99+06 analysis? Combined that is over 1,100 infants, by far the largest infant partner dataset that could exist.
  3. Interfant-06 outcomes differed by region by 10-12 points. Does partner distribution differ by region too, and if so does region confound any pooled partner comparison?
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Interfant-99 enrolled 482 infants under 1 year across 22 countries with 4-year EFS 47.0% (SE 2.6; 95% CI 41.9-52.1), and reported that all types of rearrangement in the MLL gene, very high WBC, age under 6 months and poor prednisone response were independently associated with inferior outcome - a class-level statement with no per-partner breakdown in the published abstract.

    PMID 17658395DOI
    What this evidence cannot show

    The abstract is the only part of this paper read for this entry: Europe PMC returned pmcid null for PMID 17658395 and no full text could be fetched. Whether the paper's tables contain per-partner EFS is unknown. The randomised question in this trial was late intensification, not anything about partners, so any partner analysis it contains would be a non-pre-specified subgroup.

    Evidence that argues against this (1)
    • A 497-patient pooled analysis of childhood 11q23 ALL over an earlier era found that partner distinctions did emerge in patients over 1 year old even though any 11q23 abnormality was uniformly dismal in infants, so a class-level statement in an infant-only trial is consistent with a real partner effect that infancy masks. Pui CH et al. 2002
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Interfant-06 enrolled 651 infants with 6-year EFS 46.1% (SE 2.1) and OS 58.2% (SE 2.0), and identified KMT2A rearrangement as the strongest prognostic factor for EFS followed by age, WBC count and prednisone response, without reporting EFS by individual 3-prime partner in the published abstract.

    PMID 31283407DOI
    What this evidence cannot show

    Abstract only; Europe PMC returned pmcid null for PMID 31283407 and no full text could be fetched. The prognostic factor named is KMT2A rearrangement status, a binary term, so the model as described could not have detected a between-partner difference. The trial's randomised question was myeloid versus lymphoid consolidation. Outcome differed by region within the trial (West European/North American groups 10-12 points better), which is an additional source of between-subgroup variation unrelated to partner.

    Evidence that argues against this (1)
    • A 175-infant BCP-ALL cohort that did perform the partner comparison found no difference (KMT2A::AFF1 2-year EFS 36% versus 37% for other partners, P=0.9), which suggests the missing Interfant table would probably also be null - but that is an inference from a different cohort, not from Interfant. Ilyasova K et al. 2025
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    CCG 1953 demonstrates that a cooperative-group infant trial can and did report the partner comparison explicitly, so its absence from the Interfant abstracts is a reporting choice rather than a technical impossibility.

    PMID 16556894DOI
    What this evidence cannot show

    CCG 1953 is a different, smaller and earlier trial in a different cooperative group; that it reported the comparison does not establish that Interfant could have with comparable power, nor that Interfant did not report it in its full text.

    Evidence that argues against this (1)
    • CCG 1953 reported the comparison but did not give per-arm n or per-arm EFS in its abstract either, so even the trial that did test partner did not publish the numbers a meta-analysis would need. Hilden JM et al. 2006
Connected to