Oncofusion

start / graph / concept-interfant-06-treatment-backbone

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q34-mrd-monitoring

The Interfant-06 backbone, phase by phase: what the protocol every infant reference in this graph points at actually does

also called Interfant-06 protocol · Interfant backbone · protocol IB · MARMA · OCTADA(D) · ADE/MAE · prednisone prophase

Interfant-06 is the two-year chemotherapy plan that most babies with this leukaemia are still treated on, and almost every trial that adds a new drug adds it to this backbone. It runs a one-week steroid test, a five-week induction, a consolidation block, high-dose methotrexate and cytarabine, a re-induction, and then maintenance tablets to two years from diagnosis, with a transplant offered to the highest-risk group.

The same thing, for a scientist

Interfant-06 enrolled 651 infants under 1 year with ALL or biphenotypic leukaemia across 18 national and international groups [E1]. Structure, from the registered protocol [E2]: a **7-day prednisone prophase** with intrathecal methotrexate and prednisolone on day 1; a **5-week remission induction** of dexamethasone days 8-28 with taper, vincristine days 8/15/22/29, cytarabine days 8-21, daunorubicin days 8-9, asparaginase on six days from day 15, with triple-style intrathecal therapy; then risk stratification into low risk (KMT2A-germline), medium risk and high risk. Medium- and high-risk patients in complete remission on day 33 were **randomised** to lymphoid-style **protocol IB** (cyclophosphamide, 6-mercaptopurine, low-dose cytarabine) or to two myeloid-style courses, **ADE** then **MAE** [E1][E2]. All then received **MARMA** (high-dose methotrexate with leucovorin rescue, then high-dose cytarabine and pegaspargase), **OCTADA(D)** reinduction (dexamethasone, thioguanine, vincristine, pegaspargase, cytarabine, with daunorubicin in the D variant, then a second cyclophosphamide/cytarabine part), and **maintenance** of daily 6-mercaptopurine and weekly oral methotrexate with intrathecal therapy, continuing to 104 weeks from diagnosis [E2]. High-risk patients with a suitable donor were scheduled for allogeneic transplant in first remission within 8 months of diagnosis, conditioned with busulfan, cyclophosphamide and melphalan [E2]. The randomised question was negative: 6-year disease-free survival 39.3% for ADE+MAE versus 36.8% for IB, log-rank P = .47; whole-cohort 6-year EFS 46.1% and OS 58.2% [E1].

Go deeper
The phases, in order

All drug-level detail from the registered protocol description 2.

PhaseContent
Prednisone prophaseprednisone orally or IV three times daily, days 1-7; MTX + prednisolone intrathecally day 1. Response to this defines prednisone good/poor responder
Remission inductiondexamethasone days 8-28 with 1-week taper; vincristine days 8, 15, 22, 29; cytarabine IV days 8-21; daunorubicin days 8-9; asparaginase days 15, 18, 22, 25, 29, 33; MTX IT days 1 and 29; cytarabine IT day 15; prednisolone or hydrocortisone IT days 1, 15, 29. CNS-positive patients get extra IT MTX on days 8 and 22, then weekly until clear
Randomised consolidationArm I / protocol IB from day 36: cyclophosphamide days 1 and 29, oral 6-mercaptopurine days 1-28, cytarabine days 3-6, 10-13, 17-20, 24-27, with IT therapy. Arm II: ADE (cytarabine, daunorubicin, etoposide) then MAE (mitoxantrone, cytarabine, etoposide)
MARMAPart I: oral 6-MP days 1-14, high-dose MTX 24-hour infusion days 1 and 8 with leucovorin rescue at 42, 48 and 54 hours, IT MTX days 2 and 9. Part II: high-dose cytarabine twice daily days 15, 16, 22, 23; pegaspargase day 23
OCTADA(D)Part I: dexamethasone days 1-14 with taper to day 21, oral thioguanine days 1-28, vincristine days 1, 8, 15, 22, pegaspargase day 1, cytarabine in four 4-day blocks, IT therapy days 1 and 15; OCTADAD adds daunorubicin days 1, 8, 15, 22. Part II: thioguanine days 36-49, cytarabine days 37-40 and 45-48, cyclophosphamide days 36 and 49
Maintenancedaily oral 6-mercaptopurine, weekly oral methotrexate, IT MTX weeks 1 and 15, IT cytarabine week 8, to 104 weeks from initial diagnosis
HSCT (high risk only)after MARMA or before/during OCTADA(D), in CR1, within 8 months of diagnosis. Matched sibling: busulfan days -7 to -4, cyclophosphamide -3 to -2, melphalan -1. Matched unrelated: the same plus ATG days -3 to -1. GVHD prophylaxis ciclosporin, with short-course MTX and leucovorin for matched unrelated donors

MRD sampling sits at end of induction (TP2), end of consolidation (TP4) and after MARMA (TP5); MRD was not mandatory except for medium-risk patients at TP5 to determine transplant eligibility in first remission 3.

The risk model, and a discrepancy worth recording

The registered protocol defines the three groups as 2:

  • Low risk: all KMT2A-germline cases.
  • Medium risk: KMT2A status unknown; or KMT2A-rearranged and age >6 months; or KMT2A-rearranged and age <6 months with WBC <300 x10^9/L and prednisone good response.
  • High risk: KMT2A-rearranged and age at diagnosis <6 months (<183 days) and WBC >=300 x10^9/L and/or prednisone poor response.

The published Interfant-06 abstract states the age criterion in the opposite direction, defining high risk as KMT2A-rearranged and older than 6 months with WBC >=300 x10^9/L or poor prednisone response 1. The registry 2 and the protocol's own MRD paper, which lists the high-risk factors as age <6 months at diagnosis together with WBC >=300 x10^9/L and/or poor prednisone response 3, both give age under 6 months. Two of three sources, including one from the same authors, contradict the abstract, and the direction matters because age <6 months is independently adverse in Interfant-99 4. Recorded here as a discrepancy rather than a correction: this beat did not obtain the Interfant-06 full text, which is not open access and returned no PMCID by any of the four routes.

The practical significance is that high risk in Interfant-06 is a conjunction, not a list. An infant of 4 months with a WBC of 400 x10^9/L is high risk; an infant of 8 months with the same count is medium risk. Entries in this graph that quote the criteria as a disjunction are describing a different, larger high-risk group than the protocol used.

Why this backbone is the reference frame for everything else

Three reasons the graph keeps pointing here.

It is the comparator. Interfant-06 remains the backbone on which blinatumomab, venetoclax and menin inhibitors are being added; the headline blinatumomab result in this disease is a single course of sequential blinatumomab added to Interfant-06 chemotherapy, improving 2-year disease-free survival to 82% against 49% for historical chemotherapy alone 5. A historical comparison against Interfant-06 is only interpretable if the backbone is known.

It defines the MRD timepoints. End of induction, end of consolidation and post-MARMA are protocol phases, not calendar points, so an MRD value from another protocol is not exchangeable with an Interfant one. concept-mrd-timepoints-and-decisions-infant-all depends entirely on this structure.

Its central randomisation failed, which is the most important single fact about it. Myeloid-style consolidation did not beat lymphoid-style across the cohort (P = .47), and whole-cohort outcomes did not improve on Interfant-99 1. The subsequent MRD analysis found an interaction - high end-of-induction MRD did better with myeloid-style consolidation, negative MRD did better with lymphoid-style 3 - which is the finding that turned a negative trial into a stratification hypothesis for the next protocol. This backbone is therefore not a settled standard being iterated on; it is a plateau that two consecutive international trials failed to move.

Correction: the protocol document IS retrievable, and this graph already held the link

This entry records that its beat could not obtain the Interfant-06 full text, which is not open access and returned no PMCID by any of four routes. That is true of the JOURNAL PAPER and false of the PROTOCOL DOCUMENT, which are different objects.

The 90-page SKION national protocol (version 01, April 2006) is freely retrievable and was fetched on 2026-09-20 (HTTP 200, 890,629 bytes, ~195,000 characters of text). It gives a dose, a schedule, a phase and a toxicity rule for every agent in the backbone - the detail this entry reconstructs from the registry and marks with 'no dose stated' gaps.

And this graph already contained the link. It sits in risk-anthracycline-cardiotoxicity-infant-leukemia E1, written in an earlier beat. One entry held the answer to another entry's recorded absence, and nothing in this repository looks across entries for that - check-dangling-data-refs.py finds references to nodes that do not exist, and there is no corresponding check for a resource one entry has and another says it lacks.

The phase and dose detail here can now be upgraded from the protocol itself. The statement about the JCO paper's accessibility stands unchanged and was never wrong.

Facts and figures
nctNCT00550992
eudract2005-004599-19
lead sponsorDutch Childhood Oncology Group
collaborating groups listed in registry18
enrolment published651
enrolment registered estimate445
age eligibilityunder 1 year (maximum age 1 year in registry)
risk group discrepancythe published abstract (pmid 31283407) states the high-risk age criterion as OLDER than 6 months; the registry (NCT00550992) and the protocol's own MRD paper (pmid 33405950) both state age UNDER 6 months
mrd mandatoryonly for medium-risk patients at TP5, to determine HSCT eligibility in CR1
hsct conditioningbusulfan + cyclophosphamide + melphalan; ATG added for matched unrelated donors
randomised result
p0.470p: 0.47arm_ib_dfs_6y_pct36.8arm_ib_dfs_6y_pct: 36.8arm_ade_mae_dfs_6y_pct39.3arm_ade_mae_dfs_6y_pct: 39.3n_ib161.0n_ib: 161n_ade_mae169.0n_ade_mae: 169
whole cohort
hr_group_efs_6y_pct20.9hr_group_efs_6y_pct: 20.9efs_6y_pct46.1efs_6y_pct: 46.1os_6y_pct58.2os_6y_pct: 58.2
phases
[
  "prednisone prophase (7 days)",
  "remission induction (day 8 to day 33)",
  "randomised consolidation: protocol IB vs ADE+MAE",
  "MARMA",
  "OCTADA(D) reinduction",
  "maintenance to 104 weeks from diagnosis",
  "allogeneic HSCT for high risk in CR1"
]
risk groups registry
{
  "low": "all KMT2A-germline cases",
  "medium": "KMT2A status unknown; or KMT2A-rearranged and age >6 months; or KMT2A-rearranged and age <6 months with WBC <300e9/L and prednisone good response",
  "high": "KMT2A-rearranged AND age <6 months (<183 days) AND (WBC >=300e9/L and/or prednisone poor response)"
}
randomised result — the numbers behind the chart
{
  "arm_ade_mae_dfs_6y_pct": 39.3,
  "arm_ib_dfs_6y_pct": 36.8,
  "p": 0.47,
  "n_ade_mae": 169,
  "n_ib": 161
}
whole cohort — the numbers behind the chart
{
  "efs_6y_pct": 46.1,
  "os_6y_pct": 58.2,
  "hr_group_efs_6y_pct": 20.9
}
mrd timepoints
[
  "diagnosis",
  "end of induction (TP2)",
  "end of consolidation (TP4)",
  "after MARMA (TP5)"
]
exclusions
[
  "mature B-ALL",
  "t(9;22)/BCR-ABL1",
  "relapsed ALL"
]
How this could bite us back
  • Two years of therapy including anthracyclines, high-dose methotrexate, high-dose cytarabine, etoposide and for the highest-risk group busulfan-based transplant, all delivered to a child under one year, for a 6-year event-free survival of 46.1%. The late-effects burden of this backbone is the price every added agent is measured against, and an agent that only adds toxicity to it has no route to benefit.
  • The central randomisation was negative and whole-cohort outcomes did not improve on the previous protocol. Any new agent added to this backbone inherits a comparator that two international trials failed to improve, which makes a single-arm historical comparison against it unusually easy to beat and unusually easy to misread.
  • High risk in this protocol is a conjunction of KMT2A rearrangement, young age and either hyperleukocytosis or poor steroid response. Every term is a dichotomised continuous variable except the first, so the group that receives transplant in first remission is defined by three committee decisions stacked on top of each other.
  • MRD was not mandatory except at one timepoint for one risk group. The evidence base for MRD-driven decisions in this disease therefore comes from a non-randomly sampled subset of a trial that did not require the measurement, and the next protocol is being built on it.
Still open
  1. Which age direction does the Interfant-06 high-risk definition actually use? Two sources say under 6 months and the published abstract says over 6 months. Resolving it needs the full text or the protocol document, and every downstream entry that quotes the risk model inherits the error if it is one.
  2. How much of the 10-12 percentage-point EFS gap between West European/North American groups and the rest is supportive care, and how much is protocol delivery? Interfant-06 reports the gap and does not decompose it, and it is larger than any treatment effect the trial tested.
  3. Interfant-06's outcomes did not improve on Interfant-99 despite a randomised intensification question in each. What is the argument that backbone intensification has not been exhausted in this disease?
  4. Does the MRD-by-consolidation-arm interaction replicate? Interfant-21 is testing blinatumomab, not that interaction, so the hypothesis generated by the negative trial may never be tested directly.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Interfant-06 enrolled 651 infants younger than 1 year with acute lymphoblastic leukaemia across 18 national and international study groups, defined three risk groups on KMT2A status, age, WBC count and prednisone response, randomised medium- and high-risk patients to lymphoid-style protocol IB versus myeloid-style ADE then MAE, and found no difference (6-year disease-free survival 39.3% for ADE+MAE versus 36.8% for IB, log-rank P = .47), with whole-cohort 6-year event-free survival 46.1% and overall survival 58.2% and a 6-year EFS of 20.9% in the high-risk group.

    PMID 31283407DOI
    What this evidence cannot show

    Abstract only; full text not open access and no PMCID was returned by the Europe PMC core record, so the four-route full-text test terminated at route 1. The abstract's statement of the high-risk age criterion (older than 6 months) contradicts both the registered protocol and the protocol's own MRD publication, which is recorded in the body of this entry as an unresolved discrepancy. Outcomes differed by region (6-year EFS 49.4% in West European/North American groups, 10-12% higher than elsewhere), so pooled figures are not a single population.

    Evidence that argues against this (1)
    • A single-institution Argentinian series comparing Interfant protocols against BFM-based protocols in 116 infants found comparable induction and complete-remission mortality and no reduction in relapse rate, and concluded the two strategies presented comparable results - so the backbone's superiority over general ALL therapy in infants is not established everywhere it has been tested. Pennella CL et al. 2020
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    The registered Interfant-06 protocol specifies a 7-day prednisone prophase with intrathecal therapy, a remission induction of dexamethasone, vincristine, cytarabine, daunorubicin and asparaginase with intrathecal methotrexate and cytarabine, randomisation of medium- and high-risk patients in complete remission on day 33 between protocol IB (cyclophosphamide, 6-mercaptopurine, cytarabine) and ADE then MAE, followed by MARMA (high-dose methotrexate with leucovorin rescue then high-dose cytarabine and pegaspargase), OCTADA(D) reinduction, and maintenance of 6-mercaptopurine and methotrexate to 104 weeks from diagnosis; high-risk patients with a suitable donor are scheduled for allogeneic transplant in first remission within 8 months of diagnosis, conditioned with busulfan, cyclophosphamide and melphalan, with ATG added for matched unrelated donors; and high risk is defined as KMT2A rearrangement with age at diagnosis under 6 months (under 183 days) and WBC at or above 300 x10^9/L and/or prednisone poor response.

    International Collaborative Treatment Protocol for Infants Under One Year With Acute Lymphoblastic or Biphenotypic Leukemia (Interfant-06) · Dutch Childhood Oncology Group (lead sponsor), with 18 named national and international collaborating groups · ClinicalTrials.gov 2006
    NCT00550992
    What this evidence cannot show

    A trial registry record, fetched from the raw ClinicalTrials.gov API on 2026-09-20. It states what the protocol prescribes, not what was delivered; the record's overall status is UNKNOWN and enrolment is listed as an estimated 445 against the 651 actually enrolled and published. Registry text is a protocol summary written for the registry, and dosing is absent. The whyStopped field is null.

    Evidence that argues against this (1)
    • The published report gives an enrolment of 651 against the registry's estimated 445 and states the high-risk age criterion in the opposite direction, so the registry and the publication disagree on two verifiable facts about the same trial. Pieters R et al. 2019
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the Interfant-06 MRD analysis, bone marrow samples were taken at diagnosis, after induction (end of induction, TP2), after consolidation (end of consolidation, TP4) and after MARMA (TP5); MRD analysis was not mandatory except for medium-risk patients at TP5 to evaluate eligibility for transplant in first remission; high-risk factors on the protocol are listed as age under 6 months at diagnosis and a high WBC count at diagnosis at or above 300 x10^9/L and/or a poor prednisone response; and the prognostic impact of end-of-induction MRD depended on the consolidation arm, with high MRD patients doing better on myeloid-style and MRD-negative patients better on lymphoid-style consolidation.

    PMID 33405950DOI
    What this evidence cannot show

    Read in full text. The MRD-by-arm interaction is a secondary analysis of a randomised trial whose primary comparison was negative, on non-randomly-sampled MRD data (249 of 439 eligible), and the authors present it as support for a hypothesis to be tested in the next protocol rather than as a result. The listing of protocol risk factors is background text in the introduction, not a protocol document.

    Evidence that argues against this (1)
    • The trial's own primary randomised comparison found no difference between the consolidation arms (P = .47), so the interaction rests on a subgroup analysis within a null trial and could be a chance finding. Pieters R et al. 2019
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In Interfant-99, which enrolled 482 patients aged 0-12 months across 17 study groups in 22 countries, all types of MLL/KMT2A rearrangement, very high white blood cell count, age younger than 6 months and poor response to the prednisone prophase were independently associated with inferior outcomes, and a randomised late intensification course with high-dose cytarabine and methotrexate gave no benefit (4-year disease-free survival 60.9% versus 57.0%, P = 0.81).

    PMID 17658395DOI
    What this evidence cannot show

    Abstract only; not open access and no PMCID returned by the Europe PMC core record. The abstract says very high white blood cell count without stating a threshold, so it does not itself establish where the 300 x10^9/L line came from. The randomised subset was 191 of 445 patients in complete remission, a minority.

    Evidence that argues against this (1)
    • Interfant-06, which succeeded this protocol and used its prognostic factors to build the risk model, produced whole-cohort outcomes that the authors state did not improve on Interfant-99 - so the factors were correctly identified and acting on them did not help. Pieters R et al. 2019
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    A single course of sequential blinatumomab added to Interfant-06 chemotherapy led to an improvement in 2-year disease-free survival to 82% compared with 49% from the historical chemotherapy-only approach in infant KMT2A-rearranged B-ALL.

    PMID 40584390DOI
    What this evidence cannot show

    A narrative review reporting a single-arm study against a historical comparator, not a randomised result; the randomised test is Interfant-21, which has not read out. Reviews are used here as a map rather than as primary evidence, and the figures are quoted at second hand.

    Evidence that argues against this (1)
    • The historical comparator itself is heterogeneous: Interfant-06's 6-year event-free survival was 49.4% in West European and North American groups against a whole-cohort 46.1%, a 10-12 percentage-point regional spread. A single-arm result compared against a pooled historical figure is compared against a number that does not describe any one population. Pieters R et al. 2019
  6. E-protocol-document-retrievableestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.protocol

    The Interfant-06 treatment protocol document, SKION national version 01 of April 2006, is freely retrievable as a 90-page PDF and specifies dose, schedule, phase and toxicity rules for each agent in the backbone; it is a different object from the Journal of Clinical Oncology results paper, which is not open access.

    What this evidence cannot show

    Fetched 2026-09-20: HTTP 200, 890,629 bytes. **This is the SKION NATIONAL version of an international protocol and was amended mid-trial**, so doses recorded from it are as-planned in one country rather than as-delivered across the study. A protocol states intent; the JCO paper reports what happened, and the two should not be merged. **Established for what the document contains**, not for what patients received.

    Evidence that argues against this (1)
    • A protocol document is not evidence of outcome and is weaker than the published results for anything a reader wants to conclude about efficacy. It is authoritative only about what was specified - and where the protocol's own statements have since been overtaken, as with its mitoxantrone cardiotoxicity guidance, following it uncritically would propagate an error that the results literature has corrected. Feijen EAM et al. 2019
Connected to