Pathology as it is practiced today in labs around the world is not all that different from the pathology of the 19th century. The microscope has changed relatively little since then, and Hematoxylin and Eosin (H&E) stained slides remain the bread and butter of the pathologist. Yet it is not for lack of technological development; in recent years, digital pathology whole slide scanners, digital worklists, advanced viewers, and AI technologies have all been developed to streamline and enhance pathology workflows. As with any new technology, the challenges that come with adoption and implementation, especially within a clinical setting, are what have held pathology back.
However, the field of pathology has at the same time begun facing a growing number of challenges – challenges that the profession can only fully address by taking the leap into the digital age. It is important for labs of all shapes and sizes to consider investing in the digital transformation process if they want to ensure they can keep pace with the challenges modern cancer diagnosis imposes upon strained pathologists.
The Problems in Pathology Today
First and foremost, pathologists today are facing increasing workloads. Better screening procedures have led to a jump in the number of specimens needing review, but the number of pathologists trained and available to complete that review has been in steady decline.1
In addition, as our understanding of cancer grows, the complexity of the information that pathologists have to report also grows accordingly. Some decades ago, a typical pathology diagnosis could be produced in a few short lines. Today, in the era of personalized medicine, a comprehensive diagnosis that includes not only the morphological aspects of cancer, but a detailed molecular profile must be produced so the appropriate treatment can be tailored by oncologists and other clinicians. This translates to an increase in the complexity of the work that pathologists have to report, that, coupled with the increase in workload and the decline of practicing pathologists, puts a significant burden upon labs around the globe.
The impacts of these challenges are also significant for the patients whose specimens are under review. The literature shows that pathologists, being only humans under considerable pressure, may miss cancer in up to 1-3% of prostate biopsies.2,3,4 Considering the over one million men who are diagnosed worldwide per year, this may reflect a large number of patients.5 Moreover, the subjectivity of Gleason grading and tumor quantification can result in difficulty assigning effective treatment paths.
How Going Digital Can Help
Digital transformation may be part of the solution to these problems. Digital pathology opens up new possibilities and offers advanced tools that help pathologists be more efficient. For example, sharing cases with colleagues is made much easier in a digital environment. Tagging and preparing cases for tumor boards is far less onerous than on glass slides, and can be done at the click of a button. Measuring tumor size and their distance to margins is faster and more accurate with digital measurement tools than using a ruler. The ability to view low power images with digital pathology permits pathologists to direct their attention more efficiently to the areas of interest that require further examination. In addition, finding the correct slide is easier and faster on a digital platform than on a busy desk or a large filing cabinet.
All of these advantages add up to help pathologists be more efficient on a digital workflow than on an analogue one. But perhaps the most interesting aspect of digital pathology is that it is a necessary step to implement artificial intelligence (AI) in pathology. Once glass slides are digitized, AI can help pathologists do the routine, repetitive aspects of their work. First, it can instantly read a whole-slide image and determine whether it is suspicious for cancer with a very high degree of sensitivity and specificity. It can then direct the pathologist’s attention to both the suspicious slides within a case and the suspicious areas on the tissue sample. This can help pathologists to better prioritize their worklist and spend less time on non-suspicious slides, save time finding suspicious regions on each individual slide, and increase diagnostic confidence and efficiency overall.
One clear example of how AI can help busy pathologists with reviewing prostatic biopsy specimens is by performing instant tumor grading and quantification, providing a Gleason score, and automatically calculating tumor length and percentage, thus helping pathologists be more efficient and accurate with prostate cancer diagnosis.
Importantly, multiple studies have found these AI applications to be trustworthy and accurate in a clinical setting, irrespective of pre-analytical slide variations. In a study published in Modern Pathology, when pathologists reviewing prostatic core needle biopsies were aided by AI, they showed an average 16% improvement in sensitivity, leading to ~60% reduction in diagnostic errors. In addition, pathologists aided by AI were significantly faster.6
Of course, like humans, AI is not perfect. While it can be very helpful in assisting pathologists with their increasing challenges, AI is designed to be an adjunct to the pathologist, and is in no way a replacement to human judgement. Think of it as an additional immunostain: it is a helpful tool to reach a diagnosis, but is not a substitute to the pathologist.
Case Study: Granada Hospitals Go Digital
Granada Hospitals is a network of 4 hospitals around the city of Granada, in southern Spain. In 2016, they served a population of circa one million inhabitants and were processing about 64,000 histopathology specimens per year with just 23 pathologists, leaving them struggling with many of the same problems that persist for labs all over the world today. They made the decision to take their labs digital to address these issues.
Within a matter of 4 months, the labs had completely transitioned to a digital pathology workflow. Dr. Juan Retamero, now Paige’s Medical Director, was an instrumental part of this transition, and saw firsthand the results it had for pathologists and patients alike. “Digital pathology and AI show benefits in terms of better efficiency, better user satisfaction, and more importantly, better diagnosis for the patient in terms of accuracy and improved overall patient safety.”
The pathologists at Granada Hospitals benefitted from:
- Improved case allocation that increased efficiency by 20%7
- Improved diagnostic quality, with most of the pathologists (78%) saying that the more accurate digital tools helped improve their diagnostic quality as compared to diagnoses rendered with a microscope.
- Greater pathologist satisfaction, with 100% of pathologists indicating that, given the choice, they would not want to switch back to using a microscope.8
Going digital in a lab setting is not without its challenges. But labs all over the world have proven, as is the case at Granada Hospitals, that the pros vastly outweigh the cons. In fact, they have shown that the benefits of going digital for both the care team and ultimately the patient are immediate and powerful.
Paige is dedicated to being a partner in the process of going digital. Designed by pathologists for pathologists, our platform and AI technologies are the next-generation solutions labs need to truly elevate their impact.
Learn more about how going digital can transform pathology directly from Dr. Juan Retamero in our Going Digital webinar.
1Metter, David M., et al. “Trends in the US and Canadian pathologist workforces from 2007 to 2017.” JAMA network open 2.5 (2019): e194337-e194337.
2Wolters, Tineke, et al. “False-negative prostate needle biopsies: frequency, histopathologic features, and follow-up.” The American journal of surgical pathology 34.1 (2010): 35-43.
3Carswell, B. M., et al. “Detection of prostate cancer by α‐methylacyl CoA racemase (P504S) in needle biopsy specimens previously reported as negative for malignancy.” Histopathology 48.6 (2006): 668-673.
4Yang, Chen, and Peter A. Humphrey. “False-negative histopathologic diagnosis of prostatic adenocarcinoma.” Archives of Pathology & Laboratory Medicine 144.3 (2020): 326-334.
5“Prostate cancer – statistics.” Cancer.Net (2022).
6Raciti, Patricia, et al. “Novel artificial intelligence system increases the detection of prostate cancer in whole slide images of core needle biopsies.” Modern Pathology 33.10 (2020): 2058-2066.
7Retamero, Juan Antonio, Jose Aneiros-Fernandez, and Raimundo G. Del Moral. “Complete digital pathology for routine histopathology diagnosis in a multicenter hospital network.” Archives of pathology & laboratory medicine 144.2 (2020): 221-228.
8Retamero, Juan Antonio, Jose Aneiros-Fernandez, and Raimundo G. Del Moral. “Microscope? NO, thanks: user experience with complete digital pathology for routine diagnosis.” Archives of Pathology & Laboratory Medicine 144.6 (2020): 672-673.
