When infant leukaemia comes back, it comes back early and while treatment is still running
also called relapse infant ALL · time to relapse infant · refractory infant leukaemia · on-treatment relapse
Most relapses in infant leukaemia happen within two years of diagnosis, while the child is still on chemotherapy, and they are usually resistant to everything that follows. That timing is the single most important fact for designing trials in this disease.
The same thing, for a scientist
About two-thirds of infants with KMT2A-rearranged ALL relapse within 2 years of diagnosis **while still on active treatment**, usually progressing to refractory disease with a fatal outcome [E1]. Interfant-06 quantifies the numerator: 213 of 442 patients in first complete remission relapsed, 50 of those relapses involving the CNS [E1]. CCCG-ALL-2015 saw 29 recurrences among 134 infants (21.6%) at a median of 15.7 months (range 1.6-39.2), with 5 relapsing beyond 2 years; the 5-year cumulative incidence of relapse was 23.0% (15.7-30.3) overall, 30.2% (18.8-41.6) in KMT2A-rearranged and 15.9% (6.8-25.0) in KMT2A-germline patients (p = 0.05) [E2]. Congenital ALL is the extreme: 2-year cumulative incidence of relapse 60.0% (SE 9.3) against 34.2% (SE 2.3) in older infants [E3]. Three design consequences follow. Relapse on treatment means the disease is resistant to the regimen it is being given, not that maintenance failed. A 2-year event horizon means single-arm pilots reporting 2-year outcomes are reporting almost the whole event burden, which is unusually favourable for early readout. And a high-risk group whose events occur early cannot be delivered to transplant: in Interfant-06 only 46% of intention-to-transplant high-risk patients actually received one, because of early events [E4].
The numbers1 deeper
| Cohort | Relapse burden | Timing |
|---|---|---|
| Interfant-06 | 213 of 442 in CR1 relapsed; 50 involved CNS 1 | most within 2 years, on treatment 1 |
| CCCG-ALL-2015 | 29 of 134 (21.6%); 5-year CIR 23.0% (15.7-30.3) | median 15.7 months, range 1.6-39.2; 5 beyond 2 years 2 |
| CCCG KMT2A-rearranged | 5-year CIR 30.2% (18.8-41.6) | 2 |
| CCCG KMT2A-germline | 5-year CIR 15.9% (6.8-25.0), p = 0.05 | 2 |
| Congenital ALL, Interfant-99 | 2-year CIR 60.0% (SE 9.3) vs 34.2% (SE 2.3) | 3 |
| Blinatumomab pilot | 4-year DFS 83.3% vs 44.0% historical | median follow-up 4.2 years 5 |
The CCCG relapse rate is conspicuously lower than Interfant's, and the same structural caveat applies as everywhere else in that cohort: it under-enrolled KMT2A-rearranged and under-six-month infants, and 13.4% discontinued protocol therapy, which competes with relapse as an outcome 2.
A bounded unreadable in the relapse literature
The dedicated study of outcome after relapse on Interfant-99 (Driessen et al., Leukemia 2016;30:1184-7 area, PMID 26369984, DOI 10.1038/leu.2015.246) has no abstract in PubMed and none in Europe PMC, and is not open access - checked 2026-09-19 6. It is cited in the Interfant investigators' own review as the source for the statement that relapse usually leads to refractory disease with a fatal outcome 1. So this beat can report that the study exists, that it is the primary source for the post-relapse prognosis this graph asserts, and that its numbers were not obtainable. Nothing about its contents is inferred here.
What this implies for a targeted agent
If relapse happens on treatment, then the agent must work against cells surviving concurrent intensive chemotherapy, not against a quiescent reservoir in an otherwise untreated child. That is a harder test than a relapsed/refractory adult setting, and it is the setting the Interfant-21 and COG trials are actually running. It also means the window for a post-remission targeted agent is narrow: by the time a two-year DFS curve separates, most of the events have already happened.
The converse is the beat's one piece of good news: a disease whose events are front-loaded gives informative readouts fast. The blinatumomab pilot's 4-year DFS of 83.3% at a median 4.2 years of follow-up 5 is close to a mature result on 30 patients, which in most cancers would need a decade.
| metric | relapse incidence and timing |
|---|---|
| population | infants <1 year with ALL |
cohorts — the numbers behind the chart
[
{
"cohort": "Interfant-06",
"in_cr1": 442,
"relapses": 213,
"cns_involved": 50,
"pmid": "40207717",
"title": "KMT2A-rearranged acute lymphoblastic leukemia in infants: current progress and challenges.",
"authors": "Stutterheim J, Van der Sluis IM, Vrenken KS, et al.",
"venue": "Haematologica",
"year": 2025,
"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."
},
{
"cohort": "CCCG-ALL-2015",
"n": 134,
"relapses": 29,
"relapse_pct": 21.6,
"median_months_to_relapse": 15.7,
"range_months": [
1.6,
39.2
],
"cir_5y_pct": 23,
"cir_5y_ci": [
15.7,
30.3
],
"cir_kmt2a_r_pct": 30.2,
"cir_kmt2a_g_pct": 15.9,
"pmid": "38634141",
"title": "Outcome of infants with acute lymphoblastic leukemia treated with the Chinese Children's Cancer Group Acute Lymphoblastic Leukemia 2015 study protocol.",
"authors": "Leung AWK, Cai J, Wan Z, et al.",
"venue": "Haematologica",
"year": 2024,
"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."
},
{
"cohort": "congenital ALL, Interfant-99",
"n": 30,
"cir_2y_pct": 60,
"se": 9.3,
"comparator_cir_2y_pct": 34.2,
"comparator_se": 2.3,
"pmid": "19657114",
"title": "Outcome of congenital acute lymphoblastic leukemia treated on the Interfant-99 protocol.",
"authors": "van der Linden MH, Valsecchi MG, De Lorenzo P, et al.",
"venue": "Blood",
"year": 2009,
"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."
}
]unreadable primary
{
"pmid": "26369984",
"doi": "10.1038/leu.2015.246",
"what_it_is": "outcome of relapsed infant ALL on Interfant-99",
"routes_tried": [
"NCBI efetch",
"Europe PMC core search"
],
"date_checked": "2026-09-19",
"title": "Outcome of relapsed infant acute lymphoblastic leukemia treated on the interfant-99 protocol.",
"authors": "Driessen EM, de Lorenzo P, Campbell M, et al.",
"venue": "Leukemia",
"year": 2016,
"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."
}- Relapse on treatment means any new agent must work against cells already surviving five-drug chemotherapy, which is a harder test than a relapsed/refractory setting where the patient is off therapy.
- Front-loaded events make historical-control comparisons unusually dangerous, because a small difference in how early deaths and discontinuations are handled moves the curve more than the drug does.
- A high-risk stratum in which only 46% reach the allocated transplant delivers an intention, not an exposure, so trials built on that stratum cannot attribute their results to the therapy named in the arm.
- Blinatumomab's marrow benefit came with all four pilot relapses involving the CNS, so relapse prevention in one compartment can become relapse concentration in another.
- What does PMID 26369984 report about survival after relapse in infant ALL? It is the primary source for a claim this graph makes and nobody in the project has read it.
- Why does congenital ALL relapse at 60% at two years when its induction failure rate is no different from older infants? The answer distinguishes a resistant clone from an inadequate consolidation.
- Given that most events occur on treatment, is there any post-remission window in which a targeted agent could act that is not already occupied by intensive chemotherapy?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review
About two-thirds of infants with KMT2A-rearranged ALL relapse within 2 years of diagnosis while still on active treatment, usually leading to refractory disease with a fatal outcome; in Interfant-06, 213 of 442 patients in first complete remission relapsed and 50 of those relapses involved the CNS.
KMT2A-rearranged acute lymphoblastic leukemia in infants: current progress and challenges. · Stutterheim J, Van der Sluis IM, Vrenken KS, Pieters R, Stam RW · Haematologica 2025What this evidence cannot show
A narrative review by the trial's own investigators; the primary Interfant-06 full text is paywalled and the post-relapse outcome source it cites (PMID 26369984) has no retrievable abstract. 'About two-thirds' is a characterisation, not a computed cumulative incidence with an interval. Read in the open-access full text at PMC12399962.
Evidence that argues against this (1)
- CCCG-ALL-2015 reports a 5-year cumulative incidence of relapse of 23.0% overall and 30.2% in KMT2A-rearranged patients, far below two-thirds, though on a cohort that under-enrolled the highest-risk infants and had 13.4% protocol discontinuation. Leung AWK et al. 2024
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In CCCG-ALL-2015, 29 of 134 infants (21.6%) relapsed at a median of 15.7 months (range 1.6-39.2), with 5 relapsing beyond 2 years; 23 relapses were marrow, 3 isolated CNS, 1 combined marrow and CNS and 2 isolated testicular; 5-year cumulative incidence of relapse was 23.0% (95% CI 15.7-30.3) overall, 30.2% (18.8-41.6) with KMT2A rearrangement and 15.9% (6.8-25.0) with germline KMT2A (p = 0.05).
Outcome of infants with acute lymphoblastic leukemia treated with the Chinese Children's Cancer Group Acute Lymphoblastic Leukemia 2015 study protocol. · Leung AWK, Cai J, Wan Z, Qin J, Fang Y, Sun L, et al. · Haematologica 2024What this evidence cannot show
18 of 134 (13.4%) discontinued protocol treatment, 10 on parental request, which competes with relapse and censors it. The cohort under-enrolled infants under six months and KMT2A-rearranged cases by the authors' own account. Deaths in remission (18) also compete. Read in the open-access full text at PMC11290514.
Evidence that argues against this (1)
- With 18 deaths in remission and 18 treatment discontinuations against 29 relapses, competing risks are of the same magnitude as the event of interest, so a cumulative incidence estimate here is highly sensitive to how those competing events were handled. Stutterheim J et al. 2025
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Congenital ALL treated on Interfant-99 had a 2-year cumulative incidence of relapse of 60.0% (SE 9.3) against 34.2% (SE 2.3) in older infants (p < 0.001), with 2-year event-free survival and survival of 20% (SE 9.1).
Outcome of congenital acute lymphoblastic leukemia treated on the Interfant-99 protocol. · van der Linden MH, Valsecchi MG, De Lorenzo P, Moricke A, Janka G, Leblanc TM, et al. · Blood 2009What this evidence cannot show
n = 30 congenital patients, so SE 9.3 on the relapse incidence. Not randomised against older infants; the two groups differ on every baseline factor. Abstract read; full text not retrieved.
Evidence that argues against this (1)
- Induction failure did not differ significantly between congenital and older infants (13% vs 7%, p = 0.14), so the congenital disadvantage is specific to relapse rather than to chemosensitivity, which argues against a simple 'more resistant disease' reading. Pieters R et al. 2007
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In Interfant-06 the 6-year event-free survival of the high-risk group with intention to undergo stem-cell transplantation was 20.9% (SE 3.4), and only 46% of them received a transplant because many had early events.
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. · Pieters R, De Lorenzo P, Ancliffe P, Aversa LA, Brethon B, Biondi A, et al. · J Clin Oncol 2019What this evidence cannot show
An intention-to-treat figure in a non-randomised risk group; the 46% who reached transplant and the 54% who did not are not comparable, so no inference about transplant efficacy can be drawn from it. Abstract read; full text paywalled.
Evidence that argues against this (1)
- That more than half of an intention-to-transplant group never reaches transplant means the arm tests an allocation policy rather than a treatment, and a trial designed this way cannot answer whether transplant helps. Tomizawa D et al. 2020
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
At median follow-up 4.2 years, infants receiving one post-induction blinatumomab course added to Interfant-06 chemotherapy had 4-year disease-free survival of 83.3% against 44.0% in historical controls.
Sustained Benefit of Blinatumomab in Infants With KMT2A-Rearranged ALL: Long-Term Outcomes, Toxicity, and Pharmacokinetics. · Vieira Martins M, de Lorenzo P, Kotecha RS, Attarbaschi A, Escherich G, Nysom K, et al. · J Clin Oncol 2026What this evidence cannot show
30 infants, single-arm, historical comparator selected from Interfant-06. Era effects in supportive care are not controlled, and Interfant-06 itself showed a 10-12 point regional outcome gap on the same protocol. Abstract read; full text paywalled.
Evidence that argues against this (1)
- All four relapses in the pilot involved the CNS despite no CNS involvement at diagnosis, so the marrow benefit may be partly a redistribution of where relapse occurs rather than a reduction in relapse. Stutterheim J et al. 2025
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
The dedicated Interfant-99 study of outcome after relapse in infant ALL has no retrievable abstract through NCBI efetch or Europe PMC and is not open access, checked 2026-09-19.
Outcome of relapsed infant acute lymphoblastic leukemia treated on the interfant-99 protocol. · Driessen EM, de Lorenzo P, Campbell M, Felice M, Ferster A, Hann I, et al. · Leukemia 2016What this evidence cannot show
Bounded unreadable. Title, authors, journal, year and DOI were fetched and cross-checked between efetch ArticleIdList and esummary articleids; the content was not obtained by any route. Nothing about its findings is asserted here. Availability may change.
Evidence that argues against this (1)
- This graph's post-relapse prognosis claims currently rest on a secondary citation of a paper nobody in the project has read, which is exactly the shape of error the citation rules exist to prevent. Stutterheim J et al. 2025
a cellular account of why two-thirds of infants relapse, typically during treatment: the compartment is present before the first dose
relapse incidence, timing and post-relapse prognosis
- supports →Leukaemia diagnosed in the first month of life: a different presentation and a different prognosis
congenital ALL's 60% two-year relapse incidence is the extreme of the gradient
PMID 26369984, the primary post-relapse outcome study, has no retrievable abstract
- related to →Interfant-06 — 651 infants, and the definitive answer that myeloid-style consolidation does not help
213 of 442 in CR1 relapsed; high-risk intention-to-transplant EFS 20.9% with only 46% transplanted
- related to →CNS disease at diagnosis in infants: a real excess, measured by a procedure that partly creates it
50 of 213 relapses involved the CNS