Highlight
– Replacing the ML‑DS 2006 reduced‑intensity induction with CPX‑351 (liposomal cytarabine:daunorubicin/cytarabine formulation) in the ML‑DS 2018 trial produced a significantly lower 24‑month event‑free survival (EFS) — 69% versus 90% (P <.001) — leading to early trial suspension.
– Despite higher relapse frequency, most relapsed patients achieved remission with salvage therapy and overall survival (OS) at 24 months remained similar to historical controls (88% vs 92%, P = .612).
– CPX‑351 showed an excellent safety profile with no treatment‑related mortality in this highly treatment‑sensitive population.
– Positive measurable residual disease (MRD) by error‑corrected GATA1 NGS, presence of trisomy 8, and complex karyotype were associated with increased relapse risk.
Background: clinical context and unmet need
Children with Down syndrome (DS) have a distinct predisposition to acute myeloid leukemia (ML‑DS). Historically, ML‑DS is characterized by high cure rates when treated with risk‑adapted chemotherapy; however, children with DS are uniquely susceptible to treatment‑related toxicity and mortality, which has motivated the development of reduced‑intensity regimens. The ML‑DS 2006 protocol successfully maintained excellent event‑free and overall survival while reducing toxicity, establishing a clinical paradigm of de‑escalation tailored to this population.
CPX‑351 is a liposomal formulation designed to maintain a fixed synergistic molar ratio of cytarabine and daunorubicin, with altered pharmacokinetics and tissue distribution compared with free drugs. In older adults with secondary AML, CPX‑351 improved outcomes versus conventional 7+3 therapy. Translating this agent into pediatric ML‑DS aimed to preserve antileukemic potency while minimizing systemic toxicity; the ML‑DS 2018 trial tested this strategy by replacing the reduced‑intensity anthracycline/cytarabine induction used in the prior ML‑DS 2006 protocol with CPX‑351.
Study design and interventions
The ML‑DS 2018 trial (EudraCT: 2018‑002988‑25) enrolled children aged 6 months to 6 years with ML‑DS. The trial replaced the intensity‑reduced induction and reinduction backbone of ML‑DS 2006 (cytarabine, idarubicin ± etoposide) with CPX‑351 administered at 66 U/m² for three days in course 1 and two days in course 2. Risk stratification relied on flow cytometric MRD assessment after the first induction: patients with higher risk features received higher‑dose consolidation with high‑dose cytarabine (3 g/m² every 12 hours), while standard‑risk patients received intermediate‑dose cytarabine (1 g/m² every 12 hours).
The trial was designed to maintain the excellent EFS of the ML‑DS 2006 experience while attempting to reduce toxicity. A planned per‑protocol interim analysis was conducted after accrual of 35 patients; enrollment was halted due to an unexpectedly high relapse rate.
Key findings and detailed results
Enrollment and early termination: Thirty‑five patients were accrued before the trial was stopped following the interim analysis. The safety profile of CPX‑351 was favorable: there were no treatment‑related deaths reported, and treatment‑related toxicities were generally manageable — an important consideration in DS, where chemotherapy sensitivity is heightened.
Event‑free survival and overall survival: The pre‑specified per‑protocol interim analysis demonstrated a substantially lower 24‑month EFS compared with the historic ML‑DS 2006 trial: 69% (ML‑DS 2018) versus 90% (ML‑DS 2006), P <.001. In contrast, 24‑month OS was comparable: 88% in the ML‑DS 2018 cohort versus 92% in ML‑DS 2006 (P = .612). The dissociation of lower EFS but preserved OS was driven by a high proportion of relapsed patients who subsequently achieved remission with salvage therapy, indicating retained chemosensitivity of relapsed ML‑DS disease.
Relapse and salvage response: The trial reported an unexpectedly high relapse rate after CPX‑351 induction/reinduction. Most relapsed patients responded to salvage regimens and were rescued, which explains why OS did not fall in parallel with EFS. The nature of salvage therapies, timing of relapse, and long‑term outcomes beyond 24 months will be important to define in follow‑up reports.
Prognostic biomarkers: Two biologic features emerged as risk markers for relapse. First, detectable MRD by error‑corrected GATA1 next‑generation sequencing after induction was associated with increased relapse risk. GATA1 mutations are canonical in ML‑DS and sensitive MRD detection by deep sequencing provided prognostic discrimination beyond flow cytometry. Second, the cytogenetic presence of trisomy 8 or a complex karyotype conferred higher relapse risk. These findings point to biologic heterogeneity within ML‑DS and support integrating genomic MRD and karyotype into risk stratification.
Toxicity and mortality: Importantly, CPX‑351 was well tolerated in this trial population, with no treatment‑related mortality reported — a key goal when treating children with DS. Overall toxicity appeared favorable when compared to historical expectations of anthracycline and cytarabine–based regimens in ML‑DS, though direct toxicity comparisons with ML‑DS 2006 were limited by early trial termination and small sample size.
Expert commentary: interpreting the results and biological plausibility
The ML‑DS 2018 results carry two central messages. First, dose and formulation matter in ML‑DS: although CPX‑351 offers pharmacologic advantages and a favorable toxicity profile, its use in a reduced‑intensity induction context appears to have underdelivered antileukemic efficacy for a subgroup of patients, producing higher relapse rates. Children with DS often require treatment de‑escalation to reduce morbidity, but overt de‑escalation runs the risk of under‑treating disease; ML‑DS is curable because therapy has been optimized to this delicate balance.
Mechanistic hypotheses include altered pharmacodynamics of liposomal drug delivery in very young children and in DS, differences in leukemic cell sensitivity to drug ratios versus absolute dose, and potential reductions in peak free drug exposure that could diminish cytotoxicity even if overall toxicity is lowered. CPX‑351 was developed and validated in older adults with secondary AML — a biologically distinct population — and direct translation to ML‑DS may require re‑calibrated dosing or schedule adaptations rather than one‑for‑one substitution.
Second, the success of salvage therapy in many relapsed patients suggests maintained chemosensitivity of ML‑DS and supports aggressive, effective salvage strategies. That said, relapse carries morbidity and the risk of later treatment‑related complications; preventing relapse remains preferable to rescue.
Biomarker insights are clinically actionable. Error‑corrected GATA1 NGS MRD appeared predictive of relapse and could be integrated into future risk‑adapted strategies to identify infants and children who need intensified consolidation or alternative approaches. Cytogenetic risk features (trisomy 8, complex karyotype) likewise identify patients at higher relapse risk and may guide therapy intensification or closer monitoring.
Limitations and generalizability
Key limitations include the small sample size (n = 35) and early termination, which constrain the precision of effect estimates and the ability to perform robust subgroup analysis. The analysis compared ML‑DS 2018 outcomes with the historical ML‑DS 2006 dataset rather than a contemporaneous randomized control arm; changes in supportive care, diagnostic sensitivity (e.g., MRD methods), and salvage approaches over time may influence comparisons. Longer follow‑up is required to understand late relapses, late toxicity, and durable survival beyond 24 months.
Furthermore, while the safety signal was favorable, detailed toxicity tables and grade‑specific events will be important to contextualize the trade‑off between reduced acute toxicity and increased relapse. The balance between efficacy and toxicity must be individualized for DS patients, and any broad adoption of CPX‑351 in this population should await dose optimization and validation in larger, ideally randomized, studies.
Implications for clinical practice and future research
Clinical implications:
– CPX‑351 should not replace the established reduced‑intensity induction of ML‑DS 2006 in routine practice until optimized dosing and schedules are tested; the ML‑DS 2018 results caution against wholesale substitution even when safety seems improved.
– Incorporation of highly sensitive MRD assays (error‑corrected GATA1 NGS) into routine post‑induction assessment should be considered to refine risk stratification and trigger early therapy modification for MRD‑positive patients.
– Children with trisomy 8 or complex karyotype should be considered for closer surveillance and potential intensification in future protocols.
Research priorities:
– Dose‑finding studies of CPX‑351 in ML‑DS to identify schedules that preserve low toxicity while restoring antileukemic potency.
– Prospective, randomized comparisons or thoughtfully matched registry analyses versus the ML‑DS 2006 backbone to conclusively determine benefit/risk.
– Further validation of error‑corrected GATA1 NGS MRD and cytogenetic markers as stratification tools to guide tailored therapy and to evaluate pre‑emptive interventions.
– Pharmacokinetic and pharmacodynamic studies in young children with DS to understand liposomal drug distribution, metabolism, and target tissue exposure.
Conclusion
The ML‑DS 2018 trial demonstrates that in a treatment‑sensitive pediatric leukemia population, a safer pharmacologic formulation is not automatically better if antileukemic efficacy is compromised. Substituting reduced‑intensity induction with CPX‑351 produced fewer immediate toxicities and no treatment‑related mortality, but at the cost of a significantly higher relapse rate and lower EFS. The good news is preserved OS due to effective salvage, and the identification of MRD (GATA1 NGS), trisomy 8, and complex karyotype as relapse predictors provides a pragmatic route for improved risk adaptation. Future efforts must focus on defining optimal CPX‑351 dosing, integrating sensitive genomic MRD assays into trials, and ensuring that treatment de‑escalation in ML‑DS maintains the delicate balance between safety and cure.
Funding and trial registration
Trial registration: EudraCT: 2018‑002988‑25. Funding details are reported in the primary publication (Laszig et al., Blood 2025).
References
Laszig S, Diederichs A, Salzmann‑Manrique E, Schuschel K, Goncalves‑Dias J, Issa H, Miladinovic M, Rettinger E, Wehner S, Kreyenberg H, Bremm M, Hünecke S, Kerp H, Waack‑Buchholz K, Thol FR, Goemans BF, De Moerloose B, Boztug H, Scheidegger NK, Pawinska‑Wasikowska K, Reinhardt D, Klusmann JH. CPX‑351 in Down syndrome‑associated Myeloid Leukemia: Results and Prognostic Factors from the Phase 3 ML‑DS 2018 Trial. Blood. 2025 Oct 21: blood.2025030775. doi: 10.1182/blood.2025030775. Epub ahead of print. PMID: 41118594.
AI image prompt for thumbnail (for designer/AI generation)
A compassionate clinical scene in a pediatric hematology clinic: a doctor gently reviewing charts with a parent beside a young child with Down syndrome seated on an exam table. Superimpose subtle scientific icons — stylized liposomal drug particles, a double helix, and small chromosome symbols highlighting trisomy 21 and trisomy 8. Use soft, warm lighting, calm blue and teal color palette, and a clean clinical background. Emphasize hope, scientific complexity, and pediatric care.