Preparing your lab’s automation infrastructure for the inclusion of digital pathology and AI

As digital pathology and artificial intelligence (AI) continue to transform diagnostic medicine, laboratories find themselves at a timely and critical juncture. These technologies are not simply enhancements, they represent a fundamental shift in how clinical images and data are managed, routed, shared, viewed, interpreted and presented to pathologists. Laboratories that wish to remain at the forefront of innovation must evolve beyond traditional automation and digitally replicating their manual workflows.

Instead, the lab that successfully implements these solutions must carefully evaluate its technical infrastructure to accommodate complex, image-centric workflows, multi-source data integration, and AI-assisted analysis. This evolution requires careful analysis and review of platforms that not only handle high-resolution whole slide images (WSIs) but also enable interoperable communication across diverse systems—such as scanners, LISs, EMRs, EHRs, and archives—using both native file formats and standardized protocols like DICOM.

Successful implementation of digital pathology and AI depends upon a robust informatics ecosystem that supports interoperability and real-time data exchange between your applications. This includes information that will be displayed in your viewer such as WSI’s, structured annotations, snapshots, AI overlays, and results outputs, but it also includes patient-specific case information you traditionally see in an LIS. These tools must be securely embedded within workflows, ensuring compliance with privacy regulations, while remaining scalable for multi-site deployments.

And this transformation is not solely technical. It requires training and cultivating a digitally fluent workforce, with training initiatives that enable pathologists, lab (and scanning) technicians to navigate the components of the process seamlessly and accurately.

Assess current infrastructure preparedness

Before implementing new technologies, it is essential for laboratories to assess their existing systems and infrastructure. This includes evaluating scanner compatibility, data storage capabilities, and network bandwidth to ensure readiness for high-volume imaging, robust case routing and AI integration. Understanding what's already compatible and supporting this initiative can help identify gaps, avoid redundant investments, and lay the foundation for seamless upgrades. A thoughtful audit sets the stage for informed decisions and smoother transitions into digital pathology and automation.

Key considerations include the following:

Scanner selection: Do you already have scanners in place? If so, are the digital solutions you are evaluating compatible with all of them? Do the scanners fit the needs of your organization in terms of volumes, speed, and quality of scanning?  Are they dated or current in terms of the technology? If they aren’t, or you don’t have scanners, these are all items to consider in your evaluation and acquisition of scanners. 

Keep in mind that the better digital platforms will support multiple scanner brands and models, enabling you to mix and match the scanners that best suit your organization’s needs. 

Information management system(s): Is your LIS/EMR/EHR interoperable or compatible with other systems or applications? Is it an older, legacy system that will have limitations or challenges with being able to send and receive real-time data messages through standard methods of integration? Can your digital solution providers help bridge any of these gaps or do they have ways they can work around them in order to provide a seamless workflow for your pathologists and your technologists?

If your information system or “source of truth” for the patient is current and easy to integrate with other solutions, do you have one or multiple you are looking to integrate with?  Are you a multi-site entity or single location with one LIS or one version of that LIS? 

All of these are points to consider when determining: 1) Which digital solution to select, and 2) If there are other infrastructure changes you need to make to ensure that your applications are able to speak to each other effectively.

Network bandwidth: Most robust digital solutions today are web-based, as is the storage for these large images. It is essential to understand what your current network supports and to plan to make modifications prior to implementing a digital platform. Speed and performance are a huge factor in the successful usage and adoption of your platform. 

Most digital solutions are going to be optimized for performance, but internet and network bandwidth are two things that no vendor can account for. There are certainly things like load balancing and viewer performance that are important parts of the development and tuning process of vendors, however, providing the proper bandwidth and connectivity is something that the lab — or your IT department — will need to plan on. 

Storage options: Preplanning is an important step here as well. Knowing whether you are going to archive all of your images or some of your images and for what amount of time is a key consideration. 

Also understanding your options and selecting the one (or two) that are right for you is important. Are you planning to store locally/onsite? In the web? Does your digital solution provider offer both short- and long-term storage? Are you looking to use one of the large storage vendors (Google, AWS, Azure)?  Understanding the price difference between these options, and if your digital solution provider can integrate bi-directionally with them is also a fundamental consideration.

On this note — any decision you make in this category can be easily changed.

Configure for interoperability

A successful digital pathology deployment depends upon integration, real-time data exchange, and true interoperability.

Support for image formats: Enables streamlined viewing, image sharing, and best of breed selection for your organization across platforms. Are you looking to integrate your other ‘ologies’ and have a cross-patient perspective by being able to collocate radiology and pathology images? If so, you also need to ensure that your systems support DICOM and integration with VNRs.

Open APIs and HL7: Enable seamless connectivity with LIS, AI algorithms, other lab applications and modules, and advanced analytics platforms.

Vendor-neutral solutions: Avoid vendor lock-in by selecting digital platforms that offer interoperable viewers, databases, workflows, and reporting.

Choose AI vendors and algorithms that align with your needs

AI algorithms should be carefully considered and evaluated to ensure that you address your specific clinical needs rather than adopt them as generic tools. Successful integration requires a thoughtful alignment by your digital pathology solution and the AI vendor. Those who have formed partnerships have put a great deal of thought and technical evaluation into the integration between the capabilities of their AI outputs and the digital workflow and visual representation to the pathologist. 

Whether it's assisting with tumor detection, streamlining case triage, or supporting specific diagnosis and cancer grading, AI solutions should be targeted to enhance clinical workflows and improve decision-making. By integrating AI deployment directly within your digital platform’s viewer, you are enabling your workflow to be grounded in the realities of patient care and operational demands, ensuring that your laboratory can have meaningful impact, increased efficiency, and better outcomes.

Modernizing your lab’s infrastructure isn’t a single upgrade, it’s a strategic transformation. By focusing on compatibility, clinical utility, and scalable systems, labs can future-proof their operations and unlock the full potential of digital pathology and AI.