Oncology applications
Pancreatic ductal adenocarcinoma (PDAC) · Glioblastoma multiforme (GBM)
By the numbers
hypotheses
candidates
(no existing trial)
trials reproduced
Counts include hypotheses where pancreatic cancer or glioblastoma is either the drug-source domain or the outcome domain. A single hypothesis may bridge into a neurology domain (e.g. PDAC drug → mediator → Alzheimer's).
Featured hypotheses
Selected high-quality candidates surfaced through shared molecular mediators. Each card links to the full evidence chain in the catalog.
Suggests carbamazepine (an Epilepsy drug) binds to KRAS, which is associated with pancreatic ductal adenocarcinoma.
Suggests osimertinib (a Glioblastoma drug) inhibits EGFR, which is associated with pancreatic ductal adenocarcinoma.
Suggests erlotinib (a Pancreatic cancer drug) inhibits EGFR, which is associated with glioblastoma.
Suggests valproic acid (an Epilepsy drug) downregulates PKM2, which is associated with poor overall survival.
Suggests metformin (an Alzheimer drug) inhibits mTOR, which is associated with glioblastoma.
Suggests niclosamide (an ALS drug) inhibits mTOR, which is associated with glioblastoma.
Suggests olanzapine (a Depression drug) binds to EGFR, which is associated with glioblastoma.
Suggests cetuximab (a Pancreatic cancer drug) binds to EGFR, which is associated with glioblastoma aggressiveness.
Suggests osimertinib (a Glioblastoma drug) inhibits EGFR, which is associated with AGR2-dependent oncogenic phenotypes.
Browse all pancreatic cancer hypotheses → · Browse all glioblastoma hypotheses →
Methodology note
These hypotheses were surfaced using the same Swanson ABC reasoning applied across all of Robertium's domains: identifying cases where a drug literature and a disease literature share an intermediate mediator but rarely cite each other. Full methodology is described at /method.
Why this matters
PDAC has had no major treatment breakthroughs in 15 years and a 5-year survival rate below 12%. KRAS mutations drive over 90% of PDAC, yet remained "undruggable" until recently and treatment options remain limited. Glioblastoma has a median survival of 14–18 months despite maximal treatment. Cross-domain literature analysis offers a complementary path to identifying drugs from other therapeutic areas that may target these pathways.
Open data
- Full catalog — the complete cross-domain hypothesis set is at /hypotheses, filterable by drug or outcome domain.
- Archived dataset — the hypothesis catalog is permanently archived on Zenodo: 10.5281/zenodo.20110977.
- License — source code under MIT, extracted data under CC-BY-4.0.
- Contact — if you research pancreatic cancer or glioblastoma and want to discuss specific candidates, reach out to daniel@robertium.com.