Loading
No cancer AI would be approved until it has been tested on patients from at least five different hospitals in at least two countries, none of which contributed training data.
Cleared AI devices have frequently been validated on data from one or two sites, often overlapping with development sites. The proposal sets a minimum external validation requirement (at least five independent sites, at least two countries or health systems, no training-site overlap, pre-registered analysis, subgroup reporting) for regulatory clearance of cancer AI, with the sequestered benchmarks as one accepted route.
Shares Sequestered, prospectively collected benchmark datasets that no one can train on, How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals, AI that is built but not validated or deployed.
Shares Sequestered, prospectively collected benchmark datasets that no one can train on, How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals, AI that is built but not validated or deployed.
Shares How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals, AI that is built but not validated or deployed, Regulatory divergence between regions.
Shares How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals, AI that is built but not validated or deployed.
Shares AI that is built but not validated or deployed, Regulatory divergence between regions.
Shares AI that is built but not validated or deployed, Regulatory divergence between regions.
Shares How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals, AI that is built but not validated or deployed.
Shares How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals, AI that is built but not validated or deployed.