Coming Soon · Join the waitlist
Digital Pathology

The slide contains
the diagnosis.
The label contains
the patient.

SlideIQ de-identifies whole-slide images at three surfaces simultaneously - TIFF metadata, embedded label images, and LIMS barcodes - before any WSI file reaches an AI model or research pipeline.

3
PHI surfaces
WSI
Native format
TIFF/SVS
Support
SlideIQ · WSI processing stream
TIFF metadata fields
ImageDescriptionPatient: ████ ████ | DOB: ██/██/████stripped
SoftwareAperio ImageScope 12.3.2retained
DateTime2024:08:14 09:23:11retained
MakeLeica Biosystemsretained
XMP:PatientName██████████stripped
XMP:AccessionNumber████████████stripped
XMP:ScannerIDSCP-████████pseudonymised
Label image surface
MACRO IMAGE · PHI detected in label region
redacting
The problem

Everyone de-identifies the tile.
Nobody de-identifies the label.

Whole-slide imaging carries patient identity in three distinct locations. TIFF metadata headers are the most visible - but the embedded macro image (which includes the physical slide label, often hand-written or barcoded) and LIMS integration fields are routinely missed by standard de-identification pipelines.

Typical WSI de-identification coverage
TIFF header metadata
PatientName, AccessionNumber, XMP fields
covered
Embedded label image (macro)
Physical slide label - often contains handwritten PHI
missed
LIMS barcode in macro region
Lab information system barcode links to patient record
missed
Aperio / Leica vendor tags
Scanner-specific fields not in TIFF spec
missed
Tile-level pixel data
Tissue image - does not contain PHI
covered
What SlideIQ covers

Three surfaces. One pipeline.

SlideIQ processes every location where patient identity can survive standard TIFF export.

01

TIFF & vendor metadata

Full TIFF header de-identification including ImageDescription, XMP namespaces, Aperio and Leica proprietary fields. Configurable retain / remove / pseudonymise per field.

TIFF IFD tagsXMP metadataAperio SVS fieldsHamamatsu NDPILeica SCN
02

Label image redaction

The macro region of a WSI contains a photograph of the physical slide - including the label, which often shows patient name, barcode, or handwritten notes. SlideIQ automatically detects and removes this region with a model trained on pathology slide formats.

Macro image regionLabel detectionBarcode region removalHandwritten text
03

LIMS barcode unlinking

Barcodes embedded in slide labels link back to LIMS patient records. SlideIQ pseudonymises or removes barcode values, replacing them with a consistent study-level token that preserves research cohort integrity without retaining the original reference.

Barcode pseudonymisationLIMS reference removalConsistent study tokenCohort integrity
How it works

The SlideIQ pipeline.

01
WSI Ingest
Whole-slide image received from scanner or LIS via file drop, API, or SFTP
02
Metadata Scrub
TIFF IFD, XMP namespaces, and all vendor-proprietary fields processed per profile
03
Label Detection
Automated detection identifies macro label region and confirms PHI. Region removed.
04
Policy Check
Routing rules validated. Recipient allowlist enforced. Research DUA verified.
05
Clean Output
De-identified WSI forwarded. Cryptographic audit entry written.
Technical specifications

Built for every major scanner vendor.

SlideIQ handles all major WSI formats natively, including proprietary variants. No format conversion required before de-identification.

Supported formats
TIFF / BigTIFF · SVS (Aperio) · NDPI (Hamamatsu) · SCN (Leica) · MRXS (3DHISTECH) · CZI (Zeiss)
Metadata surfaces
TIFF IFD · XMP · ICC profile · Vendor private tags · EXIF
Label detection
Trained detection model for pathology label formats
Throughput
Real-time · tile processing async
Deployment
On-prem agent · LIS / scanner direct integration · S3-compatible storage output
Audit format
Cryptographically-signed audit format · HREC exportable
Compliance
OAIC APP 11, NPAAC standards, ISO 15189, TGA SaMD
Coming Soon

SlideIQ is in development.

Join the waitlist and we will reach out when early access opens. Pathology labs and research institutions go first.