67% likely to miss · workos:user_01KSZTKKXBYCCV0QQHTN6WP9S8
target CD40 + PD-1 · moa cd40-agonism + checkpoint-inhibition · failure mode benefit-concentrated-in-unselected-subgroup
- premise Many checkpoint non-responders carry a built, stem-like CD8 reservoir stalled at the egress gate in the draining lymph node, awaiting the dendritic-cell licensing signal — the Garrison state. doi:10.5281/zenodo.20479100
- premise CD40 agonism supplies exactly that missing licensing signal, so it is the mechanistically correct lever for Garrison-positive patients. doi:10.5281/zenodo.20479100
- premise Prior CD40-agonist + checkpoint trials (selicrelumab + atezolizumab; sotigalimab + nivolumab) produced real but mixed signals — the pattern of a benefit averaged across selected and unselected patients.
- inference This trial adds agonist anti-CD40 (CDX-1140) to a pembrolizumab + bevacizumab backbone without selecting for the Garrison state, so any benefit is diluted across the non-Garrison majority.
- prediction The ORR primary is not met — no clean significant ORR improvement from adding CD40 in the unselected population; the signal, where measured, would concentrate in Garrison-positive patients.
Evidence: doi:10.5281/zenodo.20479100, doi:10.5281/zenodo.19958113