Research
Tumor-State Explorer
A transparent, rule-based, nine-axis profile — with explainability and a missing-data list.
Tumor-State Explorer turns the structured record into a rule-based profile across nine axes of tumor behavior. It is not ML, not a prediction model, and not a recommender. The rules are visible in source (services/tumor_state.py) and each output has an explainability section.
The nine axes
- Driver signaling
- Immune evasion
- Hypoxia / microenvironment stress
- Metabolic plasticity
- DNA-repair escape
- Death resistance
- Dormancy / MRD risk
- Metastatic behavior
- Treatment resistance
Per-axis outputs
- A score and level (e.g. low / moderate / high).
- Reasons — the inputs that drove the score, by name.
- Missing data — the inputs that would change the score if known.
- Uncertainty — a combined score of how much the answer rests on assumptions.
- Public-dataset references — pointers into the Data Bank when relevant.
Safety
Cautious language only — 'may be associated with', 'requires validation'. The Explorer is for research and hypothesis generation; it is not a clinical decision tool.