Suren Avunjian, can you talk about the demands exceeding the resources that labs and your own customers may be able to bring to bear?
Suren Avunjian, cofounder and CEO, LigoLab Information Systems: Laboratories today are under extraordinary strain financially, operationally, and technologically. The rising complexity of testing, workforce shortages, and tightening reimbursement compound the pressure. This environment creates a challenge and an opportunity for LIS vendors like us. The challenge is to do more with less; the opportunity lies in deploying automation and intelligence that multiply the productivity and accuracy of every team member and process across the lab’s life cycle.
At LigoLab, we see the LIS not as a passive system of record, but as a system of action, one that can orchestrate workflows end to end, from accessioning through billing, while eliminating manual, redundant steps. The future belongs to labs that unify their departments—anatomic, molecular, clinical, and billing—on a single, connected platform. Only then can they operate transparently, surface the right insights in real time, and reallocate human expertise to higher-value work.
I don’t believe incremental improvements will close the resource gap. What’s needed is a step change, automation that’s adaptive, rule driven, and deeply integrated into every stage of operations. As consolidation continues, we have an opportunity to help multisite laboratory networks standardize and scale efficiently without sacrificing visibility or flexibility. The labs that embrace this transformation will emerge stronger, with systems that not only record data but also act on it to drive quality, speed, and profitability.
Liron Pantanowitz, what is the state of play as we look into the final months of 2025?
Liron Pantanowitz, MD, PhD, MHA, chair and Maud Menten professor of pathology, University of Pittsburgh Medical Center and University of Pittsburgh: We cannot get our business done or provide the care we need to provide and be compliant without the LIS. But the usability of lab information systems from the end user’s perspective—pathologists, lab personnel, administrators—is still low. We published a study in the Journal of Pathology Informatics in 2017 [https://www.sciencedirect.com/science/article/pii/S2153353922004448] in which the researchers used a system usability score. When they surveyed end users, they found the usability score of LISs was below the benchmark. Orchard was the only exception.
We like interoperability in an LIS. We need to connect to the LIS to get our workflow to support what we do, so integration is key. I hear from digital pathology and AI vendors that the hardest part of integrating AI into a digital pathology workflow is the LIS vendor. LIS vendors will say they have an API [application programming interface] connection, but it doesn’t seem to materialize unless a particular client initiates it. The vendors could work better together, and I don’t see that happening in practice.
We have to automate more in the lab because we don’t have the workforce for manual work and testing has become very complex. In this era of personalized medicine, we have to deliver results that are more complicated than before, not only in molecular pathology but also in the large number of hematopathology and immunohistochemistry panels we run and in microbiology with mass spectrometry. Those results are so complex that the interfaces required now are complex, and the demand often falls to the LIS vendor. We need the LIS vendor to be more flexible to support complex testing and workflows and customized middleware solutions. A spanner in the works is that the cybersecurity clearance required from a vendor is so difficult that it delays getting anything done.
Ulysses Balis, we can’t just kick the demands to vendors and say, “Do something about this,” because they’re part of the same ecosystem as the labs.
Ulysses G. J. Balis, MD, A. James French professor of pathology informatics and director of the Division of Pathology Informatics and of the computational pathology laboratory section, Department of Pathology, Michigan Medicine: That’s right. It’s a multifactorial problem perpetuated by a number of issues. There’s a lack of multiple ecosystems. We’re seeing a regression to a monoculture, at least for large academic medical centers, in the form of Beaker. Not that Beaker hasn’t improved over the years; it has. But when you look at the complexity of molecular and digital pathology and the rate at which they have evolved, it’s faster than any vendor can keep up with—the nuances of user and electronic interface design and data formats and structures, whether DICOM or otherwise. The net result is an ever-present gap between what is possible with a single vendor and what can be done with niche vendors.
We once espoused the idea of best of breed by having multiple vendors, but because of the need for enterprises to consider the primary vendor approach, it’s become increasingly difficult. When you combine that with economies of scale that don’t let you have many verticals, you have an impossible situation in which the enterprise has high expectations, yet your budget is for one or two systems in addition to your LIS. It could be, for example, that Beaker creates a situation such that you simply can’t deliver. When you add that certain vendors hold the construct of the API hostage so you don’t have an orchestration layer, you’re beholden to put the entire workflow through their system, whereas other good vertical work systems exist. For example, for whole slide imaging, there are image management solutions in which sign-out can be a one-and-done process. Yet we can’t do that because one vendor holds the orchestration layer.
The perfect storm is having inadequate funding and vendors that don’t provide the API key. That creates a perpetual challenge: How do you then provide the best service? In industries like banking, commerce, business-to-business, this has been solved. We’ve had computerization in health care for more than 50 years and yet can’t solve these problems. The vendors must recognize that this needs to be a partnership. Until that happens, we’ll have a significant gap between the art of the possible and the current minimum capabilities.
Nick Trentadue, can you comment on what you’ve just heard?
Nick Trentadue, VP, laboratory and diagnostics, Epic: Our approach at Epic might not resonate at every institution in the same way. We will never be able to provide all the software you need to run a laboratory. However, we work with thousands of vendors worldwide and our approach is to follow industry standard APIs, web services, FHIR [Fast Healthcare Interoperability Resources], interfaces, so no matter what vendor you have on the back end, it will work for you.
Digital pathology and AI vendors are a great example. Epic has no product there. We have 17 digital pathology vendors that customers integrate with. For us, it’s a standard, open infrastructure to connect with whomever you want. Institutions can get boxed in if they have to work with a certain enterprise vendor neutral archive or can’t pick the best of breed system if standards aren’t followed and they need to do custom development. We want you to pick the digital pathology or AI vendor that works best for you and use standards-based integrations to run them in harmony.
Primary sign-out outside your LIMS is supported today. Each institution needs to look at what system it wants to bring in and what its capabilities are. As an example, many of our Beaker groups use SoftBank and standard HL7 interfaces. That integration works well; it’s fairly plug and play. There’s no reason other areas in the lab can’t work similarly. Sometimes it’s a matter of having the technical resources on your side to stitch them together.
Matt Folsom, give us your impressions of what you’ve heard and what your principal concerns are as you do your job on the engineering of the Xifin LIS.
Matt Folsom, senior director of software engineering for Xifin LIS: Interoperability between different systems is the biggest challenge. It’s one thing to interface just with Leica. It’s another when a customer has Leica, Ventana, and other platforms and wants them all to work together, while each needs a different workflow.
We want to have interoperability and standard interfaces anyone can use. If a customer has the technical savvy, they can hit our API and automate things themselves. Or they can use our standard way, and we handle the different vendors. If they don’t like that, it can become a challenge because labs want to follow the opportunity, where the money is, and if the LIS vendor is in the way, it can be a showstopper.
We’re looking to help customers find the best LIS solution, whether it’s standard or not. That can mean different customizations for an individual lab, especially when the lab offers testing no other lab does, it’s a new workflow, and a new, homegrown software. We work with them to come up with the best flexible strategy, realizing that implementing a new module in your system can be significant work. Labs today often look to the LIS vendor to be the solution maker for everything. In one sense, it’s true. The LIS is the heart and blood of the lab.
Gui Axus, what is your perspective on what you’ve heard, and where is Abbott in its development of digital health?
Guillaume (Gui) Axus, divisional vice president of digital health, Abbott: At Abbott we start the conversation around our customers’ pain points, many of which revolve around lab staff shortages, growing test volumes, and the complexity of data integration across the system. It is becoming a challenge, especially with the continuing consolidation in the United States, resulting in wider networks and workload distribution. An example is multiple labs within a system using different LISs—it presents many challenges for them to resolve.
Our mission is clear: to be in a global leadership position in lab digital transformation augmenting operational efficiencies for our clients. We see an opportunity in streamlining data flow across every laboratory and making data integration smoother into the LIS and EHR. We worry about workflow integration and adoption when we talk about digital solutions with our customers, so we process map our customers’ workflows around what we call the diagnosis cycle, starting from when the physician sees the patient to the blood draw to lab testing and the release of results into the LIS and EHR. Our aim is to streamline, automate, and digitize all the phases where our solutions can have an impact. We see opportunities for middleware—ours is AMS—to fill capability gaps within the lab. Our goal is to standardize the needs of the lab around QC audits and reporting.
Standards continue to emerge in importance as we develop LIS and EHR solutions because without them, regularizing and making the workflow efficient seems almost impossible. Ul Balis, can you comment on the need for robust standards to enable what everyone wants to accomplish?
Dr. Balis (Michigan Medicine): Standards are extremely useful if properly implemented. Liron [Pantanowitz] and I have had conversations about the need for standards vis-à-vis DICOM, but also for DICOM to be reflexive and responsive to the unique needs of pathology and to educate the pathology community on how to use DICOM effectively and understand its strengths and weaknesses. Part of the failure is with the DICOM community to inculcate what DICOM’s capabilities are in fields outside of radiology, cardiology, and other image-capturing specialties, given that pathology is the most recent to join the digital imaging specialties.
This is an opportunity for a broad dialogue in pathology informatics on what must happen in terms of standards—not just DICOM but other interoperability standards, image formats unique to pathology—so we can become a normal and effective consumer of the larger digital image ecosystem and have the tools and features we need. This is a transitory time for pathology awakening to DICOM being present but not entirely adequate for the needs of pathology. At the same time, pathology needs to recognize we’re not going to change the world. DICOM is not going away. We have to find a way to use it effectively.
Liron Pantanowitz, your response to that?
Dr. Pantanowitz (UPMC): We should adhere to standards. They simplify things and improve interoperability. It can be hard to understand how to implement them, and not everyone is compliant, so I get that people like customization.
One of the hardest things now is where the LIS fits into the digital pathology workflow. I have always been a proponent of an LIS-driven workflow for pathologists to sign out cases, as opposed to a digital pathology system workflow. The LIS is where we go for reporting and ordering and for orders to be compliant. At the same time, if pathologists are going to sign out cases via a digital pathology workflow, then LIS vendors need to support a digital workflow. We now have 13 CPT codes that track the use of digital pathology to demonstrate the value of scanning a slide. We want to be reimbursed for those transactions to provide the return on investment. But we struggle to get these simple tracking codes incorporated into the LIS, and having spoken to other labs, I know they do too. It’s a big deal for the business of lab medicine to be able to appropriately capture a CPT code.
If you look at the history of the introduction of digital pathology systems, there has been difficulty in deciding which digital pathology systems and vendors will try to recreate an AP system and which you want to feed efficiently into a current LIS. Gilbert Hakim, what do you think of this dilemma as this new vendor pool comes into the space?
Gilbert Hakim (SCC Soft Computer): For the integration of multivendor systems, first you need to denormalize the results. We have a data warehouse in which the data is denormalized. We can augment the data warehouse with foreign system information that is congruent with local information and expose it to the network. We can solve some problems by denormalizing the foreign system results received in the warehouse.
AI ultimately will reduce the labor and cost of testing and create accuracy. It is being used in revenue cycle management, digital pathology, autoverifications, and even client support. We’re using AI to ensure clients receive an immediate response to their questions or tickets without even contacting us. We can improve support and turnaround time while requiring less labor from the client, the hospital itself.
We have a middleware called DMI, which can reduce the amount of data when connecting to pipelines. So data reduction can be done with AI as well. Even in clinical pathology we can do detection. You must have machine learning language and AI to be able to detect bacteriuria in blood without doing a microbiology test, for example. AI reduces the utilization of resources and cuts the costs of unnecessary testing.
There are specialized areas within the laboratory, such as toxicology, flow cytometry, biochemistry, and genetics connections, where using traditional HL7 interfaces can hinder productivity, whereas using the native API of the instrument, as we do, is more efficient. Unless you use the correct application, you will not be able to reduce errors and the FTE requirement.
The control of purchasing submodules is not in the hands of lab directors or pathology chairs. These decisions have been switched to IT, and they control the purse. So implementing submodules within the institution to improve productivity and reduce errors is a dilemma for the hospital. But that’s not in the cards for most large institutions. SCC has integrated with most digital pathology vendors in the market today and is agnostic to the type of scanners.