ideasIdea

One certified open-source de-identification pipeline for scans and slides

Build and certify a single free tool that strips names and identifying marks from cancer scans and pathology slides, so every hospital stops writing its own.

Every institution that shares imaging builds or buys its own DICOM de-identification, with inconsistent handling of burned-in text, private tags and slide label images. The Cancer Imaging Archive has curation experience and there are open tools (for example the RSNA anonymizer and CTP), but no certified reference implementation for whole-slide images or radiotherapy objects. The proposal funds a maintained open pipeline with a public test corpus of adversarial cases and an independent certification.

Hypothesis

A certified reference pipeline adopted by more than 50 centres will cut time-to-share for a new imaging cohort from months to weeks and reduce identified leakage incidents to near zero in audits.

Rationale

De-identification uncertainty is one of the most common reasons legal teams block imaging release; a certified standard shifts the risk decision from each hospital to a shared, audited tool.

What would test it

Run the pipeline and three institutional pipelines against a red-team corpus of 5,000 images with planted identifiers; publish leak rates. Then measure adoption and release times among centres that switch.

Maturity

early clinical

Who has to act

engineering

Cost to try

Small (under $1M)

Years to first evidence

Bottlenecks it attacks

Data silos · Records, scans, genomes and outcomes sit in separate systems that cannot talk. Every patient's experience is lost to the next.
AI that is built but not validated or deployed · Thousands of cancer AI models are published; a handful are in clinical use, and fewer have shown they help patients.

Connected

6top

Pages like this

not linked directly; found by shared links

One certified open-source de-identification pipeline for scans and slides

Connected

Pages like this

fronts

technologies

bottlenecks