Companion Diagnostic Development
Stratify patients by tumor differentiation stage to support translational subgroup definition
The Challenge
Need for Stratification Biomarkers
Your compound shows variable activity across patients, but you need a companion diagnostic to characterize molecularly defined subgroups. Traditional biomarkers (mutations, IHC) don't capture differentiation biology.
Heterogeneous Treatment Responses
Patients with "the same diagnosis" (e.g., myeloma) show variable outcomes. Why? Different differentiation stages may have different pathway dependencies and functional characteristics.
Expensive CDx Development
Developing companion diagnostics requires substantial investment and extended timelines. Most fail because biomarkers discovered post-hoc don't validate in independent cohorts.
Unclear Biological Rationale
FDA/EMA demand mechanistic rationale for companion diagnostics. Gene signatures without biological context face regulatory skepticism and longer approval timelines.
"Our compound trial showed variable activity in myeloma, but we had no biomarker to characterize which molecularly defined subgroups showed alignment. Developing a post-hoc CDx took years and substantial investment. We should have designed the biomarker upfront."
The Helomnix Solution
Helomnix uses our in-vitro normal B-cell→plasma cell differentiation atlas as a reference to classify tumors by differentiation stage blockade. Tumor samples are profiled with scRNA-seq or bulk RNA-seq and compared to our 4-stage reference atlas (MBC → Pre-PB → PB → PC).
Here, differentiation-based classification means identifying where along the normal B-to-plasma cell trajectory a tumor is arrested, providing a mechanistic basis for subgroup definition. This reveals tumor subtypes based on differentiation biology: e.g., "MBC-like myeloma" vs. "PC-like myeloma" or "early blockade" vs. "late blockade" WM (Waldenström macroglobulinemia). These biologically-defined subtypes represent distinct functional disease states with different pathway dependencies — validated in published WM research.
The differentiation-based classifier becomes your companion diagnostic: classify each patient's tumor by differentiation state, providing mechanistic context for translational interpretation and enrichment design. Mechanistic biological rationale (differentiation stage) strengthens FDA approval case vs. empirical gene signatures.
Unique Differentiator
Differentiation-based CDx has clear biological mechanism (stage of arrest) unlike black-box gene signatures. Published validation in WM (Waldenström macroglobulinemia) research shows MBC-like vs. PC-like subtypes have different outcome profiles and pathway dependencies.
How It Works
Normal Differentiation Reference
Use Helomnix in-vitro B→PC differentiation scRNAseq atlas as the biological reference. Defines normal transcriptional roadmap.
Tumor Profiling & Classification
Profile patient tumor samples (scRNA-seq or bulk RNA-seq). Project onto normal differentiation trajectory to classify: MBC-like, Pre-PB-like, PB-like, or PC-like.
Subtype-Outcome Correlation
Correlate differentiation-based subtype with clinical outcomes. Characterize which functional disease states are associated with distinct outcome profiles.
CDx Design & Validation
Design qRT-PCR or NGS panel measuring differentiation stage markers. Validate on independent cohorts. Support regulatory strategy with mechanistic rationale for differentiation-based stratification.
Real-World Application
Published WM research used Helomnix in-vitro differentiation reference data to classify WM patients. Question: Are there biologically-defined WM subtypes with different prognosis?
Before
Traditional approach: Cluster WM patients by mutation status (key driver mutations). Found heterogeneity but limited prognostic value and unclear treatment implications.
After
Compared WM tumor cells to Helomnix normal B→PC differentiation atlas. Discovered 2 distinct subtypes: MBC-like WM (early blockade, worse prognosis) vs. PC-like WM (late blockade, better prognosis).
Outcome
Subtypes published in peer-reviewed journal. MBC-like subtype showed distinct pathway dependencies and outcome profiles. Differentiation-based classification provides mechanistic context for translational interpretation.
View Published ResearchValue to Your Organization
CDx Development Speed
Differentiation-based biomarkers identified upfront (during Phase II) vs. extended post-hoc discovery timelines.
Trial Enrichment
Enriching trials for differentiation-matched subtypes supports improved stratification compared to unselected populations.
Regulatory Advantage
Clear biological mechanism (differentiation stage) provides stronger FDA/EMA case than empirical gene signatures, accelerating CDx approval.
Our Methodology
Data Inputs
- Patient tumor samples (bone marrow, lymph node)
- ScRNA-seq or bulk RNA-seq data
- Clinical outcomes (response, survival, relapse)
- Treatment history (drug, dose, duration)
- Disease type and stage
AI/ML Techniques
- Integration with Helomnix in-vitro differentiation reference atlas
- Pseudotime trajectory analysis for tumor classification
- Gene set enrichment for differentiation stage signatures
- Patient clustering by differentiation blockade patterns
- Treatment response correlation analysis
- Multi-cohort validation (training + validation sets)
- QRT-PCR or NGS panel design for clinical deployment
Deliverables
- Differentiation-based patient subtype classification
- Outcome profiles by subtype
- Companion diagnostic gene signature
- QRT-PCR or NGS panel design recommendations
- Validation report with ROC curves (AUC, sensitivity, specificity)
- Mechanistic summary for FDA/EMA submission
- Enrichment design considerations for translational and clinical workflows
Discuss a Translational Application
We welcome discussions about how this approach can support your translational research.