Oncofusion

start / graph / trial-interfant-06

Trialstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q6b-therapeutics2

Interfant-06 — 651 infants, and the definitive answer that myeloid-style consolidation does not help

also called Interfant-06 · NCT00550992 · EudraCT 2005-004599-19 · DCOG-INTERFANT-06

Interfant-06 is the chemotherapy plan most babies with KMT2A-rearranged leukemia are still treated on. Its main experiment — swapping in AML-style drug blocks after induction — made no difference, and overall survival did not improve compared with the previous protocol.

The same thing, for a scientist

Interfant-06 enrolled 651 infants under 1 year with ALL across 18 national and international groups and randomised medium- and high-risk patients to lymphoid-type consolidation (protocol IB: low-dose cytarabine, 6-mercaptopurine, cyclophosphamide) versus two myeloid-type courses (ADE then MAE). The randomisation was negative: 6-year DFS 39.3% (SE 4.0, n=169) for ADE+MAE versus 36.8% (SE 3.9, n=161) for IB, log-rank p=0.47. Whole-cohort 6-year EFS was 46.1% (SE 2.1) and OS 58.2% (SE 2.0), which the authors state did not improve on Interfant-99. KMT2A rearrangement was the strongest prognostic factor for EFS. Interfant-06 remains the backbone on which blinatumomab, venetoclax and menin inhibitors are now being added.

Go deeper
The risk model, which every later trial inherits

Interfant-06 is where infant ALL risk stratification became formal, and every subsequent trial (Interfant-21, NCT06317662, TINI 2) uses a version of it.

  • Low risk (LR): KMT2A germline.
  • High risk (HR): KMT2A-rearranged and age under 6 months and (WBC ≥300 ×10⁹/L and/or prednisone poor response).
  • Medium risk (MR): all other KMT2A-rearranged cases.

The published abstract states this in condensed form and the ClinicalTrials.gov record (NCT00550992) gives the fuller eligibility text; the two are not verbatim identical, and the registry's detailed criteria should be treated as the more reliable statement of the boolean logic 12.

The randomised question, and the answer2 deeper

Lymphoid-style versus myeloid-style consolidation means exactly this comparison, in MR and HR patients who achieved remission:

ArmWhat it isn6-year DFS
IB (lymphoid)low-dose cytarabine, 6-mercaptopurine, cyclophosphamide — a standard ALL consolidation block16136.8% (SE 3.9)
ADE + MAE (myeloid)cytarabine/daunorubicin/etoposide, then cytarabine/mitoxantrone/etoposide — standard AML induction-intensity blocks16939.3% (SE 4.0)

log-rank p = 0.47 1. No subgroup in which myeloid intensification won is reported in the abstract, and we could not access the full text to check the subgroup tables.

The logic being tested was seductive and wrong in the same way twice. KMT2A-rearranged infant leukemia is lineage-ambiguous, often co-expresses myeloid antigens, and switches to AML under pressure — so treating it with AML drugs should help. It does not. Lineage plasticity is not the same as drug-class sensitivity. That distinction, established across Interfant-99 and Interfant-06 in 1,133 infants, is one of the most robust negative results in paediatric oncology and it deserves to be stated plainly before anyone proposes another intensification.

Whole-cohort outcomes
EndpointValue
6-year EFS, all patients46.1% (SE 2.1)
6-year OS, all patients58.2% (SE 2.0)
6-year EFS, West European / North American groups49.4% (SE 2.5)
6-year OS, West European / North American groups62.1% (SE 2.4)
6-year EFS, high-risk group20.9% (SE 3.4)

The abstract reports standard errors, not 95% confidence intervals; we therefore do not state CIs for these figures, and neither should anything downstream of this node. The 10-12 percentage point gap between West European/North American groups and the rest is a supportive-care and access signal, not a biology signal, and it is larger than any treatment effect the trial tested.

What the high-risk group tells us about transplant

High-risk infants were intended to proceed to stem-cell transplant in first remission. Only 46% of them actually received one, because many had early events 1. Their 6-year EFS was 20.9% (SE 3.4).

This is the most important practical fact about HSCT in infant KMT2A-rearranged ALL: the problem is not whether transplant works, it is that more than half of the children it is intended for relapse or die before they can get there. Any intervention that simply holds a high-risk infant in remission long enough to reach transplant — which is precisely the argument for blinatumomab and for menin inhibitors as bridges — is doing something the protocol itself could not.

Facts and figures
nctNCT00550992
eudract2005-004599-19
phasePhase 3, international randomised
statusCompleted; still the standard chemotherapy backbone in practice and in current trials
n651
populationInfants <1 year with newly diagnosed ALL or biphenotypic leukemia
pediatric eligibletrue
open nowfalse
readoutNegative for myeloid-style consolidation; outcome did not improve versus Interfant-99
primary result
randomised_p0.470randomised_p: 0.47os_se2.000os_se: 2efs_se2.100efs_se: 2.1efs_6y_pct46.1efs_6y_pct: 46.1os_6y_pct58.2os_6y_pct: 58.2
high risk result
se3.400se: 3.4efs_6y_pct20.9efs_6y_pct: 20.9received_hsct_pct46.0received_hsct_pct: 46
regional result
os_se2.400os_se: 2.4efs_se2.500efs_se: 2.5west_europe_north_america…49.4west_europe_north_america_efs_6y_pct: 49.4os_6y_pct62.1os_6y_pct: 62.1
risk groups
{
  "low": "KMT2A germline",
  "high": "KMT2A-rearranged AND age <6 months AND (WBC >=300 x10^9/L and/or prednisone poor response)",
  "medium": "all other KMT2A-rearranged"
}
arms
[
  {
    "name": "Protocol IB (lymphoid consolidation: low-dose araC, 6-MP, cyclophosphamide)",
    "n": 161,
    "dfs_6y_pct": 36.8,
    "se": 3.9
  },
  {
    "name": "ADE + MAE (myeloid consolidation)",
    "n": 169,
    "dfs_6y_pct": 39.3,
    "se": 4
  }
]
primary result — the numbers behind the chart
{
  "efs_6y_pct": 46.1,
  "efs_se": 2.1,
  "os_6y_pct": 58.2,
  "os_se": 2,
  "randomised_p": 0.47
}
high risk result — the numbers behind the chart
{
  "efs_6y_pct": 20.9,
  "se": 3.4,
  "received_hsct_pct": 46
}
regional result — the numbers behind the chart
{
  "west_europe_north_america_efs_6y_pct": 49.4,
  "efs_se": 2.5,
  "os_6y_pct": 62.1,
  "os_se": 2.4
}
How this could bite us back
  • Essential-function toxicity: ADE and MAE deliver anthracycline plus mitoxantrone to a patient under one year old, for no survival benefit — the clearest case in this disease of harm without gain.
  • Late effects: cumulative anthracycline in infancy carries lifetime cardiomyopathy risk, and the negative randomisation means some of that exposure was paid for nothing.
  • Delivery/attrition: the 46% transplant-receipt rate in high-risk patients shows that a treatment plan can be defeated by disease kinetics before any pharmacology matters.
  • Compensation: two independent negative randomisations of intensification in over 1,100 infants argue that cytotoxic dose intensity is saturated in this disease; the remaining levers are mechanistic and immunological.
Still open
  1. Why is the high-risk group's 6-year EFS 20.9% when 54% of them never reached the transplant the protocol intended? Separating 'transplant does not work' from 'we cannot get them to transplant' is the central unresolved question in high-risk infant ALL.
  2. Is the 10-12 point regional outcome gap entirely supportive care, or does it contain a biology signal (age distribution, WBC at presentation, time to diagnosis)? It is larger than any randomised effect the trial found and nobody has decomposed it.
  3. What are the low- and medium-risk group outcomes with confidence intervals? These drive risk-adapted decisions in clinic and we could not verify them from the primary source.
  4. Did the myeloid arm help anyone? A pre-specified interaction test between fusion partner (AFF1 vs MLLT3 vs MLLT1) and consolidation type has never been reported, and partner-specific myeloid bias is real.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Interfant-06 enrolled 651 infants under 1 year with ALL across 18 groups; randomisation of medium/high-risk patients to myeloid-type ADE+MAE versus lymphoid-type protocol IB gave 6-year DFS 39.3% (SE 4.0, n=169) versus 36.8% (SE 3.9, n=161), log-rank p=0.47; whole-cohort 6-year EFS was 46.1% (SE 2.1) and OS 58.2% (SE 2.0); high-risk 6-year EFS was 20.9% (SE 3.4) with only 46% receiving the intended stem-cell transplant; KMT2A rearrangement was the strongest prognostic factor and outcome did not improve versus Interfant-99.

    PMID 31283407DOI
    What this evidence cannot show

    Taken from the primary abstract; the full text is paywalled and was not retrieved, so 95% confidence intervals (the abstract reports standard errors only), low- and medium-risk subgroup EFS/OS, the toxic death rate, and hazard ratios with CIs for KMT2A rearrangement are NOT asserted anywhere in this node. The comparison against Interfant-99 is historical, not randomised. The randomised comparison was restricted to MR/HR patients who achieved remission, so it says nothing about myeloid-style therapy given earlier or to induction failures.

    Evidence that argues against this (1)
    • The strongest published argument against 'more intensive chemotherapy cannot move KMT2A-rearranged infant ALL' is the JPLSG MLL-10 trial, which reported 3-year EFS 66.2% (SE 5.6) in KMT2A-rearranged infants (n=75 IR+HR) using early intensification with high-dose cytarabine and HSCT restricted to high risk - against Interfant-06's 6-year EFS 46.1% whole-cohort and 20.9% in high risk. MLL-10 also eliminated HSCT for a subset, the opposite of the Interfant-06 high-risk strategy in which only 46% of patients reached their intended transplant. Counterweight: different countries, different eras, no randomisation between them, different follow-up horizons (3 vs 6 years), and MLL-10's high-risk group still did badly (3-year EFS 56.6%). Tomizawa D et al. 2020
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry

    Interfant-06 is registered as NCT00550992 (sponsor Dutch Childhood Oncology Group, 18 collaborating groups, start January 2006) with secondary identifier EudraCT 2005-004599-19 and organisation study ID DCOG-INTERFANT-06.

    NCT00550992
    What this evidence cannot show

    The registry's enrolment field (445) does not match the 651 infants analysed in the publication; this most likely reflects a partial registration snapshot rather than a discrepancy in the paper, but it is a reason to cite the publication for N and the registry for identifiers only.

    Evidence that argues against this (1)
    • This registry record is stale and its enrolment figure is wrong by a third, so it should not be used as a source of trial facts. Raw ClinicalTrials.gov v2 API, retrieved 2026-09-19: NCT00550992 has overall status UNKNOWN, was last updated on 2019-07-29, has hasResults FALSE, carries a phase field of 'NA', and lists enrolment 445 as ESTIMATED. The published international phase 3 report of Interfant-06 states that 651 INFANTS were included, with 330 randomised between myeloid-type (ADE+MAE, n=169) and lymphoid-type (IB, n=161) post-induction courses. Registration metadata for academic consortium trials of this era is frequently never reconciled with the final cohort; the publication is the citable record. Pieters R et al. 2019
Connected to