Enabling ‘the magic’ in hematology—eyes on what labs need

Ken Childs (CellaVision): Yes, we have seen that, and people are becoming more comfortable looking at a screen to view what they typically looked at under the microscope. But having access to that information was not easy in the past.

Artificial intelligence networks are trained by looking at millions of different cells. AI brings a consistency to the differential that you can’t get even by asking dozens of technologists to do the same differential, and it applies that standardization upfront. It still requires the operator to review the information and make a proper diagnosis or recommendation. Having better information provided to you by using artificial intelligence has been critical and game-changing in standardizing differentials in hematology.

Jennifer, do you think we are at the beginning of these algorithms being useful to us?
Jennifer Starks (Sysmex): I hope so. I think physicians will be looking to diagnostics for more interpretive information to be able to figure out what’s going on with their patient.

Fernando, can you comment on the autoverification element in hematology reporting and instrumentation?
Dr. Chaves (Siemens Healthineers): It’s a critical aspect. Given the challenges with the shortage in labor and expertise and the need for standardization and consistency in results and to drive down manual differentials, there is no question that autoverification is a key aspect of any successful solution in hematology.

Linda, the importance of refining reference ranges for patient populations is a first step to get to useful autoverification and artificial intelligence. Are you thinking about dealing with the mass of data to help refine reference ranges in hematology?
Linda Garlaus (Beckman Coulter): Yes. We would like to use data mining to gather data fairly easily versus having to do specific studies.

The other component of autoverification is trying to mimic what the laboratory is doing as part of its operating procedure. What’s critical for them? What is their action point? It’s not the same across different laboratories.

Jonathan, will the increasing subspecialization in pathology play into improved autoverification and understanding of reference ranges? Where are you in your practice in terms of special reference ranges for the population you serve?
Dr. Galeotti (UNC School of Medicine): Yes, I hope subspecialization will mean improved autoverification and understanding of reference ranges.

How to appropriately define our reference ranges is something we deal with and think about often. People do it in different ways. A lot of factors go into it. It’s critical to know what is normal and what is abnormal. You can’t do one without the other. It’s important to us in the lab to have good reference ranges we believe in and have evidence to support.

Fernando, are vendors increasingly taking on some of that data mining and refinement as part of their offering?
Dr. Chaves (Siemens Healthineers): We are taking the steps to prepare for that. You need to have solutions that are digitally integrated across networks, with similar interfaces. At the stage where we are in the industry, it is critical for us to set up the platforms that will enable all of this magic to happen.

When it comes to reference ranges, it’s as Dr. Galeotti said—you can only know what is abnormal once you know what is normal, and knowing what is normal changes depending on your population. So we need to create platforms that will enable the AI algorithms, the digital work that is coming. It will be a tremendous value if the background noise of the data can be recognized and eliminated.

It’s a new era in terms of what can happen in hematological data, and we are moving in that direction, as is the rest of the world—all industries are using artificial intelligence now. We in the industry need to keep that in mind so that we create platforms that enable this innovation.

Jennifer, how is Sysmex reacting to the issues we’ve raised today, and what is your focus in research and development and in educating clinicians?
Jennifer Starks (Sysmex): We’ve been focusing recently on software solutions. The history of hematology and the laboratory in general was: We buy an analyzer, it has the software loaded on that operates that analyzer, and then we use the brains of the pathologists and technologists to interpret the data.

We’re starting to see more emphasis on having software that can interpret the data as well as tease out information. The challenge becomes leading the laboratory toward identifying the value to be gained by investing in software solutions that can provide much more help and enhancement to the data they’re currently producing with their hematology analyzers.

Linda, what is your reaction to what you’re hearing, and what is Beckman Coulter doing to address these issues?
Linda Garlaus (Beckman Coulter): Automation solutions are needed for the small and midsize laboratories; our DxA 5000 Fit is one way we are helping support the staffing challenges labs are facing today. In addition, there is a need for more help in clinical decision support, whether it’s technology or putting in complex algorithms that look at the data as well as the patients’ results.

Combining the knowledge of those proprietary markers with the common practices that clinicians go through in their assessment, putting them together to help with clinical decision support, is a big emphasis.

We are also focused on reducing manual steps and are looking at what technologies in the future can help. Instruments on the market today provide accurate results, but they’re using conventional measurements, whether it’s light scatter or fluorescence, to indirectly classify cells through correlations of identified populations. We’re also thinking about what could be a cost-effective technology to address the needs to simplify workflow and provide more valuable information.

Ken, do you see within the CellaVision offering new parameters, new depth you could achieve with image analysis?
Ken Childs (CellaVision): We’re always improving and adding parameters to the AI analysis. We continue to add them as we develop the capability.

Some of the growth we’ll see, I believe, is in using artificial intelligence integrated with the cell counter. There’s more possibility there than just improving our hardware or adding another parameter.

The future may lie in integrated software solutions, because when you have an integrated system and workflow, you solve the labor shortage and knowledge problems. Workflow can mean a lot of things—it’s not only hardware but also integrating the software and artificial intelligence into the process.

Fernando, you have a new offering on the way. What can CAP TODAY readers expect to see coming from Siemens hematology?
Dr. Chaves (Siemens Healthineers): We have the Atellica Hema portfolio, which is already available outside the United States. It is a CE-marked product. It’s designed to meet key customer needs in hematology—ease of use, reliability, scalability, flexibility. Customers love how easy it is to use—few reagents, easy to load and unload reagents, walkaway startup and shutdown, few manual steps—which addresses the shortages of labor and skilled professionals.

In terms of integration, we have Atellica Data Manager, our middleware solution, in which customers can have a hematology solution under the same middleware as chemistry and immunoassay. The same technologist who is operating a chemistry system doesn’t need to learn a new software or become familiar with a new environment to operate the hematology system. All of this is at the instrument level. At the solution level, the clinical content level—that’s where the digital content and artificial intelligence is, and that’s the platform we want to enable so clinical innovations can also become a reality.

Jonathan, can you share with these industry figures what is top of mind for you in terms of your desires in commercial hematology offerings?
Dr. Galeotti (UNC School of Medicine): Cybersecurity issues have been a hot topic in the academic world. The laboratory is a point of access to the EHR and the network, so that has come up in a lot of discussions.

The other is point-of-care testing in hematology. There’s been a push from smaller clinics or specific clinics to get more bedside testing to get results faster for various patient populations.

Jennifer, can you comment on the point-of-care issue from the Sysmex perspective?
Jennifer Starks (Sysmex): Sysmex has a segment of analyzers designed to fit in the point-of-care customer segment. One is the XW-100, which is a CLIA-waived CBC analyzer. It fits easily into physician offices or small clinics that maybe don’t have the ability to have a moderate-complexity analyzer.

With our partnership with Cella­Vision, we have the smaller Cella­Vision DC-1 analyzer, which can bring that level of automation to smaller facilities.

Ken, more sophisticated technologies and automation are popping up in places we’ve not seen before. Is that true for CellaVision’s offering?
Ken Childs (CellaVision): Yes, and we introduced a new system, the DC-1 that Jennifer mentioned, in the United States last year. It can run one slide at a time and it’s perfect for smaller clinics and places where you are closer to the patient and need expert consultation immediately. It is essentially identical to our larger system, but price- and size-wise it is suited to a smaller clinic or laboratory that can’t afford and has no need for a high-volume system.

Jonathan, is there anything you’d like to add in closing?
Dr. Galeotti (UNC School of Medicine): I’m excited that people are thinking forward and trying to use new technology to help physicians and the laboratory staff make decisions and support us in a way that is productive and helpful. I hope we can all work together to make better solutions that help patients and the hospitals that treat them.