• There are no suggestions because the search field is empty.

GLP and AI-assisted image analysis: what to take into account

Good Laboratory Practice (GLP) ensures the quality of non-clinical laboratory studies. Learn how AI-assisted pathology image analysis fits in the GLP workflow.
Written by Emilia Lönnberg

The first part of this article describes Good Laboratory Practice, its benefits, and how to be GLP-compliant. The second part dives deeper into how AI-assisted image analysis fits into the GLP-compliant environment. If you are familiar with GLP as a concept, click here to jump straight to part 2.

What is Good Laboratory Practice?

Good Laboratory Practice is a quality control system that covers the organizational process and conditions of non-clinical laboratory studies assessing chemical and product safety. It aims to uphold data quality and validity by ensuring that studies are planned, performed, monitored, recorded, reported, and archived according to the standards defined in the OECD Principles of GLP and 21 CFR Part 58 in the US. The GLP principles aim to uphold data quality and validity in non-clinical histopathology studies assessing chemical and product safety.

The safety of new drugs and chemicals must be tested to avoid toxicity to humans, animals, and the environment. Thalidomide is a known example from the late 1950s when things went tragically wrong due to the lack of proper testing protocols. As a result of the tragedy, changes were made to the way drugs were marketed, tested, and approved across the world.

Nowadays, to prove that the medicine is safe and not toxic, pharma companies need to provide regulatory bodies with traceable and reliable data on the safety evaluation. The primary benefit of following GLP is the creation of a document trail that provides traceability. This results in increased confidence in the quality and reliability of study data. Another benefit of widely implemented GLP standards is that everyone in the industry speaks the same “language,” as the United States, UK, and EU all have similar GLP requirements, being members of the OECD.

What is required to be GLP compliant?

Test facilities that run GLP studies are monitored by GLP Compliance Monitoring Programmes. Test facilities that work under the GLP environment increasingly collaborate with external providers, such as Software as a Service (SaaS) companies. Even though many SaaS software are GLP compliant, ultimately, it is the test facility's responsibility to ensure that their studies meet the quality standard. 

When a test facility decides to implement new software into their workflow, it defines an acceptable risk level and validates the software products according to that level. Validation ensures that the product meets regulatory, functionality, and safety requirements.

Usually, a service level agreement (SLA) is created between the test facility and the external service provider. Its purpose is to define the responsibilities and expectations of both parties, and it should address many relevant GLP aspects. The external service provider is responsible for the delivery in a way that ensures that the test facility is able to fulfill GLP requirements. It is important to remember that in the end, the test facility is responsible for GLP compliance for the lifecycle of their computerized systems.


AI-assisted image analysis: what to consider in a GLP setting

The preclinical phase of drug discovery involves extensive safety evaluations of the potential therapeutic intervention, often using cell and animal disease models. These studies commonly involve examining numerous histopathological samples, which makes the process time-consuming and labor-intensive. AI can empower pathologists and scientists to make the most of a digitized workflow and fully harness their expertise by automating repetitive tasks and increasing the speed and accuracy of study reviews.

When choosing a vendor for AI-assisted image analysis in a GLP-compliant environment, a couple of things should be considered: 

  1. How does the platform support the GLP workflow
  2. What are the advantages of a cloud-based platform
  3. How to integrate the platform into existing laboratory systems

Let’s cover these one by one. ↓

How the platform supports the GLP workflow

It is important to evaluate whether the platform provider has experience working with customers who operate in a GLP-compliant environment, what kind of support they provide, and if they have designed their product to support the GLP workflow.

To ensure that drugs and medical devices are tested in controlled studies so that the products are safe, developers need to be accountable and results auditable. Ensuring sufficient documentation in a traceable manner is one of the most important but maybe the most unpopular tasks in the laboratory. 

Aiforia® Studies is designed to fulfill the requirements for GLP-compliant studies. Its user-friendly UI guides the user in performing GLP-compliant actions. Studies users can create their own AI models or use Aiforia’s existing AI models, and they can benefit from seamless management, viewing, and sorting of images based on study-defined metadata – all within a GLP-compliant environment. With Aiforia's AI models, quality control of images and scanning and other quality control measures can be integrated into the workflow, boosting efficiency in the laboratory.


What are the advantages of a cloud-based platform

A GLP safety study can consist of up to thousands of histological tissue samples, and managing such studies in the cloud can provide substantial benefits.

The advantages of choosing a cloud-based image analysis platform are: 

  • Easy access to data in a secure environment
  • Extensive traceability through automated audit trail
  • Being able to manage, view, and share study data efficiently
  • Flexible integration capabilities into other systems, e.g. LIMS
  • Consistent image analyses using AI models
  • Storing data in the cloud or exporting to other locations
  • No need to install and maintain on-premise


How to integrate the platform into existing laboratory systems

Viewing images and AI analysis results does not require a high-bandwidth network. There are minimum recommended client requirements (hardware) for training the AI models and when uploading large files into the Aiforia® Platform. However, the most important requirement for successful image analysis is that the laboratory image quality is good enough. 

When laboratories consider adopting digital image analysis tools, one concern often is their integration into existing systems, such as the scanners and laboratory information systems (LIMS) in use. 

The Aiforia® Platform can be seamlessly integrated into the workflow of a pathology laboratory:

  • Images and metadata can be automatically transferred from the scanner/LIMS to the Aiforia® Platform. All major scanners and file formats are supported. 
  • Aiforia® Studies will be used to conduct GLP-compliant image analysis and study evaluation within the platform. 
  • Image analysis results can be automatically transferred to LIMS and/or viewer.



If you want to know more about how the Aiforia® Platform can benefit your laboratory's workflow, book a demo with one of our experts.

Listen to our webinar to learn more about AI-assisted image analysis in preclinical research →