Seamless automation: within reach for AP?

Anne Paxton

August 2014—A familiar optical illusion uses a drawing of a vase that makes your eyes play tricks. First you see the vase, then two faces gazing at each other, then again, the vase…two faces…ad infinitum. It’s a concept that comes to mind when thinking about “tracking” in the anatomic pathology laboratory. Does it refer to a physical track—a conveyor belt to automatically transport and sort specimens—or to a system for “tracking”—that is, electronically keeping tabs on specimens?

As the AP laboratory steadily closes the automation gap with the clinical laboratory, it’s not clear whether machine-age automation or digital-age information technology is more important. But AP laboratory directors and others interviewed by CAP TODAY suggest that IT is increasingly an essential part of freeing the AP lab from reliance on manual workflow processes.

Most feel that while AP lab automation is catching up with automation in the clinical lab, serious differences keep full AP automation just out of reach. At Sonora Quest Laboratories, in Phoenix, Ariz., for example, bar coding and electronic tracking by Ventana Roche Diagnostics have been in place in the histology laboratory for a little more than a year and half, says William DeSalvo, BS, HTL(ASCP), system production manager for anatomic pathology. But he still considers the laboratory semiautomated.

Unlike in the clinical lab, where specimens remain in the same tube throughout the preanalytical process, “our specimens get transformed multiple times,” DeSalvo explains. “A specimen arrives in a cup with fixative, is moved, after being grossed, to a processing cassette or cassettes, and is manually loaded into a tissue processor. From there it’s taken to an embedding center and then to the microtome, where the tissue is cut and sections placed on a slide, then stained and coverslipped. So I can’t take a cassette and put it on an instrument at the beginning and have it come out as a slide at the end.”

Sonora Quest, which handles more than 150,000 cases per year, has two workflows to handle specimens. The pathologist or pathology assistant performing the gross dissection decides whether a specimen will undergo rapid or conventional processing. On the laboratory’s Sakura Xpress continuous-load rapid tissue processor, one to 40 cassettes can be added every 20 minutes. Processing takes one or two hours. The conventional batch processors can take up to 200 cassettes at a time but may require four to 12 hours to complete, depending on tissue types.

The system is efficient, but no tracking of specimens takes place on the tissue processors. The laboratory manually transports the samples from the gross dissection bench because Ventana does not have a tracking module for this process,” DeSalvo says.

As a result, it’s difficult to say that when you purchase any existing automation options for histology, it will reduce the number of units working through your lab. “We have some things that release us from manual tasks,” says DeSalvo, who is past president of the National Society for Histotechnology, “but typically there is a large volume of cassettes and slides that must be manually produced and worked through a multi-task system.”

Sonora Quest is a multi-vendor laboratory—it uses special staining instruments from Dako—but DeSalvo does not feel there are a lot of options for middleware that will connect all of the instrumentation. “There are a couple of middleware systems out there now, but they can be costly because you need to purchase a lot of hardware and software to work with them.”

Based on his work as a consultant to labs that are considering automation, DeSalvo believes the trend in histology across the United States, to reduce FTEs, is to replace smaller histology labs with a core histology laboratory in an integrated system. In fact, that is the model his lab has followed for the past decade. “Our majority partner, Banner Health, has acquired additional hospitals, then we bring in that work and/or integrate their small histology lab into our core lab.” Even though the laboratory’s outreach business also continues to grow, this acquisition pattern has kept the outreach/hospital work ratio at 50/50.

In the process, Sonora Quest has gone from 60-plus FTEs several years ago to a much higher volume and a broader test menu with 50-plus FTEs, relying on attrition to reduce staff. “The initial response to automation is that we’ll be paying for it through FTE reduction, which is not always the case. It’s always difficult for the histology world to look at the bigger picture, but there are other soft costs that have to be considered: quality, risk mitigation, overall process improvement. And you cannot always fund these by just bringing in automation and reducing staff.”

More systems are needed to link processing and embedding with robotics, he says. “We should be able to place a cassette on a tissue processor and have it moved to an automated tissue-embedding instrument and come out ready for cutting or microtomy. The Sakura Tissue-Tek AutoTec does help move technologists to other tasks. But the best improvement would be having a way to connect some of the semiautomated instruments. The difficulty is that because of the variation in quality of the specimen tissue, the process requires manual manipulation.”

The other area where DeSalvo would like to see more efficient processing is with a fully automated microtomy instrument. “That would be where I put the paraffin block in, and through manipulation of the software, it cuts that block to produce quality slides in high volume.” One or both of those process improvements should be available in the next five to 10 years, he predicts.

The histotechnology field is increasingly heading toward cancer diagnostics, DeSalvo says. “With the increased need to identify as early as possible any abnormalities, we’re going to see more and more diagnostic-type biopsies and fewer resection-type tissue samples through the lab. And the small diagnostic biopsies are the ones that require the most additional testing, such as special staining, IHC, FISH probes, and next-generation sequencing.”

So he doesn’t foresee a slowdown in histology. “We’re going to be busier with the samples we work with. That’s why we need the newest technology in the semiautomated areas, and we need true automation to allow us to continue expanding our use of technologies in the histology lab to provide pathologists with a full picture of information.”

Digital pathology, too, promises to make the overall AP laboratory process higher quality and more efficient, but it’s going to require a period of adjustment. “It’s another intrusive technology that will require a lot of hands-on work by the trained histotechnologists,” DeSalvo says.

His laboratory is preparing to move to new scanners for breast digital pathology. “Part of our problem is the histology lab is not completely connected. We’ve always been ‘connected’ through our surgical pathology report. But we are now wired and connected digitally, and it’s a big stumbling block for a lot of histology labs to talk about cabling, bandwidth, and servers. Those are absolute necessities for you to work in the digital age.”

When the AP laboratory of David Smith, MD, started looking for a bar-code solution in 2011, it wasn’t planning for more automation than that. But the laboratory soon found itself stepping into a much fuller automation package.

“We were probably on our third-generation automated immunostainer from Ventana Roche, and they had a bar-coding solution called Vantage and offered us a total, or turnkey, software solution. So we added two Symphony automatic H&E stainers first,” says Dr. Smith, president of the pathology group and medical director of the laboratory at Riverside Regional Medical Center, Newport News, Va.

These replaced two “dip-and-dunk” stainers and a coverslipping instrument that histotechnologists had spent an inordinate amount of time adjusting because it left air bubble artifacts under the coverslip. “The Symphony works with a whole different process,” he explains. “The histotechnologist cuts the paraffin block, puts a section in the slide, puts them into a specially designed tray, then pushes one button on the instrument, and on the other end they get a stained, coverslipped slide.”

The laboratory, which was already a customer of Roche and Cerner, then implemented simultaneously the Cerner CoPath Plus advanced bar-coding and tracking solution and Ventana workflow solution. “We decided to purchase both the Vantage and the CoPath tracking solutions because we wanted bidirectionality of the bar-coding solution.”

The laboratory also added a digital pathology solution from Ventana. Verification took several months, but since the system went live about a year and half ago, Dr. Smith says, it has saved the histotechnologists all kinds of time.
“Having a total solution of the software really ties it all together very nicely,” he adds. “Before, if I wanted to add extra IHC stains to the platform, I would order them and my CoPath system orders would print out, but you would have to retype the information in the computer so proper labels would print; then the case could run. Since we implemented the software solution, orders are automatically received into the special staining and don’t have to be retyped. So it has speeded up the process considerably.”

The laboratory, which now handles about 24,000 surgical pathology specimens a year, has seen its AP volume rise 44 percent in the past two years, due to a growing clinical volume and two hospitals having been added to the health system, and it has not had to add staff. “We credit the automation with our significant growth,” Dr. Smith says.

He points to two main advantages of the increased automation, financed through reagent rental with the vendors. “No. 1 is the assurance that the bar-code solution provides positive patient identification throughout each of the steps in the lab and patient ID errors are driven out. Second, it’s significantly increased the productivity of our histotechnologists in the lab.”

The new system has also eliminated duplicative data entry and allowed secretaries, through the tracking solution, to give clinicians requesting a status report accurate information on the spot. As a pathologist, Dr. Smith says, “I can sit at my computer, order my special stains, and have the assurance they are correct. And I can see how many minutes I have left before the stains are complete.”

It’s a replay, in some ways, of the improvements that took place when the system’s clinical laboratory added automation and robotics, he says. “That was quite a few years ago, but at this point we’ve caught up with the clinical lab.”
Next to be added at Riverside are a whole-slide scanner called I-Scan Coreo and image-management software called Virtuoso, both from Ventana. Software integration is the key technological advance that has brought AP automation into its own, Dr. Smith says.

If you are implementing multiple solutions at the same time, he adds, that’s a serious and significant management decision. “It can be a strain on the IT department and your staff. You have to decide how to phase in implementation of the various solutions.”

Dr. Smith’s advice for any AP laboratory considering stepping up its automation: “Going automated is going to require attaining both hardware and software solutions. If you’re going to have multiple vendors or one vendor, think about the end product and have a serious planning session. Will there be an overarching solution? Or will you have to develop homegrown interfaces to tie everything together?”

USC Keck Medical Center, in Los Angeles, uses the Leica Cerebro tracking system for its outreach services. Cerebro tracks and verifies the identity of every specimen at each point of tissue transfer within the lab. “It works fantastically,” says John Vallone, MD, director of informatics and digital pathology. “We use the tracking functionality of Cerebro in conjunction with a cloud-based AP LIS called PathCentral.” The medical center chose Leica because the company agreed to develop a bidirectional interface with the outreach LIS, he says.

At every point from accessioning to archiving, the patient specimen is validated. The system allows for print-on-demand services, so cassettes and slides are specific to the bar-coded parent container and are generated only as needed—that is, when tissue is transferred from container to cassette or at the microtome when tissue is transferred from block to slide.

Dr. Vallone’s laboratory is now planning to track specimens from their points of origin—in endoscopy suites, surgical rooms, and clinics, for example. “We’ve relooked at how pathology functions within the hospital and we’re incorporating processes that used to occur when specimens hit our gross room into the general hospital workflow,” he says. “We don’t want to know the specimen exists when it hits our desk. We want to know it exists when it is removed from the patient.”

By replacing paper requisitions with electronic requisitions, nurses will be able to generate a tissue request like they would any other order, linking the tissue to the patient’s EHR in real time. Written errors will be prevented and specimen standardization supported, Dr. Vallone says, by allowing nurses to choose from  predefined back-end dictionaries to populate specimen name, specimen location, and other fields.

“The new process will improve efficiency by alleviating the need for nursing to notate dates, procedure type, clamp times for breast cases, and formalin times for immunohistochemical processing because that information is populated as the specimen is acquired,” Dr. Vallone says. Accountability for patient information will reside with the person responsible for generating the information. “And its fidelity is enhanced via programming services that can be automated in the EHR and AP LIS,” he adds.

Once the electronic requisition is complete, a bar-coded label that includes patient demographics and specimen information is generated and placed on the container. “Before the specimen reaches pathology, we will know the specimen exists, when it was generated, where it resides, how many specimens are present, and 95 percent of the accessioning process has already been completed,” Dr. Vallone says. When the specimen arrives in the gross room, its bar code will be scanned, which automates the time received and pulls the information from the electronic requisition within the EHR into the AP LIS. A lab technician will verify the information in the AP LIS, and “the accessioning process is reduced from two to three minutes per case to five to 10 seconds.”

The Department of Pathology will know in real time the number of specimens coming from each location. “This makes it possible for us to plan our personnel for increased specimen loads and makes us immediately aware of specimens that haven’t arrived in pathology,” Dr. Vallone says. The process will continue beyond accessioning, “to track all specimen containers to gross benches where cassettes are printed on demand, into and through histology processing, and ultimately to the pathologist to whom cases are assigned.” Once complete, they will be scanned into archives.

Denise Bland-Piontek, CTBS(AATB)HTL(ASCP)QIHC, senior technical director for histopathology in the pathology service at Massachusetts General Hospital, is a believer in health information technology driving automation.

Her pathology service has experienced significant budget savings by insisting on an open bar-code system, instead of being locked into one vendor. Clinical pathology has long had open bar coding, while some AP vendors initially offered only proprietary systems, she says. “That limited some laboratories’ freedom to think about their ability to improve workflow,” she adds, noting that this is not a common pattern now.

After putting out a request for proposal to vendors of immunohistochemistry automated instruments, MGH chose Leica Biosystems’ BOND-III IHC and in situ hybridization stainers to process the roughly 90,000 specimens it receives per year. “That was a very large RFP,” says Piontek, “and quality was our No. 1 requirement, with turnaround time second, but the LIS interface was third: Do you have a bar code we can interface with?”

This element of workflow was crucial, she says. “We label our slides once with the LIS label, and there’s no double labeling, which can cause major patient ID issues. The order goes right across the interface to the Leica instruments from our LIS without using middleware. The pathologists see every move the lab makes on a case real time in the LIS, including when the slides were put on the stainers.”

As a result of its experience with the BOND-III and Leica’s service, quality, and interface, “we’re actually pushing for other Leica platforms,” Piontek says, though she notes that no company has a big enough portfolio for the range of testing MGH runs, and she expects that multiple vendors will continue to be the norm.

“When I arrived five years ago, we had just entered a 10-year contract with Sunquest CoPath to develop the LIS, including the tracking system, so we’re halfway through that, and it’s been terrific.” MGH will move, over the next 18 months, into the next phase of its Epic EHR system, which will include laboratory testing. The AP laboratory will also expand digital pathology. “We’re in a 10-year agreement with Phillips to design the best digital pathology system available,” Piontek says.

For any AP lab considering automation, cost can be a barrier, she agrees. “It’s quite an undertaking to bring in a new platform, and you need to make sure you are leaning on the vendor as much as possible to do the tasks they are promising for you—the vendor should act like a partner.” MGH works almost completely on a capital purchase basis. “Most hospitals have found that what they’re actually paying in reagent rental agreements is astronomical,” Piontek says, “and they really should have budgeted the expense up front.”

But she promises the current plans won’t be the last word. “We do see automation increasing drastically over the next five years because it’s a patient safety issue. Just what’s been done in the last few years has been amazing. We’re finally starting to catch up to clinical pathology, and my mind almost runs wild with what can be possible 10 years down the road.”

She predicts that for some subspecialties or disease processes, in vivo microscopy and digital gross will significantly change workflow in surgical pathology. “Grossing could occur in completely different ways that may not involve human interaction. Grossing, processing, embedding, and cutting may become a one-step rather than a four-step process through automation. We finally see automated fixation stations and decalcification platforms on the market; you can see the gaps being filled,” says Piontek. “Staining is already mostly automated with attached coverslipping; you can see the line process forming.” She has seen some of the automated microtomes on the market and believes they do a decent job, though they have a ways to go before they can replace histotechnologists.

AP automation at Indiana University Health Pathology Laboratory has also advanced significantly in recent years, as the laboratory has acquired new IHC platforms and pretreatment modules, and especially with the Dako Link software solution and True Positive ID AP tracking system that came online more than three years ago, says Eyas M. Hattab, MD, medical director of the immunohistochemistry laboratory at the Indianapolis facility.

However, AP automation in general is “barely scratching the surface” when compared with the level clinical pathology has achieved, Dr. Hattab says. “Tissue testing presents unique challenges, partly because we need to preserve the architecture of the tissue, in contrast with the testing of fluids and blood products that clinical pathology labs perform. So it’s not really fair to compare the two.”

Dr. Hattab’s laboratory is part of IU Health, a large academic medical center spread across a large geographic area and one that has acquired many partners over the past 10 years. So the IUH Pathology Laboratory has been expanding and increasing its volume. “In the next three months, a couple of our partners will close down their IHC labs for cost-cutting purposes and will be transferring their work to us, both IHC and histology. And we do have the infrastructure set up to handle that.”

As far as automation is concerned, IHC is probably head and shoulders above all other laboratories in AP, he says. Tissue embedding, sectioning, and microtoming are still largely done manually and lag far behind IHC automation. “A few years ago, most special stains were done by hand, but nowadays essentially all our special stains are automated. Our H&Es have been automated for many years.”

The IHC laboratory, which handles about 100,000 specimens per year, has two components, the clinical service and research, and both are equally automated, he says.

In the early 1990s, “we decided that in order for us to stay competitive and provide quality patient service in a timely manner, we needed to automate as many of our procedures as possible. So right now we have a large surgical volume—about 85,000 specimens a year—and there’s absolutely no way we could do this without automation.”

Because the laboratory has been growing, it has added staff over the past five years, Dr. Hattab says. “This may seem counter to the idea of automation, but we’ve become more efficient, our arsenal of IHC tests in the lab and other services we provide have expanded, and our volume has increased by 30 percent.”
Partnering with Dako, an Agilent Technologies company, as a beta testing site for the past several years, his laboratory has been able to experiment with platforms and other AP solutions not available to all. Over the past several weeks the laboratory has been transitioning from the Dako Autostainer Link system to Dako Omnis.

The laboratory opted to switch to Omnis because it offers better quality IHC staining and would provide the laboratory with a better workflow. But the upgrading process can be stressful, Dr. Hattab says. “It’s a labor-intensive installation process that is potentially disruptive to our routine workflow, and that’s inevitable when you are changing the way you’re doing things. Repeat testing, for example, is something that spikes during a new installation process, until our IHC protocols have been optimized to the new standard.”

Switching platforms is also a costly endeavor that often requires large initial capital. “Answering the cost question, ‘How will it impact our bottom line, especially in this new era of lower reimbursement rates?’ is one of the most challenging issues,” Dr. Hattab says. In the past, his institution has worked on a reagent rental basis, but with the Omnis, it has moved to capital purchase because it opted for a different business model. That’s meant increased expenditures.
When should an AP laboratory be considered “automated”? “Would I look at a lab without an integrated workflow solution and tracking system and say this is a state-of-the-art automated lab? No, I would not,” Dr. Hattab says. “I would say an AP lab is automated if it has a fully automated IHC lab, if its special stains are automated, and if it has an integrated workflow solution that enables specimen tracking throughout the AP lab and manages the flow of lab information from one AP instrument to another. I believe this is essential nowadays to any AP lab.”

In his view, moreover, AP labs considering automation can no longer rely on a cost analysis only. “If you are looking for this to be cost-neutral, then you may never move toward automation. But you have to keep in mind the additional benefits that automation provides: better quality control, more reproducible testing environments that don’t vary from one technician to another, fewer chances for error, and, first and foremost, improved patient care.” Nevertheless, buy-in from the institution’s pathologists and technologists is essential. “This is inevit-ably going to be a very laborious and involved process that will test everyone’s patience.”

Dr. Hattab is confident that progress in automation will continue. “Automation in AP has achieved major leaps in the last decade, and I believe we will continue to achieve technological advances that will move us closer to our goal.”

[hr]

Anne Paxton is a writer in Seattle. A product guide for anatomic pathology automation will be published in February 2015.