Oncofusion

start / graph / concept-total-body-irradiation-and-the-age-four-boundary

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q53-transplant

Total body irradiation and the age-four boundary - the best conditioning in childhood ALL has never been tested in the children who most need a transplant

also called TBI · total body irradiation · fractionated 12 Gy TBI · TBI/etoposide conditioning · irradiation-free conditioning · chemoconditioning · TBI age cut-off

Whole-body radiation before a transplant works better than drugs alone in children with leukaemia - it was tested in a large randomised trial and it won clearly. But the trial only enrolled children aged four and over, because radiation damages a growing brain and body, so nobody knows whether it would also win in babies, and babies are given the drug regimen that lost.

The same thing, for a scientist

FORUM randomised 417 children aged 4-21 years at HSCT to fractionated 12 Gy TBI plus etoposide versus fludarabine/thiotepa with busulfan or treosulfan. It stopped early on its futility rule: 2-year overall survival was 0.91 (95% CI 0.86-0.95) with TBI versus 0.75 (0.67-0.81) with chemoconditioning, relapse 0.12 versus 0.33, and treatment-related mortality 0.02 versus 0.09 - all favouring irradiation, and the authors recommended TBI plus etoposide for patients over 4 years old [E1]. The eligibility floor of 4 years is the whole point of this entry. Infants are excluded from TBI on developmental grounds that predate the trial and that the Interfant-06 protocol states explicitly: 'Total body irradiation is not used because of its neurotoxic side effects in these young children' [E2]. So the group with the worst prognosis in paediatric ALL receives, by protocol, the conditioning strategy that the only randomised comparison found inferior - and the decision not to test the alternative in them was made on toxicity grounds, not on evidence of equal efficacy.

Go deeper
What FORUM actually showed, and to whom1 deeper
TBI 12 Gy + etoposidechemoconditioning (FLU/THIO + BU or TREO)
n randomised212201
2-year overall survival0.91 (0.86-0.95)0.75 (0.67-0.81)
2-year cumulative relapse0.12 (0.08-0.17)0.33 (0.25-0.40)
2-year treatment-related mortality0.02 (<0.01-0.05)0.09 (0.05-0.14)

All p-values favoured TBI; the stopping rule was applied on 2019-03-31 and the trial's conclusion is a recommendation for 'patients > 4 years old with high-risk ALL undergoing allogeneic HSCT' 1.

Read the eligibility line, not only the result: 'Patients <= 18 years at diagnosis, 4-21 years at HSCT'. Every number above is from a child who was at least four years old on the day of transplant. The relapse difference - 0.12 against 0.33, nearly a threefold gap - is the single most consequential conditioning finding in paediatric ALL, and it has no infant arm.

The exclusion is a value judgement and it is stated as one

Interfant-06's own reasoning, from section 1.3.9: 'Two recent publications might suggest that the use of SCT contributed to a favourable outcome in infant ALL but these studies did not have a control arm in which patients only received chemotherapy and the data were not corrected for waiting time to SCT and in one of these studies total body irradiation was used which leads to severe late effects in infants (Sanders 2005, Kosaka 2004).' And then, flatly: 'Total body irradiation is not used because of its neurotoxic side effects in these young children.' 2

The study it names is real and its results were good: 40 infants transplanted at under 12 months, 39 of 40 conditioned with cyclophosphamide plus TBI, disease-free survival 76% in first remission, and 'having the MLL gene was not a factor predictive of outcome' 3. The infant TBI experience exists, it is favourable on the efficacy axis, and it was set aside on the late-effects axis.

That is a defensible trade. It is not a finding, and the field has since stopped generating the data that would let anyone revisit it.

What the late effects actually look like, in the children who did get irradiated young

Seventeen survivors of transplant performed at under 3 years of age, eleven of them TBI-conditioned, assessed at 3.25 to 22.33 years after transplant 4:

abnormalityprevalence
growth hormone deficiency59%
dyslipidaemia59%
hypothyroidism35%
osteochondromas24%
decreased bone mineral density24%
second malignancy2 patients

And the part that cuts the other way, in the authors' own words: neuropsychological testing revealed 'average intelligence quotient with attention deficits and other weaknesses for most patients', quality of life did not differ from population norms, and 'even in this group of patients where the majority was exposed to TBI at a very young age, most are functioning at an average or above-average level.'

So the endocrine and metabolic burden is heavy and near-universal, and the neurocognitive catastrophe that is the stated reason for avoiding TBI in infants was not observed in this small series. Seventeen patients, 17 of 33 eligible, recruited across two decades of changing technique - far too few to overturn the policy, and enough to say the policy's stated premise has thin supporting data.

Is a busulfan exposure equivalent to irradiation? One dataset says maybe

The FORUM pharmacokinetic substudy makes a claim worth isolating: patients whose cumulative busulfan exposure fell inside the optimal window of 73.3-98.0 mg*h/L 'showed no differences in EFS, GRFS, overall survival, or relapse with post hoc matched total-body irradiation (TBI) recipients (n = 51 in each group)', and the authors suggest therapeutic drug monitoring 'might be validated for use as an alternative to irradiation' 5.

If that holds, the FORUM headline is partly a dosing result rather than a modality result: chemoconditioning lost because a large share of patients were outside the busulfan window, not because alkylation is intrinsically weaker than radiation. That would be a very good outcome for infants, who cannot have the radiation.

It should be held loosely. It is a post hoc matched comparison of 51 against 51 inside a trial whose randomised answer went the other way, with the window derived in the same dataset. But it is the most direct argument available that infant chemoconditioning can be made to work, and it points at a concrete, cheap intervention: mandatory busulfan exposure targeting, which Interfant-06 made optional 2.

What is bounded here and what is not

Bounded. FORUM's floor is 4 years at HSCT, stated in its own eligibility criteria 1. The standing review of irradiation-free conditioning states the boundary as a field norm: TBI 'is generally regarded as the gold-standard for children >4 years of age with acute lymphoblastic leukaemia' 6.

Not bounded, and deliberately not asserted. This entry does not claim that no infant anywhere is given TBI, or that no guideline permits it. The exact threshold varies between 2, 3 and 4 years across institutions and eras, and a systematic review of institutional practice was not performed by this beat. What is established is that (a) the randomised evidence base starts at 4 years, (b) Interfant-06 forbids TBI by protocol and states its reason, and (c) the historical infant TBI series that exist are pre-2005, single-centre and uncontrolled.

Facts and figures
modalityfractionated total body irradiation, 12 Gy, with etoposide
randomised evidence age range4-21 years at HSCT (FORUM eligibility); no randomised TBI-versus-chemoconditioning comparison exists below 4 years
infant policyInterfant-06 forbids TBI by protocol: 'Total body irradiation is not used because of its neurotoxic side effects in these young children'
historical infant tbi experienceSanders 2005: 40 infants, 39 conditioned with cyclophosphamide plus TBI, DFS 76% in CR1, MLL rearrangement not prognostic; uncontrolled, single centre, 1982-2003
the structural problemThe group with the worst prognosis in paediatric ALL receives, by protocol, the conditioning modality that the only randomised comparison found inferior - and the decision not to test the alternative in them rests on late-effects reasoning rather than on evidence of equivalent efficacy.
boundednessFORUM's age floor of 4 years at HSCT is stated in its published eligibility criteria and is the bound on every number in this entry. This entry does NOT claim that no infant anywhere receives TBI; institutional thresholds of 2, 3 and 4 years all exist and were not systematically surveyed by this beat.
forum result
{
  "n_randomised": 417,
  "tbi_2y_os": "0.91 (95% CI 0.86-0.95)",
  "chemo_2y_os": "0.75 (95% CI 0.67-0.81)",
  "tbi_2y_relapse": "0.12 (0.08-0.17)",
  "chemo_2y_relapse": "0.33 (0.25-0.40)",
  "tbi_2y_trm": "0.02 (<0.01-0.05)",
  "chemo_2y_trm": "0.09 (0.05-0.14)",
  "median_followup_years": 2.1,
  "stopped_early": "futility rule applied 2019-03-31"
}
late effects under three years
{
  "growth_hormone_deficiency_pct": 59,
  "dyslipidaemia_pct": 59,
  "hypothyroidism_pct": 35,
  "osteochondromas_pct": 24,
  "low_bone_mineral_density_pct": 24,
  "second_malignancies": 2,
  "n_evaluated": 17,
  "n_eligible": 33,
  "neurocognitive": "average IQ with attention deficits; quality of life not different from population norms"
}
How this could bite us back
  • The children who need the best conditioning are given the conditioning that lost. There is no infant arm of FORUM and there will not be one, so the gap between 0.12 and 0.33 relapse is a permanent unknown for the population with the worst prognosis.
  • The exclusion propagates. Once infants are out of the randomised trial, they are also out of every subsequent secondary analysis, every exposure-response substudy and every regulatory label derived from it - which is why the treosulfan approval stops at 1 year and the busulfan label's dosing sentence starts at 12 kg.
  • Avoiding radiation does not avoid late effects; it substitutes them. Busulfan and melphalan produce infertility, growth failure and second malignancy through alkylation rather than through ionising damage, and the infant regimen contains three alkylators in eight days.
  • The stated reason for the policy is the endpoint with the weakest data. Endocrine and metabolic late effects after infant TBI are documented and severe; the neurocognitive catastrophe that the protocol names as its reason is supported by 17 survivors in whom it was not found.
  • An early-stopped trial with 2.1 years of follow-up cannot see the harm the debate is about. FORUM measured efficacy and stopped; the late-effects arm of the same comparison will not read out for twenty years, by which time the conditioning it compared will be obsolete.
Still open
  1. Would low-dose or organ-sparing irradiation - total marrow and lymphoid irradiation, for instance - recover part of the relapse advantage in infants without the growth and endocrine cost of 12 Gy whole-body exposure? This is the only route by which the FORUM result could ever reach this population.
  2. Is the neurotoxicity premise on which infant TBI was abandoned actually supported? The one long-term neurocognitive series in children transplanted under 3 years, most of them irradiated, reports average IQ and normal quality of life in 17 survivors - which is thin evidence in both directions and is currently the only evidence there is.
  3. If busulfan exposure targeting can make chemoconditioning indistinguishable from TBI in matched older children, what fraction of the infants transplanted on Interfant-06 were inside that window? The protocol made the measurement optional, so this may be unanswerable from the trial data.
  4. Does the relapse gap FORUM found (0.12 vs 0.33) depend on fusion status? A KMT2A-rearranged subgroup analysis of FORUM has not been published, and the entire rationale for conditioning choice in this graph's index disease turns on whether it does.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    FORUM randomised 417 patients aged 4-21 years at HSCT to fractionated 12 Gy total body irradiation with etoposide versus fludarabine, thiotepa and either busulfan or treosulfan; the futility stopping rule was applied and 2-year overall survival was 0.91 (95% CI 0.86-0.95) with TBI versus 0.75 (0.67-0.81) with chemoconditioning, cumulative relapse 0.12 versus 0.33 and treatment-related mortality 0.02 versus 0.09, leading the authors to recommend TBI plus etoposide for patients over 4 years old.

    Total Body Irradiation or Chemotherapy Conditioning in Childhood ALL: A Multinational, Randomized, Noninferiority Phase III Study. · Peters C, Dalle JH, Locatelli F, Poetschger U, Sedlacek P, Buechner J, Shaw PJ, Staciuk R, et al. · Journal of Clinical Oncology 2021
    PMID 33332189DOI
    What this evidence cannot show

    Median follow-up was only 2.1 years at the reported analysis, and the trial stopped early, which inflates the apparent effect size relative to a completed trial. Late effects of TBI accrue over decades and cannot appear at 2 years, so the comparison as published measures only the efficacy half of the trade-off the whole debate is about. Eligibility was 4-21 years at HSCT with an HLA-compatible related or unrelated donor and complete remission pre-HSCT, so nothing in it applies to infants, to mismatched or haploidentical grafts, or to patients transplanted with active disease.

    Evidence that argues against this (1)
    • Within the same trial's pharmacokinetic substudy, patients whose cumulative busulfan exposure fell in the optimal 73.3-98.0 mg*h/L window were indistinguishable from post hoc matched TBI recipients (n=51 each) on EFS, GRFS, overall survival and relapse, and the authors propose that busulfan therapeutic drug monitoring 'might be validated for use as an alternative to irradiation'. On that reading, the trial's headline is partly a measure of how many patients missed the busulfan window rather than an intrinsic superiority of radiation. Ben Hassine K et al. 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.book

    Interfant-06 prohibits total body irradiation in its conditioning regimen, stating 'Total body irradiation is not used because of its neurotoxic side effects in these young children', specifies busulfan/cyclophosphamide/melphalan instead, and separately notes that one of the two studies suggesting a transplant benefit in infants used TBI 'which leads to severe late effects in infants'.

    INTERFANT-06: International collaborative treatment protocol for infants under one year with acute lymphoblastic or biphenotypic leukemia (protocol document, SKION version 01, April 2006) · Pieters R, de Lorenzo P, Valsecchi MG, Peters C, and the Interfant-06 study group · Interfant-06 study group / SKION protocol document 2006
    What this evidence cannot show

    Sections 1.3.9 and 6.5 of the 2006 SKION protocol document, retrieved and read on 2026-09-20. The protocol cites no primary study for the neurotoxicity claim at the point it makes it; the supporting reference given a paragraph earlier is to a study (Sanders 2005) whose own long-term neurocognitive data, published separately, do not show the catastrophic picture the phrase implies. This is a protocol writing committee's judgement, correctly labelled as such, not a cited finding.

    Evidence that argues against this (1)
    • The long-term follow-up of children transplanted under 3 years of age, the majority of them TBI-conditioned, reports average intelligence quotient with attention deficits rather than global impairment, quality of life no different from population norms, and the authors' summary that 'even in this group of patients where the majority was exposed to TBI at a very young age, most are functioning at an average or above-average level'. The endocrine burden was severe; the neurocognitive premise on which infant TBI is banned is less well supported than the ban implies. Perkins JL et al. 2007
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In 40 infants diagnosed before 12 months of age and transplanted between 1982 and 2003, 39 of whom were conditioned with cyclophosphamide plus total body irradiation, disease-free survival was 76% for those transplanted in first remission versus 45% in CR2/CR3 and 8% in relapse; 26 of 33 with cytogenetic data had MLL gene rearrangement, 14 of these were in CR1 and 11 survive in remission, and MLL rearrangement was not predictive of outcome.

    Allogeneic hematopoietic cell transplantation for infants with acute lymphoblastic leukemia. · Sanders JE, Im HJ, Hoffmeister PA, Gooley TA, Woolfrey AE, Carpenter PA, Andrews RG, Bryant EM, et al. · Blood 2005
    PMID 15637143DOI
    What this evidence cannot show

    Single centre, 40 patients accrued over 21 years, no chemotherapy-only comparator, and no correction for waiting time to transplant - so the CR1 group is selected for having survived long enough to be transplanted in first remission, which is precisely the guarantee-time bias the Interfant investigators name when they decline to act on this study. The 76% DFS rests on 17 CR1 patients. Cytogenetics were available for only 33 of 40. The authors' own conclusion that 'consideration for transplantation in CR1 is warranted' is not a comparative claim.

    Evidence that argues against this (1)
    • The largest retrospective intergroup analysis of 11q23-rearranged childhood ALL reached the opposite conclusion in the biggest subgroup: among 256 patients with t(4;11), ANY type of transplantation was associated with significantly WORSE disease-free survival (HR 1.61, 95% CI 1.10-2.35, p=0.014) and overall survival (HR 1.76, 1.08-2.45, p=0.004) than chemotherapy alone, and even HLA-matched related or unrelated donor transplantation tended to be associated with a worse outcome than chemotherapy alone. Pui CH et al. 2002
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In 17 evaluated survivors of haematopoietic cell transplant performed for ALL or AML at less than 3 years of age (11 TBI-conditioned, surviving 3.25-22.33 years), prevalence was 59% for growth hormone deficiency, 59% for dyslipidaemia, 35% for hypothyroidism, 24% for osteochondromas and 24% for decreased bone mineral density, with two second malignancies; neuropsychological testing revealed average IQ with attention deficits, and quality of life did not differ from population norms.

    Long-term follow-up of children who underwent hematopoeitic cell transplant (HCT) for AML or ALL at less than 3 years of age. · Perkins JL, Kunin-Batson AS, Youngren NM, Ness KK, Ulrich KJ, Hansen MJ, Petryk A, Steinberger J, et al. · Pediatric Blood & Cancer 2007
    PMID 17474113DOI
    What this evidence cannot show

    Only 17 of 33 eligible patients were evaluated, and survivors who were doing badly, who had died, or who were lost to follow-up are systematically absent - a survivorship bias that pushes every reassuring finding upward and the neurocognitive conclusion in particular. Transplants span 1981-2000 with heterogeneous conditioning and radiation technique. Eleven of 17 received TBI, so the six who did not are folded into the same prevalence figures. No control group. 'Average IQ' is a group mean and does not exclude severe impairment in individuals.

    Evidence that argues against this (1)
    • The graph already records the late-effects burden in infant ALL survivors at higher severity than this small series suggests, and the same paper's own framing is double-edged: 'typical late effects seen after radiation exposure are common', with dyslipidaemia in more than half and a stated concern about early cardiovascular disease. A study reporting 59% growth hormone deficiency and two second malignancies in 17 survivors is not a reassuring study; it is a study in which the neurocognitive endpoint specifically failed to show the expected harm. Perkins JL et al. 2007
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Within FORUM, 145 paediatric ALL patients receiving fludarabine-busulfan-thiotepa conditioning had an optimal cumulative busulfan AUC of 73.3-98.0 mg*h/L, and those inside that window showed no differences in event-free survival, GVHD-free relapse-free survival, overall survival or relapse compared with post hoc matched total body irradiation recipients (n=51 in each group).

    Refining busulfan exposure enhances pediatric ALL HSCT outcomes: insights from the International FORUM study. · Ben Hassine K, Gloor Y, Dupanloup I, Uppugunduri CRS, Mlakar V, Gonzales F, Guengoer T, Ifversen M, et al. · Blood Advances 2026
    PMID 41779961DOI
    What this evidence cannot show

    A post hoc matched comparison of 51 versus 51 patients, nested inside a randomised trial whose primary result went the other way, using an exposure window derived from the same dataset. Matching cannot recover randomisation, and 'no difference detected' in 102 patients is weak evidence of equivalence - the confidence intervals would admit substantial differences. The authors phrase the conclusion conditionally ('might be validated'), and this entry does not strengthen it. The population is aged 0.5-19.5 years and is not stratified by age, so the claim has not been tested in infants.

    Evidence that argues against this (1)
    • The randomised comparison in the same trial, which does not depend on post hoc matching, found TBI plus etoposide superior on overall survival (0.91 vs 0.75), relapse (0.12 vs 0.33) and treatment-related mortality (0.02 vs 0.09), and the trial was stopped early for futility of the chemoconditioning arm. A matched-subset analysis showing no difference in 51 pairs does not overturn a randomised result in 417 patients; at most it suggests a mechanism for part of the difference. Peters C et al. 2021
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    TBI-based conditioning is generally regarded as the gold standard for children over 4 years of age with acute lymphoblastic leukaemia, and survivors suffer long-term sequelae including second malignancies and neurocognitive, endocrine and cardiometabolic effects, which motivates continued optimisation of irradiation-free myeloablative conditioning.

    Total Body Irradiation Forever? Optimising Chemotherapeutic Options for Irradiation-Free Conditioning for Paediatric Acute Lymphoblastic Leukaemia. · Ben Hassine K, Powys M, Svec P, Pozdechova M, Versluys B, Ansari M, Shaw PJ · Frontiers in Pediatrics 2021
    PMID 34956984DOI
    What this evidence cannot show

    A narrative review, cited here for its explicit statement of the age-4 field norm and as a map of the alternatives, not as primary evidence. It does not enumerate institutional age thresholds, and the boundary is stated as a convention rather than derived from a study of outcomes by age. Published 2021, so it predates the FORUM busulfan exposure analysis and the treosulfan US approval.

    Evidence that argues against this (1)
    • The age-4 norm as stated is partly circular: it is the gold standard for children over 4 because that is the population FORUM randomised, and FORUM randomised that population because the norm already excluded younger children. No study cited by this beat derives the threshold from an age-stratified comparison of TBI outcomes, so the boundary is a convention with a physiological rationale rather than a measured cut-point. Peters C et al. 2021
Connected to