The Path to Digital Pathology: 3 Obstacles, 3 Opportunities
It was about 2000 when Yale pathologist John Sinard, MD, PhD, first heard the prediction. “In five years, we won’t be using microscopes,” a respected peer quipped. “We’ll be examining all our slides as digitized images on computer monitors.”
Nearly a quarter-century later, Sinard reports: “I’m at my workstation, and my microscope is sitting right here next to me.”
Sinard’s mistaken colleague shouldn’t be faulted for repeating hype that started circulating around the turn of the 21st century. The only real error in the prediction was the timeframe.
Which is to say that whole slide imaging will indeed transform pathology everywhere pathology is practiced. It’s just taking longer to fully arrive at a national scale than early enthusiasts expected.
Soon to catch up
The pace of digital pathology adoption at Yale has been deliberately slow—cautious, even. Yale-New Haven Hospital is consistently ranked as one of the top academic medical centers in the country. Yale School of Medicine’s pathology operation appears on numerous listings of the best pathology programs in the country.
“A lot of people try out whole slide imaging for teaching or research tasks before bringing it into clinical workflows.”
Eder Lagemann, Co-lead Author, KLAS Report on Digital Pathology
“The reality is that [even well-resourced] medical centers generally don’t have tens of millions of dollars floating around for investments with uncertain returns,” says Sinard, who founded Yale’s pathology informatics program and serves on the Information Technology Leadership Committee at the College of American Pathologists. “Whole slide imaging costs a lot of money and has a very poorly defined ROI. It’s a tough sell for the leadership of both medical schools and their affiliated hospital systems.”
That said, Sinard knows pathology’s future is digital.
“We were starting down the path in a concerted way when COVID-19 hit,” he recalls. “The disruption put us behind the eight ball. Today we have a couple of slide scanners and some limited data storage, and we’re slowly deploying some individual use cases.”
Yale’s staged entry into the digital pathology arena offers an apt starting point to explore the proposition that, for those mulling digital transformation but not yet moving toward it, every obstacle can be recast as an opportunity. Here are three examples.
OBSTACLE 1: Financially speaking, digital pathology—aka whole slide imaging—is hugely expensive and notoriously ROI-elusive.
OPPORTUNITY 1: There’s no need to go “all in” all at once. Many pathology practices and departments have demonstrated the feasibility—and the strategic soundness—of digitizing in stages undertaken over many months, years or even decades.
In a 2023 KLAS Research study looking at 14 early adopters of digital pathology, most respondents—71%—said they’re digitizing in incremental phases. Only 29% had opted for a “big bang” approach.¹
Among the benefits a phased approach brings, according to the KLAS report:
- Pathologists’ comfort with digital pathology develops organically, leading to easier buy-in from a diverse group of pathologists.
- Issues are addressed in smaller groups before broader rollout.
- Approach requires less capital expenditure up front (organizations can begin with fewer slide scanners).
- Less-disruptive approach to workflows; scanning can be gradually increased.
“Most pathology departments purchase the components of digital pathology separately” as a comfortable way to spread out the cost of entry, says Eder Lagemann, co-lead author of the KLAS report.
Lagemann says slide scanners of the size needed for large academic medical centers run roughly $300,000 apiece, and most departments need several.
Other obvious costs include an image management system (IMS), often called PACS for picture archiving and communication system or multi-department enterprise imaging system, which render the scans for viewing on a monitor. Image-viewing software needssome serious compute power, so it’s usual to have to buy new computers as well as medical-grade monitors.
The image management system may have some storage capacity, but for long-term archiving of the very large files generated by whole slide imaging, dedicated storage solutions—on-prem, cloud or hybrid—are must-haves.
For a large academic medical center, the initial price tag for all of this is probably going to be around $5 million, Lagemann estimates. “Then again, those who want to adopt digital pathology in phases can buy a small scanner for maybe $50,000 and learn the technology that way,” he says. “A lot of people try out whole slide imaging for teaching or research tasks before bringing it into clinical workflows.”
OBSTACLE 2: Members of hospital leadership perceive the digitization of pathology as driven more by intra-specialty peer pressure than by sensible business planning.
OPPORTUNITY 2: Yes, ROI data is sparse at the specialty-wide level. However, there’s enough anecdotal evidence to build a winning business case for supporting the journey to digital pathology and, with it, pathology AI.
In the 2023 KLAS study on the state of digital pathology, 93% of the 14 early adopters included reported significant time savings. In addition, 77% saw improved clinical collaborations, 69% used the technology to mitigate a pathologist shortage and 64% improved their diagnostic accuracy.¹
Importantly, the study authors note a strong positive correlation between diagnostic accuracy and AI deployment. In fact, the association is more than strong—it’s exclusive. Most organizations that aren’t leveraging AI report that pathologist accuracy is no better with digital slides than with glass slides, the authors report.
“So far there’s no AI that produces real dollars. The main argument for going in this direction is that AI is coming and will revolutionize the field.”
John Sinard, MD, PhD, Founder, Yale pathology informatics program, Yale School of Medicine and Member, Information Technology Leadership Committee, College of American Pathologists.
AI isn’t doable without digitization, of course. This point alone may be enough for the many healthcare C-suites that have already gotten comfortable envisioning clinical AI applications as essential to healthcare’s future.
Meanwhile, those who follow the investment markets in pathology probably already know about one pointedly AI-motivated acquisition announced in November 2023. That’s when Australia-based Sonic Healthcare said it was buying “unprofitable” Pathology Watch of Murray, Utah, for $150 million.²
Sonic CEO Colin Goldschmidt called the acquisition “an enormously exciting step in Sonic’s transition to digital pathology and pathology AI,” adding that the transaction “offers multiple avenues for value creation.”
Commenting on the deal, Robert Tessier, co-founder of the Panel of National Pathology Leaders, points out that Pathology Watch has current yearly revenues of “only” $15 million and has barely been breaking even.
“Sonic Healthcare must believe it’s going to have a very significant return on its $150 million investment over time,” Tessier says. “I believe they expect AI will boost productivity in pathology in a time of potentially worsening pathologist shortages.”
Yale’s Sinard points out that, so far, the only pathology AI application that has FDA approval and is in relatively wide use is Paige Prostate, which flags areas of interest on prostate biopsy images.³ But he’s not skeptical about AI’s likelihood to penetrate pathology across the board.
“You really can’t argue against the sort of efficiency claims that say, ‘If we scan whole slides, tumor boards will be easier to do,’” Sinard offers. “Or ‘We won’t have staff constantly pulling glass slides. Slides won’t get lost as frequently. And in emergency situations, we’ll be able to view slide images elsewhere.’”
Until CMS issues a full set of CPT codes for reimbursing the use of AI in pathology, the technology’s ROI will come in “a lot of these soft things,” Sinard says. “So far there’s no AI that produces real dollars. The main argument for going in this direction is that AI is coming and will revolutionize the field.”
That’s a hard argument to counter, he acknowledges. And this alone makes the case for taking steps toward digital pathology sooner rather than later, he suggests.
“If one has not built the infrastructure and acquired the experience of working with whole slide images,” Sinard says, “one is going to be left way behind when that [AI revolution] happens. AI will not replace pathologists, but I definitely see a future in which pathologists work with algorithms to deliver higher quality patient care.”
OBSTACLE 3: The necessary spend on IT alone—enterprise-level infrastructure and storage as well as department-specific computers and software—is exorbitant.
OPPORTUNITY 3: The overall spend for digital pathology can be strategically navigated between a.) the best scanner supplier, b.) the best image management system vendor and c.) the best storage and IT infrastructure partner.
It’s true that IT tends to consume the largest single area of budget expense for digital pathology transition. This was borne out in a recent single-site study at a large tertiary cancer center in New York City.⁴
Relatedly, getting a full picture on infrastructure planning calls for factoring in network bandwidth, as moving around large digital files can present a special problem for departments housed in old buildings with outdated cabling. Pathology leaders at a large integrated healthcare system made note of this need as an important aside when they catalogued their saving of around $12.4 million through digital pathology implementation. They credited the financial gain mostly to improved pathologist productivity and histopathology lab consolidation (from several hospital-based labs into two main labs).⁵
“The answers to some key questions around storage aren’t as clear yet for digital pathology. There are still a lot of proprietary image formats floating around, so there’s the question of how to bring in proprietary image formats. That’s one key question right there. We think the answers will come as providers work through the pathology portion of their enterprise imaging strategies.”
Sidney Tate, Co-lead Author, KLAS Report on Digital Pathology
The good news is that the researchers and other experts agree: All challenges related to IT and infrastructure for digital pathology can be met by picking the right industry partners.
For starters, slide scanner purchases and service contracts can be straightforward, budget-driven decisions.
Shopping for image management and storage solutions, on the other hand, is more involved. It entails looking for industry partners who will work closely with pathology physicians and staff and IT teams over time.
A good place to start is with the 2023 KLAS Research report on digital pathology. The concise document looks at KLAS-validated IMS vendors by name and offers tips on vendor selection and management.
On storage, KLAS says most early adopters of digital pathology use tiered storage plans to allow image access for the life of the patient. Generally speaking, the older the images, the less readily accessible—and thus the less costly—they can be. Cloud storage is common for slower-access files, which are often sent to the cloud after living in on-prem hardware for 90 days post-pathologist interpretation.
“Outside of whole slide images, there is much more that can be stored for digital pathology,” the authors note. “Some early adopters have found it advantageous to store everything pertaining to a patient’s case—including microscopic images, gross images and scanned documents—for easy access across the organization.”
To this point, study co-lead author Sidney Tate adds that KLAS is turning its attention to enterprise imaging strategies—and finding pathology second only to point-of-care-ultrasound in interest levels among healthcare provider organizations.
Pathology naturally comes up in planning for enterprise imaging, Tate believes, because of the massive sizes of the files.
“A lot of organizations are asking: ‘Can my VNA [vendor neutral archive] handle this?’” Tate notes. “‘Can I really store these big pathology files on top of my images from all the other ologies?’”
A growing number of radiology PACS and enterprise imaging vendors are adding pathology to their PACS and viewer capabilities, Tate points out. So many pathology departments are joining in on the enterprise imaging capabilities already being used by radiology, cardiology, orthopedics and others.
“But the answers to some key questions around storage aren’t as clear yet for digital pathology,” he adds. “There are still a lot of proprietary image formats floating around, so there’s the question of how to bring in proprietary image formats. That’s one key question right there. We think the answers will come as providers work through the pathology portion of their enterprise imaging strategies.”
At Memorial Sloan Kettering, one course-altering lesson involved the need for dedicated staff who could identify digital pathology technologies to acquire and then evaluate proposals from bidding vendors. These team members also led the effort to secure institutional funding.
Facilitating this slate of duties meant accepting an additional category of labor cost. In the end, the researchers suggest, the technological deliverable was worth the human investment.
“These innovators,” Ardon et al. write, “dedicated time and efforts that resulted in the early adoption of digital pathology at Memorial Sloan Kettering.”
This story was created by the editorial team at Health Imaging, Radiology Business and HealthExec that are all part of Innovate Healthcare’s news sites. Educational support for the series was provided by Pure Storage.
References:
- Eder Lagemann, Sidney Tate, et al., “US Digital Pathology 2023: A Deep Dive into Early Clinical Adopters’ Successes and Lessons Learned.” KLAS Research, Nov. 10, 2023.
- Mark Tullis, “Pathology Watch Acquired by Sonic Healthcare for $150 Million.” TechBuzz, Nov. 16, 2023.
- U.S. Food & Drug Administration press release, “FDA Authorizes Software that Can Help Identify Prostate Cancer.” Sept. 21, 2021.
- Orly Ardon, Matthew Hanna, et al., “Digital pathology operations at a tertiary cancer center: Infrastructure requirements and operational cost.” Journal of Pathology Informatics, June 7, 2023.
- Jonhan Ho, Anil Parwani, et al., “Can Digital Pathology Result In Cost Savings? A Financial Projection For Digital Pathology Implementation At a Large Integrated Health Care Organization.” Journal of Pathology Informatics, Aug. 28, 2014.
To read the other two articles in this series, please click here:
Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.
