Overview
Cross-system analysis of cancer biology has generated several testable hypotheses about drug response, treatment selection, and cross-cancer connections. These build on decades of oncology research from centers worldwide, synthesizing patterns that may have been overlooked due to specialization.
Foundation
These hypotheses emerge from analysis of existing published research. The underlying biology was established by thousands of cancer researchers. Our contribution is noticing potential connections between their findings.
The Hypotheses
1
Oncogene Hierarchy
Observation: Certain oncogenes appear to function as "master" drivers.
Prediction: Tumors driven by ABL or MET mutations will show exceptional response to targeted inhibitors compared to other oncogene-driven tumors.
Testable: Compare objective response rates across oncogene-defined tumor types treated with matched TKIs.
2
Checkpoint Drug Clustering
Observation: Different checkpoint inhibitors targeting the same pathway show correlated response patterns.
Prediction: Patients who respond to pembrolizumab will likely respond to nivolumab or cemiplimab if switching is needed.
Testable: Correlate response across PD-1 inhibitor switch cohorts.
3
Chemotherapy Signature Matching
Observation: Certain chemotherapy drugs show similar efficacy profiles.
Prediction: Tumors responding to oxaliplatin may show cross-sensitivity to doxorubicin and etoposide.
Testable: Analyze cross-drug sensitivity in cell line panels and patient cohorts.
4
Cross-Cancer Drug Repurposing
Observation: Different cancer types show similar molecular characteristics.
Prediction: Drugs effective in one cancer type may show activity in others with shared characteristics, independent of tissue of origin.
Testable: Basket trials testing drugs across multiple tumor types selected by molecular markers.
5
TKI Generation Effects
Observation: Different generations of TKIs show distinct response patterns.
Prediction: Patients failing first-generation TKIs may respond differentially to second vs third generation based on tumor characteristics.
Testable: Compare outcomes across TKI generations stratified by baseline tumor features.
6
PARP Inhibitor Optimization
Observation: Different PARP inhibitors show varying efficacy and toxicity profiles.
Prediction: Patient characteristics can predict optimal PARP inhibitor selection beyond BRCA status.
Testable: Develop and validate a biomarker panel for PARP inhibitor selection.
7
Immunotherapy-Chemotherapy Synergy
Observation: Certain chemotherapy drugs may enhance immunotherapy response.
Prediction: Specific chemotherapy agents will show synergy with checkpoint inhibitors based on shared biological characteristics.
Testable: Compare outcomes in chemo-IO combinations stratified by chemotherapy agent.
8
Tissue-Agnostic Response Prediction
Observation: Some biomarkers predict response across cancer types.
Prediction: A unified biomarker panel can predict immunotherapy response better than tissue-specific approaches.
Testable: Develop pan-cancer predictor and validate vs tissue-specific models.
Research Priority Matrix
| Hypothesis |
Data Required |
Feasibility |
Impact |
| H1: Oncogene hierarchy |
Basket trial data |
High |
High |
| H2: Checkpoint clustering |
Registry data |
High |
Moderate |
| H3: Chemo matching |
Cell line + clinical |
Moderate |
High |
| H4: Cross-cancer repurposing |
Basket trials |
Moderate |
Very High |
| H5: TKI generations |
Sequential therapy data |
High |
Moderate |
| H6: PARP optimization |
BRCA cohorts |
Moderate |
High |
| H7: IO-chemo synergy |
Combination trials |
Moderate |
High |
| H8: Tissue-agnostic |
Pan-cancer datasets |
Moderate |
Very High |
Potential Impact
Cancer kills 10 million people per year.
If these hypotheses improve treatment selection by 10%:
- 1 million additional lives saved annually
- 5+ million with extended quality survival
- Reduced trial-and-error treatment cycles
Related: Microbiome & Immunotherapy Response
Akkermansia muciniphila and Faecalibacterium abundance predict checkpoint inhibitor response. See our Microbiome & Immunity hypotheses for testable predictions on pre-treatment microbiome profiling and FMT for non-responders.
Collaboration Invitation
We seek research partnerships with:
- NCI/NCTN clinical trial networks
- TCGA/GENIE genomic databases
- Real-world evidence platforms
- Pharmaceutical partners with basket trial data