The slide is de-identified.
The label is not.
Paige.AI and PathAI want your SVS, NDPI, and MRXS files. The scan is clean. The embedded label image - the one with the printed patient name and barcode - is not. SlideIQ handles all three PHI surfaces before any file reaches an AI vendor or research archive.
Before and after SlideIQ.
Everything between your lab and a Paige.AI or PathAI data-sharing agreement.
Everyone de-identifies the scan.
Nobody de-identifies the label.
Whole-slide images carry patient identity across three distinct surfaces: the TIFF/NDPI metadata, the embedded slide label (often a barcode with printed name), and the linked LIMS case record. Standard export pipelines address at most one. SlideIQ addresses all three.
The SlideIQ pipeline.
Three things that change immediately.
Deployed in your lab environment
We connect SlideIQ to your scanner output and LIMS. Half-day installation. No disruption to the scan-to-report workflow your team already runs.
Three-surface coverage active
Metadata, embedded label images, and LIMS case IDs are all processed from the first slide. Paige.AI or your research partner receives governed data immediately.
HREC-ready export log
Every WSI movement is logged with processing provenance - methodology version, surfaces addressed, audit hash. Ready for HREC submission or institutional review from day one.
What SlideIQ covers.
Three-surface de-identification
SlideIQ is the only system that addresses all three PHI surfaces in a whole-slide image in a single pass: TIFF/NDPI metadata, embedded slide label images (including barcodes), and linked LIMS case records. Each surface uses a different technique - metadata scrub, automated label redaction, and pseudonymous key replacement.
Label image redaction
Automated detection identifies slide labels, barcodes, and printed patient text embedded in the slide image tile.
LIMS case unlinking
Case IDs replaced with pseudonymous references. Linkage maintained for re-identification under controlled conditions.
Policy-gated routing
Each AI vendor or research archive receives only the fields and studies permitted by their data-use agreement.
Immutable audit chain
Every de-identification decision logged. Exportable for HREC, TGA, or institutional review.
Built for the formats your lab actually uses.
SlideIQ integrates with your existing LIMS and scanner workflow. No changes to how your lab operates.
What pathology teams usually ask us.
Paige.AI wants our whole-slide images. What is the actual privacy risk?
The risk is the label, not the tile. The TIFF pixel data - the scan itself - carries no identity. But every WSI has an embedded label image (often a photograph of the physical slide with a printed patient name and barcode), plus TIFF metadata and a LIMS case ID linkage. Paige.AI, PathAI, and similar vendors will receive all of that unless something processes it first. SlideIQ handles all three surfaces.
We already de-identify TIFF metadata before export. Isn't that enough?
Metadata de-identification covers the TIFF tag fields - ImageDescription, DateTime, Artist, and similar. It does not address the embedded label tile (a separate image embedded in the file), the LIMS case ID in the filename or associated metadata, or slide barcode text that may be captured in the label photograph. These are the surfaces that standard export tools miss and that most researchers and AI vendors inadvertently receive.
Does SlideIQ work with our LIMS system?
Yes. SlideIQ integrates with HL7 v2, FHIR R4, and direct database connectors to handle LIMS case delinking. Case IDs are replaced with pseudonymous references. The mapping is retained under controlled conditions so slides can be re-identified if required - for example, for a clinical follow-up after a research finding.
What about the barcode on the physical slide label?
SlideIQ's label detection model is trained specifically on slide label formats - printed barcodes, printed patient names, handwritten annotations, and QR codes embedded in label images. It detects and redacts the label tile within the WSI file. The WSI tile data (the scan itself) is untouched.
How do we satisfy HREC requirements for pathology data sharing?
HREC approval for pathology data sharing typically requires documented evidence of de-identification methodology, an audit trail, and alignment with OAIC APP 11. SlideIQ produces a signed processing manifest for every file - methodology, processing version, surfaces addressed, and audit hash - exportable in the format your institution's research governance team requires.
What is the throughput for a WSI batch through the pipeline?
Processing time depends on file size and label complexity. A typical SVS or NDPI file at 40x runs 2–5 seconds. For large batch exports - moving a cohort of slides to a research archive - SlideIQ queues and processes in the background with no impact on routine lab operations.