AML Relapse Risk Stratification
Characterize relapse-associated molecular signatures at diagnosis to support risk stratification beyond standard classifiers
The Challenge
High Relapse Rates After Remission
A majority of AML patients achieve complete remission with induction chemotherapy, but a substantial proportion relapse within two years. Relapsed AML is often incurable — preventing relapse is critical.
Consolidation Intensity Dilemma
High-intensity consolidation carries meaningful treatment-related toxicity, while lower-intensity approaches carry higher relapse risk. Matching consolidation intensity to molecular risk remains a central challenge in AML management.
ELN Risk Stratification Is Insufficient
European LeukemiaNet (ELN) risk categories (favorable, intermediate, adverse) are based on cytogenetics and mutations, but a substantial proportion of intermediate-risk patients relapse, highlighting the need for more granular molecular stratification.
Minimal Residual Disease Comes Too Late
MRD provides relapse risk information but is available weeks after diagnosis. Earlier molecular characterization at diagnosis would provide timelier risk stratification.
"Our cohort included patients classified as favorable risk by standard criteria who nevertheless relapsed. Retrospective molecular analysis revealed relapse-associated signatures that were not captured by conventional classifiers — suggesting that molecular stratification at diagnosis could have informed a different risk assessment."
The Helomnix Solution
Helomnix integrates multi-omics data (RNA-seq, proteomics, exome) at diagnosis to characterize relapse risk more comprehensively than ELN risk categories alone. In this context, relapse risk characterization refers to identifying molecular patterns present at diagnosis that have been historically associated with disease recurrence, independent of treatment decisions. Our models identify molecular signatures of residual leukemic stem cells, therapy resistance pathways, and immune evasion that are associated with relapse.
Patients with elevated molecular risk signatures (i.e., molecular patterns historically associated with disease recurrence in reference cohorts) can be distinguished from those with favorable molecular profiles at diagnosis. This molecular characterization provides an additional stratification layer beyond standard risk classifiers, supporting more granular risk assessment for downstream translational and clinical research workflows.
This at-diagnosis molecular characterization provides a complementary stratification layer, supporting more refined risk assessment in contexts where standard classifiers leave substantial heterogeneity unresolved.
Unique Differentiator
We integrate leukemic stem cell signatures, therapy resistance pathway activity, and immune dysfunction features — molecular dimensions that standard cytogenetic classifiers do not capture.
How It Works
Diagnosis Multi-Omics Profiling
Obtain bulk RNA-seq, exome sequencing, and proteomics (if available) from diagnostic bone marrow or peripheral blood sample. Standard of care already collects these.
Relapse Risk Modeling
Models trained on curated AML cohorts with long-term outcome annotations identify molecular features associated with relapse, including leukemic stem cell signatures, therapy resistance pathway activity, and immune dysfunction markers.
Risk Stratification
Classify patients into molecular risk tiers (elevated, intermediate, favorable) based on integrated molecular features. Complements existing ELN risk categories with additional molecular dimensions.
Molecular Risk Report
Deliver a molecular risk characterization report including integrated risk tier, key molecular features driving the assessment, Digital Twin Map positioning, and survival association analysis.
Real-World Application
An AML patient classified as intermediate-risk by standard ELN criteria achieved complete remission. Standard classifiers provided ambiguous risk assessment, leaving molecular risk unresolved.
Before
Standard approach: ELN intermediate-risk provides limited molecular resolution, with a significant proportion of patients in this category experiencing relapse despite achieving initial remission.
After
Integrated multi-omics profiling identified elevated leukemic stem cell and therapy resistance signatures. Digital Twin Map projection positioned the patient within a high-risk molecular neighborhood associated with relapse in the reference cohort.
Outcome
Molecular risk characterization identified elevated risk features not captured by ELN classification alone, providing additional stratification evidence that complemented the clinical assessment.
Value to Your Organization
Risk Characterization Timing
Characterize molecular risk at diagnosis rather than waiting for post-induction assessments, providing earlier risk stratification for translational and research workflows.
Stratification Resolution
Identify molecular risk features missed by ELN alone, providing complementary stratification that resolves heterogeneity within standard risk categories.
Subgroup Definition
Define molecular risk subgroups within ELN categories, supporting more granular cohort definition for translational research and trial design.
Our Methodology
Data Inputs
- Diagnostic bone marrow or peripheral blood sample
- Bulk RNA-seq (required)
- Whole exome sequencing (mutations, CNVs)
- Proteomics (optional, enhances characterization)
- Clinical data (age, WBC count, prior MDS, cytogenetics)
- Treatment and outcomes (for model training/validation)
AI/ML Techniques
- Leukemic stem cell signature scoring
- Therapy resistance pathway analysis
- Immune dysfunction signatures (T-cell exhaustion, checkpoint)
- Digital Twin Map projection for risk neighborhood identification
- Ensemble machine learning models trained on curated AML cohorts with outcome annotations
- Integration with ELN risk categories (cytogenetics + mutations)
- Survival association analysis for relapse-free and overall survival
Deliverables
- Relapse risk score (structured risk assessment)
- Risk stratification (High/Intermediate/Low)
- Molecular risk tier report complementing standard risk classifiers
- Key molecular features driving risk assessment (stem cell signatures, resistance, immune)
- Digital Twin Map showing patient position vs. relapse/no-relapse cohorts
- Survival analysis (RFS and OS characterization)
- Risk tier contextualization relative to reference AML cohort outcomes
- Structured molecular risk characterization report
Discuss a Translational Application
We welcome discussions about how this approach can support your translational research.