Aug 22 / J. Greenlee

Automation in Histology: Friend or Foe

Artificial intelligence, or AI, and other forms of automation have been headlining many articles and news stories these days.  Everyone is abuzz with not only what can be done today through automation, but also what the future will bring.  In this blog post we are going to examine automation’s role in the world of histology in both the past and present as well as what it may mean for the future.  Of course, automation is not just about robots whizzing around and looking cool, so as we look at automation in histology, it is important to discuss what value it brings in terms of criteria such as quality, safety, and efficiency.

Compared to other laboratory disciplines, it can be argued that most histology laboratories are quite behind the times in terms of automation.  If we were to visit clinical chemistry laboratories around the country, what we would typically see is a whole host of automated analyzers, and in many places even tubes traveling on automated tracks onto or between instruments.  Even microbiology, who used to be histology’s fellow brother-in-arms when it came to manual processes, has been largely transitioning to technology like automated plating systems.  Automation has become common in many of these laboratories due essentially to necessity.  With huge workload volumes, shortened expected turnaround times, and the lack of qualified staff, laboratories running thousands of blood tubes per day and millions of tests per year while patients are waiting for their results needed to adapt.  Automation has been crucial to meet the workloads and goals for these laboratories through increasing efficiency while at the same time reducing the risk to quality and safety.

Did those challenges clinical chemistry laboratories face sound all too familiar?  They likely did because histology laboratories have also been struggling with those same issues.  In contrast to what we might see in a common clinical chemistry laboratory, though, if we were to visit most histology laboratories today, even in some of the most renowned cancer centers in the country, what we would most often see are significant manual processes. Automation is not completely foreign to histology, though.  Most laboratories are using automation for certain parts of the process like tissue processing, staining, and coverslipping, but often in some of those cases, the instruments being used may be decades old.  As some of you are reading this, the laboratory where you work very well could be processing patient samples routinely on a tissue processor that is 10 to maybe even 30+ years old.  As histology professionals, if we were to explain to the general public that the potentially life-altering diagnosis of their irretrievable biopsy relied on the flawless operation of a machine that may have been in operation for almost as long or, in some cases, longer than some of us have been alive, they would probably have a difficult time believing us.  Yet, that is the reality. 

The same goes for some of the manual histology processes.  Imagine what the general public would think if they knew what the traditional embedding process entailed for their tissue.  What stands between receiving their diagnosis and possibly not is an individual sitting at an embedding station manually manipulating each piece of their tissue with a pair of forceps, sometimes with the help of a magnifying glass.  Not only that, but that individual must properly orient their tiny biopsy possibly on edge and hold it there as paraffin wax cools just enough to hopefully prevent the biopsy from falling over or floating up.  Also, if that individual were to get distracted, lose concentration, or maybe even just squeeze their forceps just a bit too tight, that patient’s biopsy might spring up in the air and possibly land on the floor where hopefully that individual’s coworkers are able to hunt it down while the embedder freezes in place and hopes for the best.  Once again, if the general public knew, they would probably be appalled.  Yet, that is the everyday reality for most laboratories.     

So, why is there a significant discrepancy between some laboratory disciplines and histology when it comes to the transition to automation?  There is no one single answer, but rather it is a combination of many reasons.  Some of it has to do with volume.  Laboratories running several thousand blood tubes per day were forced to scale up quicker than histology laboratories that may have a couple of hundred cases per day.  Histology laboratories have not had the same pressures to scale up quickly, but that has been changing.

Another reason is the significant difference in process transitions making histology a bit more of a challenge in some ways to automate.  A blood tube received in the laboratory may not really change form before being analyzed.  For example, serum can be analyzed straight from a serum separator tube.  In contrast, histology is a multi-stage process where the input and output of the product changes multiple times throughout.  In histology, a submitted sample may go from a formalin jar to a cassette then to a block then to a slide where it is finally stained and coverslipped before going on to the pathologist.  Transitioning all these processes that have been traditionally done manually to a fully automated workflow is clearly more challenging. 

The expected turnaround time typically for the processes is also a factor.  From the time the patient is drawn, a STAT CBC result may be expected to be delivered within 30 minutes to an hour.  For circumstances like these, the fast turnaround time expectation on top of the volumes has made automation necessary.  For most tissues for histology, the formalin hasn’t even fully penetrated the tissue in that amount of time let alone be fixed.  Many laboratories are still using conventional tissue processing protocols that are 8+ hours long or done overnight.  Because of this, it used to be common for the turnaround time expectation for histology laboratories to be measured in days, not hours.  Some of you may have caught that I said “used to” in the previous sentence.  That is because turnaround time expectations for histology laboratories have shortened significantly in more recent times.  Turning around patient results quickly is beneficial not only for the patient, but also for the business-side of the laboratory.    

Clearly there are multiple reasons why automation has not been so automatic in histology laboratories, but as outlined in the previous paragraphs, the dynamics of the histology world have changed, and the excuse of “but that is how we’ve always done things” is often no longer meeting expectations.  Workloads are growing, not only in the number of cases but also in the tasks and responsibilities of histology staff.  New regulations and requirements often come with new tasks for staff, and with the growing number of IHC, molecular, and other tests along with the adoption of digital imaging, histology staff are being asked to do more and more.  Digital imaging opens a world of new possibilities for the industry, but what is often glossed over is that it requires new tasks and skills for the laboratory to take on. 

Growing with the workloads and adapting would be all well and good if not for the case that histology laboratories around the country, even globally, are experiencing ongoing and critical shortages of qualified staff.  This is in no way a new phenomenon either.  You can go back more than 20 years and look at the data to see that the more things change, in the case of staffing shortages at least, the more they’ve stayed the same.  How many times have you heard or maybe even said yourself, “Well, if we could only get fully staffed…”  The problem with this mantra of hope is that it is ignoring the reality of the situation.  First, for many laboratories, their approved headcount is already less than what they likely need, so even if they were to get fully staffed, they are still behind the eight ball.  Second, no sooner does a laboratory get fully staffed, it is only a matter of time before someone retires or leaves.  People will come, and they will also go.  It is inevitable.  Even when a new person is hired, there is a learning curve to get up to speed, and the less experienced the person is and the more training is required, the more time and resources in the laboratory are spent to get them up to speed.  Finally, there is the reality that there simply are not enough qualified staff to fill needed positions.  According to the U.S. Chamber of Commerce, the healthcare industry has some of the highest number of job openings, and there are currently more unfilled openings than there are unemployed people with the necessary experience to fill them.  Tack on the fact that retirement rates in histology are high, 21.3% of supervisory-level histotechs are expected to retire over the next 5 years based on the 2020 ASCP vacancy survey, and that there are only 23 states in the country with at least one NAACLS-accredited histology program, and it becomes evident that the problem is likely to get worse before it gets better.  Even if all the education and training obstacles could begin to be addressed today, which needs to happen, it would still take years for changed to make a significant impact. 

This all is not being brought up to be a downer but to be realistic about the situation.  W. Edwards Deming, one of the grandfathers of Lean, said, “Hopes without a method to achieve them will remain mere hopes.”  Histology laboratories can simply hope that their current challenges will magically disappear, but the reality is that unless actions are taken and changes are made, well, hopes will remain mere hopes.

It has been established that laboratories cannot simply staff their way out of their challenges.  Workloads are likely to continue to grow, not shrink.  Laboratories are expected to improve quality and reduce risk.  Turnaround time expectations are likely to get shorter rather than longer.  The pressure to reduce costs is also not going to go away any time soon either.  So, what can be done to help mitigate these challenges?  The only real option remaining is updating and changing laboratory processes to become more efficient, and automation can be a tool to do that.

There he goes!  He said it!  The “A” word!  For those who may also follow the Histology Professionals Facebook page, you likely have noticed that every time someone posts something about automation, there is usually a flurry of posts from folks talking about how they are going to be replaced.  Many a “They took our jobs!” South Park gifs and memes immediately follow.  We have already established, though, that most laboratories cannot find qualified staff and that qualified staff do not simply grow on trees.  With staff burnout due to being short-staffed being such a big issue in histology right now, the risk of anyone losing their job because of histology automation is practically non-existent.  In fact, introducing new types of automation may be what finally gives some laboratories a bit of breathing room after suffering from being chronically understaffed.  Automation is a tool to be used to better utilize the value of human resources.  There are plenty of things a trained histotech could be doing with his or her time and experience rather than scraping paraffin off the side of cassettes after embedding, yet for many laboratories without automation, that type of task is a necessary evil wasting the time of staff.

That said, not all automation fits every circumstance.  Laboratories must do their due diligence, and continuous process improvement projects like bringing in new automation require planning and a commitment to seeing the project succeed.  Expectations must be appropriately set up front as well, which requires going into the project with an open mind but also with open eyes.  What laboratories should be careful to avoid is unrealistic expectations.  Sometimes out of desperation and exhaustion, laboratories bring in a piece of equipment hoping that it will be just a plug-and-play magic bullet that instantly solves all its problems.  The reality is that automation is a tool just like any other, with both benefits but also limitations, and it must be implemented correctly with compatible processes.  Unrealistic expectations and the lack of commitment to the change have doomed many a good project.

As mentioned before, automation is not entirely new to histology.  Automated processors and other equipment are commonplace, but it is important to remember that when those tools were first introduced, there was likely a very similar pushback as there is with new forms of automation today.  Some of you may have heard the tales, or spun those tales yourself, of longer tenured histotechs regularly performing manual coverslipping back in their day or sharpening their own microtome blades.  Even today there are laboratories manually labeling cassettes and/or slides rather than using printing technology.  If you talk to most techs about these more commonly adopted technologies and forms of automation now, they will tell you they would never want to go back to doing it the old way.

That leads to what new automation or technologies are out there right now that could help histology laboratories.  There are well-known technologies that are still not quite universal in laboratories like the cassette and slide printers just mentioned.  Printing versus handwriting is a great way to reduce risks of mislabels and other errors, especially if IDs are also connected to an LIS or interpreted through barcodes.  Printing of barcodes also enables other technologies to come into play like specimen tracking systems through the LIS or other means.  Even with printing, there are process changes that can be done to also make tasks more efficient and even further reduce risks.  Transitioning from batch, pre-printing of cassettes or slides to printing cassettes on-demand at grossing or slides on-demand at microtomy from printers localized at each station can significantly reduce the risk of mix ups, reduce waste, and save time from eliminating the need to use redundant checks to catch potential mistakes.  Another great W. Edwards Demming quote is, “Cease dependence on inspection to achieve quality. Eliminate the need for inspection on a mass basis by building quality into the product in the first place.”  By eliminating the chance for a mistake to occur, like two printed cassettes getting mixed up from being pre-printed and batched, there is no longer a need to establish time-consuming inspection steps.

Rapid tissue processors are another form of automation that has been around for a long time, but the term “rapid processing” is often mischaracterized or misunderstood.  Outside of fixation, tissue processing is one of the most time-consuming steps in the histology process, so improving processing efficiency can make a significant improvement to turnaround time.  Rapid processors have been established now for years but many laboratories continue to make do with conventional processors.  It does not help that if you read the brochures or listen to the sales pitches about most conventional processors these days, they often include lingo about “rapid processing” or other claims that are exaggerated at best.  Often laboratories may think they are doing “rapid processing” when they were actually sold another conventional processor that is just using an updated protocol versus the 30-year-old protocol they were using before.  If the processor uses traditional reagents (alcohol, xylene, paraffin – or substitutes), the same traditional processing variables (time, temperature, pressure/vacuum, and some form of agitation), and reagents are pumping in and out of a retort requiring the use of a cleaning cycle between runs, then you have yourself a conventional tissue processor regardless of the spin the sales rep may have given.  True rapid processors use differentiating variables that help speed up processing.  For example, they may use specialized or combination reagents, microwave-assist, or features that enable continuous processing.  For small tissue like GI or prostate cores, there is not going to be a major separation between the time processing takes between a conventional or rapid tissue processor.  Both can processor small tissues in that 1–2-hour range give or take, but the process flow of a rapid processor may provide other significant benefits.  Where the separation of processing time really becomes noticeable is for processing tissues that are larger or thicker.  A conventional processor may take a 6+ hour protocol to properly process thicker tissue that a rapid processor can likely do in around 2-3 hours.  That difference of several hours in concert with the potential for continuous loading can have a significant impact on the time for tissue processing for laboratories.

Automated embedding is another newer technology that is still underutilized in laboratories.  This technology has been available for more than a decade, but that is still relatively new in comparison to other histology technology and processes.  Part of the underutilization for automated embedding is that there are not widespread vendor options, and implementation takes a commitment to workflow changes.  Manual embedding, like microtomy, is one of the most labor-intensive tasks in histology.  The traditional embedding centers that everyone is familiar with do no actual production for embedding.  The human engine sitting in the seat, manually manipulating each and every tissue, is who creates embedded blocks.  Automated embedding is a complete paradigm shift in that regard, which can be intimidating.  With that major shift, though, comes a significant amount of value for the laboratory for time, human resources, and risk reduction.  There is no other technology readily available right now that can provide more significant benefits for staffing shortages, and arguably patient safety, than automated embedding.  The equivalent in terms of impact may be when the first automated tissue processors and stainers came about.  Automated embedding can enable laboratories to free up technicians normally shackled to embedding stations and utilize them in other areas like microtomy or anywhere else their talents and skills can be better used.  The process also provides many risk reduction benefits.

There are also newer technologies just starting to break though.  For example, there are already semi-automated microtomes that range from automating the turn of the handwheel to even automatically aligning the chuck with sensors, but, in more recent years, there has been the emergence of fully automated, robotic microtomes as well.  The microtomist in a box if you will.  Although not widely utilized and still with some limitations, they demonstrate the direction histology is going. 

Then there are all the newer technologies that standardized, high-quality histology processes help enable.  Digital imaging has been around many years at this point, but especially with the rules having relaxed because of COVID as it pertains to remote diagnostics, many healthcare systems got a taste of the benefit of utilizing digital imaging.  Add to that the additional benefits that are being developed using the digital slide image, like Paige Prostate, which is the first FDA-approved diagnostic tool that uses clinical-grade AI to improve efficiency in diagnosing prostate core samples.  Paige Prostate boasts a significant reduction in detection errors in prostate biopsies.  With new tools like Paige Prostate assisting and enhancing the pathologists’ own skills and experience, patients win.  These technologies come after the creation of the slide, though, so all the benefits that may come for patients, pathologists, etc. using these new technologies are lost if the laboratory struggles with producing a timely and quality slide. 

So, is histology automation a friend or a foe?  Like most things in life, the answer is not always black and white.  Automation implemented in the wrong way and for the wrong reasons, can easily become a foe for the laboratory, but the right automation implemented with commitment and clear expectations, can be a game changer for not only allowing laboratories to catch their breath and grow but also enhance patient care.  Laboratories can be certain that challenges are not going to magically disappear, and those challenges and changes in the world do not care about “but that’s the way we’ve always done it” excuses.   Change is inevitable.  Laboratories have the choice to either embrace change and control their own destiny by being proactive and properly evaluating, implementing, and committing to new technologies when they make sense, or the laboratory can forfeit its control and get dragged kicking and screaming and reactively deal with the inevitable changes later.  The challenges histology laboratories are facing are real and will only become more painful and costly for those laboratories that dig their heels in and resist adapting.  Bringing in new automation and changing long-held processes takes work, and change can be uncomfortable.  That short-term discomfort, though, can often alleviate a whole lot of chronic current and future pain.  In the end, whether automation is a friend or foe is our choice.
Created with