Scanners and ingest
- Scanner output watched and validated on arrival
- Conversion of vendor formats to DICOM whole slide microscopy objects
- Barcode and accession matching against the LIS
Digital Pathology
Digital pathology software development
Your lab or product team needs whole slide images stored, viewed and connected to the LIS, and AI that a pathologist still controls. Agnotic Technologies builds digital pathology software: DICOM WSI storage, browser viewers, lab integration and AI with a pathologist in the loop.
DICOM WSIDICOMwebHL7 v2 and HL7 FHIRHIPAA Security Rule
Trusted by global innovators
What is digital pathology software?
Digital pathology software is the set of systems that scan glass slides into whole slide images, store and manage them in DICOM format, let pathologists view and annotate them, connect the work to the laboratory information system and, increasingly, run AI models with a pathologist in the loop. Agnotic Technologies builds these systems for laboratories and digital health companies.
Whole slide images are large, tiled, multi-resolution files. Most of the engineering is about moving, storing and serving them fast enough for a pathologist to work without waiting, and about fitting the viewer into the case workflow the lab already runs.
Agnotic builds the software; we do not sell scanners or a pathology product. AI tools that could be medical devices are built with your regulatory advisors from the start.
What do we build for digital pathology?
Six capabilities, usually delivered in phases starting with storage and the viewer.
Whole slide image viewer
Browser-based viewing of tiled, multi-resolution slides with pan, zoom, annotations, measurements and side-by-side comparison.
DICOM WSI storage and image management
Ingest from scanners, conversion to DICOM whole slide microscopy objects, tiered storage, retrieval over DICOMweb and lifecycle rules.
LIS and lab workflow integration
Case and specimen data from the laboratory information system over HL7 v2 or HL7 FHIR, worklists per pathologist and results back to the LIS.
AI model integration
Running tissue detection, quantification or triage models on slides, with results shown as overlays the pathologist reviews, accepts or rejects.
Access control and audit
Role-based access, audit logging of every view and change, and encryption, designed around the HIPAA Security Rule safeguards.
Archive, sharing and consults
Long-term archive, de-identified sharing for research and teaching, and secure external consults.
Reference architecture
How does a digital pathology platform fit together?
Described in layers rather than a diagram so it reads on any screen. Each layer is a separate deployable with its own owner.
Storage and image management
- Object storage with hot, warm and archive tiers
- DICOMweb services for query and tile retrieval
- Lifecycle rules by case status and retention policy
Viewer and worklist
- Browser viewer that streams tiles at the current zoom level
- Pathologist worklists driven by case assignment
- Annotations stored as DICOM objects, not screenshots
Integration
- HL7 v2 or HL7 FHIR interfaces to the LIS and EHR
- Results and sign-out written back to the LIS
- Identity through the lab's single sign-on
AI services
- Models run as separate services on slide regions or whole slides
- Outputs stored as overlays with model version and confidence
- Pathologist review required before anything reaches the report
Security and audit
- Role-based access and audit logging of every slide view
- Encryption in transit and at rest
- De-identification pipeline for research and teaching exports
FDA context
Which FDA rules apply to pathology AI tools?
Where a function sits decides the documentation and validation it needs. Your regulatory advisors make the call and we build the evidence. This is not regulatory advice.
01
Image management and viewing
Example: Storing, retrieving and displaying whole slide images for review
Approach: Built as validated infrastructure with audit and performance testing; your advisors confirm whether a device pathway applies to viewing for primary diagnosis
02
Non-device clinical decision support
Example: Software that shows a pathologist the basis for a recommendation they can independently review
Approach: Designed against the non-device criteria in FDA's Clinical Decision Support Software guidance, with the basis of each output visible
03
Device software function
Example: An algorithm that analyzes slide images to detect or quantify disease and drives the diagnosis
Approach: Built under a quality system with design controls, traceability, validation data and the documentation a submission needs, with your regulatory team
Failure points
Why do digital pathology projects stall?
Five problems we see in labs and products, and the design decision that prevents each.
Challenge
Slides load too slowly for a pathologist to work, so the viewer is abandoned.
Agnotic approach
Tiled, multi-resolution storage with DICOMweb retrieval and caching, tested against real slide sizes before launch.
Challenge
Every scanner writes its own format and the archive becomes a mix.
Agnotic approach
Conversion to DICOM whole slide microscopy objects on ingest, so the archive has one format and one retrieval API.
Challenge
The viewer is not connected to the LIS, so cases are matched by hand.
Agnotic approach
Accession and barcode matching on ingest and HL7 v2 or HL7 FHIR integration with the laboratory information system.
Challenge
AI results appear in the report without a pathologist review step.
Agnotic approach
Overlays that a pathologist accepts, edits or rejects, with the decision and the model version logged.
Challenge
Regulatory questions came up after the build.
Agnotic approach
The FDA classification question is asked at scoping, and the documentation is produced during the build, not after.
How we work
How does a digital pathology build run?
Four steps, with pathologists reviewing working software throughout.
1.
Scoping call
We learn the scanners, the LIS, the case volume and whether AI is in scope, and agree a fixed first release.
Scope2.
Workflow and data discovery
We shadow the sign-out workflow, measure real slide sizes and map the LIS messages.
Discover3.
Build with pathologist review
Viewer, storage and integration are built with weekly demos on de-identified slides, reviewed by pathologists.
Build4.
Validate and go live
Performance, integration and, where needed, model validation are documented before the first live case.
ValidateScope and cost
What drives digital pathology software cost?
We do not publish prices. These drivers decide the scope of a first release.
Scanner formats
Each vendor format adds conversion and validation work.
Slide volume and size
Storage tiers, retrieval performance and archive rules scale with volume.
LIS and EHR integration
Each interface adds mapping, testing and go-live work.
AI scope
Model integration, the review workflow and validation evidence add scope beyond the viewer.
Regulatory pathway
A device software function needs design controls and documentation that a viewer does not.
Research and sharing
De-identification and external sharing add a pipeline and consent handling.
Related services
Which related services connect to digital pathology?
Imaging, integration and AI services that share patterns with pathology work.
Medical imaging and PACS software
Radiology imaging platforms that share storage and viewer patterns with pathology.
DICOM and PACS integration
DICOM and DICOMweb interfaces for storing and retrieving images.
Laboratory integration
HL7 v2 and HL7 FHIR interfaces to laboratory information systems.
AI diagnostic imaging
AI models for imaging with a clinician in the loop.
Who it is for
Who builds digital pathology software with us?
The first release looks different for each.
Pathology laboratories
Moving from glass to digital sign-out with a viewer connected to the LIS.
Digital health and AI companies
A platform for a pathology AI product, with review workflow and validation evidence.
Academic and research groups
Archives, de-identified sharing and annotation tools for research and teaching.
Health systems
Pathology imaging alongside radiology PACS, under one identity and audit model.
Facts and sources
Which standards and rules shape digital pathology software?
Short facts we design around, each linked to its primary source.
DICOM is the international standard for medical images and related information, maintained by NEMA and also published as ISO 12052.
Source: DICOM Standard (NEMA)The DICOM VL Whole Slide Microscopy Image IOD (DICOM PS3.3, section A.32.8, added by Supplement 145) specifies a multi-frame visible light whole slide microscopy image divided into a set of tiles.
Source: DICOM PS3.3, A.32.8 VL Whole Slide Microscopy Image IODFDA's guidance on device software functions focuses oversight on software that meets the definition of a medical device and could put patient safety at risk if it does not work as intended.
Source: FDA, Policy for Device Software Functions and Mobile Medical Applications (2022)FDA's Clinical Decision Support Software guidance explains which decision support functions meet the non-device criteria in section 520(o)(1)(E) of the FD&C Act and which remain device software functions.
Source: FDA guidance, Clinical Decision Support SoftwareThe HIPAA Security Rule requires administrative, physical and technical safeguards for electronic protected health information.
Source: 45 CFR Part 164 Subpart C (eCFR)
Compliance-First Healthcare App Development Services Backed by Global Standards
01
HIPAA
Health Insurance Portability and Accountability Act
Protect PHI with privacy-first architecture, encrypted storage and transmission, strict access controls, and traceable audit logs.
02
GDPR
General Data Protection Regulation
Implement lawful consent flows, data minimization, retention controls, and secure processing for sensitive health data.
03
FHIR
Fast Healthcare Interoperability Resources
Enable standardized health data exchange across apps, care teams, and systems through standards-based FHIR APIs.
04
HL7
Health Level Seven International
Support enterprise-grade interoperability with HL7-based integrations for records, events, and clinical messaging workflows.
05
HITRUST
Health Information Trust Alliance
Align security programs to healthcare-specific control and risk management practices trusted by providers and ecosystem partners.
06
HITECH
Health Information Technology for Economic and Clinical Health Act
Design with breach notification readiness, digital record safeguards, and operational controls that support regulated care programs.
07
SaMD
FDA Software as a Medical Device
Plan software quality, traceability, and documentation pathways for products that may require SaMD review and submission.
08
MDR (EU)
Medical Device Regulation (European Union)
Prepare EU market-ready processes for risk classification, evidence tracking, and lifecycle governance under MDR expectations.
09
SAMHSA
Substance Abuse and Mental Health Services Administration
Apply confidentiality controls and consent-aware sharing models for behavioral and mental health data experiences.
Frequently Asked Questions
It depends on the scanners, the LIS integration and whether AI is in scope. We agree a fixed first release on the scoping call and plan the schedule from that scope.
Yes. We convert scanner output to DICOM whole slide microscopy objects on ingest and connect case and result data to the laboratory information system over HL7 v2 or HL7 FHIR, tested against the vendor's sandbox before go-live.
Yes. A common first release is a viewer and DICOM storage for one scanner and one workflow, connected to the LIS, with AI added in a later phase.
Some functions are and some are not. FDA's guidance on device software functions and on clinical decision support draws the line; your regulatory advisors decide where your product sits, and we build the documentation and validation that pathway needs. We do not give regulatory advice.
We integrate models and, where you want, train them with your data science team. The product decision is the review workflow: every model output is shown as an overlay a pathologist reviews before it reaches the report, with the model version logged.
DICOM for whole slide images and annotations, DICOMweb for retrieval, HL7 v2 and HL7 FHIR for LIS and EHR integration, and the HIPAA Security Rule safeguards for access, audit and encryption.
Yes. The viewer streams tiles at the current zoom level over DICOMweb, so a standard browser on a lab workstation can open large slides without installing software. Display requirements for diagnostic use are agreed with the lab.
Engineers use AI tools for drafts, tests and reviews, and every change still goes through human code review, automated tests and a release checklist. For the product itself, no model output reaches a report without a pathologist's review.