A clinical risk score can't see the biology. Can a tumor's metabolism?
Burkitt lymphoma is several biologically distinct diseases sharing one name. Standard prognostic tools score risk from clinical variables alone. M2‑BLAIPI asks whether metabolic pathway activity — inferred from gene expression and anchored to genomic subgroups — adds prognostic signal beyond that clinical baseline.
Two patients can share a risk score and not share an outcome.
The Burkitt Lymphoma International Prognostic Index reads risk from clinical variables, so it can't distinguish tumors that are biologically different underneath. Genomic studies show the disease splits into distinct subgroups — and metabolism is the downstream layer where that biology becomes measurable.
Clinical risk (BL‑IPI)
The established index built from clinical variables. Useful, widely used, and the benchmark every addition has to beat.
Genomic subgroups
Burkitt lymphoma divides into distinct molecular subgroups (DGG‑BL, IC‑BL, Q53‑BL) with different drivers and transcriptomes.
Pathway activity
Metabolic pathway scores inferred from gene expression turn oncogenic biology into quantitative, interpretable features.
Does biology add signal?
Whether metabolic and subgroup features improve risk stratification beyond clinical variables — tested, not assumed.
Three aims, built in sequence
Each aim produces a concrete research artifact that the next one depends on — from pathway biology, to engineered features, to a risk model benchmarked against the clinical standard.
Pathway signatures
Identify which metabolic pathways are reproducibly associated with each genetic subgroup, separating subgroup biology from EBV status.
Feature engineering
Translate the literature into composite metabolic indices linking oncogenic drivers (MYC, EBV, BCR, TP53) to measurable phenotypes.
Risk modeling
Develop and evaluate a multivariate model integrating clinical, subgroup, and metabolic features — benchmarked against BL‑IPI.
A staged plan — and an honest scope
The work runs across four phases. This public site describes the approach and aims; it does not report findings until they have been reviewed and approved for release.
Foundations
Subgroup biology and oncogenic driver–metabolism literature synthesized into a feature specification.
Signatures
Pathway activity compared across subgroups with multiple-testing correction and EBV stratification.
Features
Composite metabolic scores and endemic-context indicators assembled into a modeling-ready matrix.
Model
Multivariate survival modeling with nested cross-validation, benchmarked against the clinical baseline.
No results are published on this site yet. Statements about risk separation and model performance describe the project's hypothesis and plan — not measured outcomes — until reviewed findings are released.