OnCo
ideasIdea

Turn the map of immune cells inside a tumour into a standardised test

Whether immune cells are next to cancer cells matters more than how many there are. Turning that spatial picture into a reliable, standardised test would predict response better.

Spatial features such as the distance from cytotoxic T cells to tumour cells, the presence of immune-excluded phenotypes and B-cell aggregates outperform PD-L1 and tumour mutational burden in retrospective series. No spatial assay has been standardised as a companion diagnostic, and platform variability, staining variability and analytical drift are unaddressed. Reference tissue standards and a locked analysis pipeline are the prerequisites.

Hypothesis
A locked spatial signature with defined reference materials predicts checkpoint benefit with a higher c-statistic than PD-L1 immunohistochemistry across at least three tumour types, and is reproducible between laboratories to within a defined tolerance.
Rationale
PD-L1 scoring itself became clinically usable only after harmonisation efforts and defined scoring rules, despite its known weakness. Spatial biology has better biological content and is following the same path, but has not yet done the metrology.
What would test it
A multi-laboratory ring study on shared reference slides to quantify reproducibility, then retrospective validation of a locked signature in banked trial tissue with pre-registered thresholds.
Maturity
early clinical
Who has to act
industry
Cost to try
Medium ($1M to $50M)
Years to first evidence
5
Bottlenecks it attacks

Connected

15top

Pages like this

not linked directly; found by shared links