A Realistic Approach to Pathology Automation

Vol. 17 • Issue 11 • Page 46

The complexity involved in many anatomic pathology technical processes makes automating the pathology laboratory challenging on many fronts. Still, taking advantage of those steps that can be automated is vital to maximize efficiency and ensure skilled pathologists are able to devote their time to the most important steps-not caught up in paperwork or documentation.

Critical drivers for pathology automation are numerous and succinctly summarized by Becky Fitzgerald, HTL, MS-MOT, product manager, McKesson Life Sciences. Patient safety is of fundamental importance across all of healthcare, she says, but pathology processes are not robust. “The many manual hand-offs and data translations that exist in most laboratories are system and process problems that automation can address. Staff shortages drive the need for process improvements and efficiencies. Simultaneous reduction in reimbursements and rising costs of providing diagnostic services results in the need to do more work with fewer resources, the historic calling card of automation,” Fitzgerald explains.

The need to automate is evident, but each lab’s needs must be assessed on an individual level and the process must be approached realistically when considering implementation. Technologies exist that are simply too expensive for most laboratories-and setting unrealistic goals for increasing automation will lead to wasted time and efforts. Focus on attaining what is achievable for your lab, rather than trying to create a fully automated dream lab on a shoestring budget. Assess the lab’s operational needs to determine where automation is most important, then establish practical goals. No matter the manner in which the lab chooses to implement automation, any steps taken should assist in making pathology diagnoses more objective, reproducible, quantitative, reliable and standardized, says Tsafrir Kolatt, PhD, Applications Development and program director, Applied Spectral Imaging (ASI).

First Things First

Information systems that automate order entry, billing and results reporting for pathology are well-established and can meet the needs of various size labs and budgets. “By replacing mind-numbing, repetitive, manual activity with electronic capture of that work, users may focus their attention on the truly value-add activities while being protected from errors caused by poor processes and system designs,” Fitzgerald says.

“A key goal of pathology informatics automation in the anatomic pathology lab information system (AP-LIS) is automation of the order from the point of generation to the AP LIS via an electronic orders interface,” says J. Mark Tuthill, MD, division head of Pathology Informatics at the Henry Ford Health System. He also lists the extraction of billing information for particular cases and CPT and ICD9 code generation, automated tissue cassette and barcode labeling as processes that have benefited from automated informatics in pathology.

Chain of Custody

An information system should assist in completing the pure chain of custody, stresses Nasar Qureshi, MD, PhD, medical director, QDx Pathology Services. This ensures the lab is running at its most efficient and reliable level. The chain of custody begins when the sample is collected and reaches the pathology lab, and is not complete until slides have been filed and you can be certain that each slide produced for the case is signed off and in the storage cabinet, he says. The ability to do this is directly tied to a functioning information system, including barcode systems and automatic generation of labels for slides and cassettes that allow for positive identification.

Digital Advantage

A discussion of pathology automation must include digital pathology and the valuable advantages it presents. “Digital imaging transforms the fragile physical element, the glass slide, into an electronic element, the digital image. Glass slides must be physically transported to the pathologist, or the pathologist to the glass slides,” introducing the significant risk of broken slides, Fitzgerald explains. “Laboratories mitigate this risk by performing recuts.” But that necessitates cutting deeper into the block and risks missing the pathology of the original slide. She adds, “Glass slides are physically separate from the electronic record and related paperwork” that allow many avenues for error.

“Digital images, on the other hand, are electronically tied to the patient record. Pathologists review digital images and record results into the electronic record,” Fitzgerald continues. When physical elements are not changing hands, opportunities for errors are significantly reduced.


Already, the pathology lab has benefited from automation in microwave tissue processing, immunostaining and coverslipping. Dr. Tuthill lists automatically generating labeled tissue cassettes at the point of accession and generating slide labels from cassettes when cutting tissue in the microtome as previously highly manual areas prone to error.

But it’s instrumentation that presents some of the most significant obstacles in bringing pathology automation full circle. Dr. Qureshi expresses the frustration pathologists feel when trying to streamline operations-there’s a lack of communication and integration among the steps along the way. “The pieces are there, but the integration is missing,” he says.

Art, Not Science

Diagnosis, of course, remains an area that requires a knowledgeable, trained eye; it’s hard to imagine pathologic diagnoses ever being fully automated. Histotechnologists do not all produce the same slides; every tissue does not look the same, says Dr. Qureshi. “Replacing a histotechnologist, in my mind, is nearly impossible,” he says. “That’s not science, that’s art, and an art cannot be mechanized.”

However, artificial intelligence and computer-aided technology (CAT) are promising routes not for completely automating diagnosis, but for diagnostic assistance. Dr. Kolatt suggests that perhaps some diagnostic parameters will be replaced by more meaningful, reliable and measurable parameters to assist in decision-making.

“We all think CAT must march in and conquer the arena,” Dr. Kolatt asserts, but “we are not sure whether the pathology market will make one big leap ahead directly into the digital pathology realm or undergo slow evolution based on manual and semi-manual CAT.”

A quick transition to an intense new technology, however, is unlikely, especially in an industry that already tends to be slow to embrace change. A number of first adopters followed by an increasing number of users is more likely.

Fitzgerald points out some diagnostic assistance capabilities made possible with digital pathology. A pathologist viewing preparations on very large, multi-image panels can correct for vision changes, intensity and clarity dialing and may color-correct for poor staining and mark and enhance certain features.

“By color-contrasting, tissue boundaries may be highlighted, allowing for faster margin assessment,” she says. Algorithms applied to estrogen-receptor and progesterone-receptor and HER2/neu stains provide quantifiable values.

There’s also automation opportunities in areas such as cytopathology, where a Pap smear can be run through and declared normal or flagged, Dr. Tuthill suggests. “Diagnosis still defies automation,” he says, but there are likely some future opportunities as far as assisting diagnosis.

According to Dr. Kolatt, progress is needed in definitions and quantification of diagnostic parameters that are often subjective before we see this kind of automation come to fruition. Nevertheless, in some specific diagnoses, CAT is already in a position to enable more accurate, faster and repeatable results.

Challenging Change

While opportunities are exciting, a few obstacles remain at many points along the pathology automation spectrum-many of which have to do with the adoption of new technology, Dr. Tuthill reiterates. “Change management and getting people to think in new ways is always a challenging issue,” he says. To help ease resistance, seek staff input when initially considering new automation systems. Assess what will be most beneficial to the pathology lab as a whole. The benefits to the organization and patients are usually obvious, but make sure the personal benefits are clear as well-i.e., the pathologists will get to spend more time focusing on specimen accruals and making diagnoses rather than entering results or making cassettes.

Innovation vs. Integration

Challenge also lies in the inability to take all of the pieces discussed here and bring them together to come full circle, or what Dr. Qureshi refers to as “pieces of a puzzle that simply do not fit together.”

Companies strive for innovation, and in the process often make integration with other systems and processes very difficult, if not impossible. For those companies that will assist in integrating the pathology lab’s automation systems, it can be a very expensive process-and laboratories are usually left to bear all of the costs of integration, claims Dr. Qureshi.

This needn’t be the case, though, as some integration protocols are already in place-the HL7 interface software language communication standard, for example.

Thus, the next major breakthrough, Dr. Qureshi hopes, will be from a company deciding to place the focus on integration in the automated pathology laboratory-developing software that speaks to the major market players and providing a cost-effective mechanism of utilizing automation’s full potential. An open, flexible system from a company willing to work with individual labs is needed, says Dr. Qureshi. It should be customizable and able to grow, without being cost-prohibitive for so many labs.

A Realistic Approach

Technical issues may also reflect opposition. Technologies like whole-slide scanning can be too much for networks and storage capabilities in the typical hospital, explains Dr. Tuthill. “It may not be a problem in advanced places like pharmaceutical companies or cutting-edge research institutions, but in research hospitals or institutions that are constantly cash-strapped and struggling to keep up, the technology itself is an obstacle. Technology implementation tends to lag behind what’s actually possible.”

Thus, it’s necessary to assess your infrastructure capabilities when looking to implement automation. Have your facility’s information technology teams involved, and bear in mind the importance of setting realistic goals. Not all pathology laboratories can-or need to-implement every system available. Start with the basics, and work your way up.

Automation in pathology is necessitated by two things, asserts Dr. Tuthill-improving efficiency and eliminating error. The two are clearly related because “when people have time to do their job well, it eliminates error. People pressed for time make more mistakes.

“Automation eliminates defects and information systems are keys to implementing automation systems. By eliminating defects, we improve pathologic diagnoses through improved quality and efficiency, resulting in increased patient safety,” Dr. Tuthill concludes.

Kelly J. Graham is assistant editor.