Unleashing the Power of Digital Morphology
We spoke with Ahmed Bentahar and Amy Meitus on how digitizing microscopy can transform hematopathology
Amy Meitus, Ahmed Bentahar | | 7 min read | Discussion
The microscope is the greatest revolutionary tool of the lab, but how are today’s cutting edge technologies reshaping microscopy as a whole? And how do we use new tech to unlock untold potential for fields of pathology? We asked Ahmed Bentahar, Medical Director at Beckman Coulter Diagnostics and Amy Meitus, Chief Product Officer at Scopio Labs, those very questions.
What are the downsides to traditional manual microscopy? Conversely, does it have any strengths that are not yet replicated in digital pathology?
For at least the last two centuries, microscopy has been the mainstay of laboratory medicine upon which detection and diagnosis of blood-related diseases are built. Perhaps the biggest advantage of the manual microscopic process is that it enables 100X magnification and makes it possible for trained users to see intracellular details. However, it is dependent upon the individual experience of the user and is vulnerable to inter- and intra-user variability (1). It is also labor-intensive, time-consuming, and can be limited in its ability to provide a comprehensive understanding of the underlying condition (2). Additionally, manual microscopy requires experts to be present in the lab. Samples are not easily shared, as each lives solely on the slide. To share samples – for collaboration, consultation, or teaching purposes – experts must send the samples to others or invite them into their labs. This can hinder the diagnostic process and limit teaching capabilities.
Thus, there is increased demand for automated digital microscopy systems and decision support, capable of scanning and performing morphological analysis of blood smears (3). Labs with digital morphology systems that offer 100X magnification and a full-field view are uniquely poised to benefit from the classic advantages of the microscope, as well as the increased speed, standardization, and ease of collaboration provided by digitization.
How can digitalization improve hematopathology for pathologists and patients?
Digital pathology has emerged as a very promising technology in clinical hematopathology and it holds tremendous potential in replacing traditional microscopy-based diagnosis in the near future (4). Recent advancements in artificial intelligence and slide scanning techniques are revolutionizing the field of hematology, bringing digital solutions to a traditionally manual industry.
The adoption of digital hematopathology is bringing improvements to the diagnostic workflow, greater flexibility of resource allocation, potential cost savings (5), and ease of both clinical and research collaboration. It also enables professional interactions to take place remotely, facilitating consultation and training across multiple geographical sites as well as alleviating the effects of the critical shortage of lab personnel worldwide. Taken together, these contribute to more rapid diagnosis and treatment initiation, while enhancing the patient experience.
What are the main challenges you foresee in the digitalization of hematopathology?
Digital pathology is on the verge of becoming a mainstream option for routine diagnostics thanks to the significant advances in imaging and digitization. Implementation on a large scale may be challenging on technical, logistical, and financial levels, depending on the specific application and its technology.
Some of the main challenges in digital pathology are technical. These include long-term storage of high-resolution images, data transmission, data security and privacy. Prior to making decisions and investments, current system requirements should be specified, and future ones anticipated.
Important considerations for digital pathology adoption
- The scope of the application; does it provide a fully digital workflow or rather a partial one? Consider the volume of expected slides to be scanned and the scanning speed
- System requirements; for example, adequate storage space and IT resources
- Workflow integration; assess required handling and user-friendliness, laboratory information system integration, and the ability for remote connectivity
- Regulatory issues; is the platform/application FDA-cleared and/or CE-marked, does it meet the regulatory requirements of your diagnostic setting?
- Extent of training required
- Medical and clinical aspects; assess the availability of supporting clinical data and accumulated laboratory experience as well as the levels of repeatability and reproducibility.
Finally, it should be noted that the acceptance of digital pathology generally appears to be high, although affinity for digital work can vary among pathologists.
How important is it to make AI and other technologies usable and understandable to laboratory professionals?
For a decision support system, it’s important for users to be able to easily approve results, as well as make traceable changes on their own. For example, a user may need to manually change a cell’s category or add/remove cells from a category. Artificial intelligence has the potential to transform this process by pre-classifying or suggesting information based on data, ultimately improving the delivery of care. It can support improvements in clinical outcomes, patient experience, and access to healthcare services. It can also increase productivity and the efficiency of healthcare delivery, mainly by accelerating the time to diagnosis (6).
With that said, as laboratory medicine continues to undergo digitalization and automation, laboratory professionals will likely be confronted with the challenges associated with evaluating, implementing, and validating AI algorithms. Understanding both the capabilities and limitations of AI – and how it can be applied – will be useful to practicing laboratory professionals and clinicians (8).
How much do pathologists and laboratory medicine professionals need to know about the technical aspects of digital pathology and AI?
Like most technology, if digital pathology and AI are well integrated, those who use it are spared from needing to know much about the technical aspects. Take, for example, a smartphone – when everything is functioning correctly, the user doesn’t need to know how an LCD screen or the processor works. In fact, users don’t even need to know what these are. Similarly, laboratory experts using advanced technology don’t need to learn the technical processes of the devices, beyond basic functioning. But their expertise is very much needed on the clinical side; after all, they are charged with interpreting the results of scans and drawing conclusions from the data.
How much do they need to know about the ethical considerations (data protection, sustainability, accuracy) of an increasingly digital lab?
Laboratory experts need to know that the tools they are using are reliable, trustworthy, and accurate – just like the microscope. They need to be confident that the data is protected through a secure hospital network. Additionally, they need to ensure that their devices adhere to regulatory standards like ISO, FDA, or CE mark for safety and that the systems they use do not lead to any environmental sustainability issues.
One significant ethical consideration is the value being provided to patients. Labs should strive to contribute to ongoing improvement in healthcare delivery through the application of efficient high-quality technology. Another consideration is how varied, representative, and equitable the data is from which AI decision support tools were trained and developed. It is paramount to the performance of these tools in real-world settings and for fostering trust.
Is cost a concern for labs looking to upgrade their technologies? Do you have any recommendations for labs with limited resources?
The demand for clinical laboratory services continues to expand due to population growth, aging, and an increasing prevalence of chronic illnesses and comorbidities. Additionally, the critical shortage of lab professionals and the need for improved precision, calls for continuous optimization of laboratory processes. Digitization is ultimately a way to increase limited resource efficiency and improve laboratory processes.
Before considering the adoption of digital pathology, it is crucial to conduct a thorough analysis that examines the cost-benefit ratio. This analysis will provide valuable insights for making a long-term investment decision in a potential new standard of care. Aggregating a business case for digital pathology is not purely related to direct vendor costs. There are numerous quality improvements that encompass patient care, safety, consistency and standardization of the results, workflow enhancement, efficiency, and turnaround times.
Under the existing pressures, departments that do not adapt may find it difficult to provide quality services and solutions to both their healthcare professionals and patients (7).
Do you have any examples or illustrations from your own experience of how digitization has improved hematology labs?
There are many examples of how digitization has improved hematology labs, such as workflow benefits, decreased sample turnaround time, and increased collaboration. One specific example is the improved weekend workflow of a hematology laboratory – prior to implementing our platforms, the weekends were a challenging time for the laboratory. Patients that would come in over the weekend had to wait until the beginning of the work week for test results, and the staff was overwhelmed by an increased workload at the beginning of every week. With remote digital morphology, experts could review samples over the weekend without having to physically go to the lab, improving patients’ experiences and lessening the Monday workload. Additionally, due to increased productivity over the weekend, the lab was able to cut one shift a week.
- Z Tian et al., “Blood Cell Analysis: From Traditional Methods to Super-Resolution Microscopy,” Photonics, 9, 261, (2022).
- EJ Cho et al., “The efficient workflow to decrease the manual microscopic examination of urine sediment using on-screen review of images,” Clinical Biochem, 56, 70, (2018). PMID: 29655959.
- A Kratz et al., “Digital morphology analyzers in hematology: ICSH review and recommendations,” Int J Lab Hematol, 41, 437 (2019). PMID: 31046197.
- ReportLinker, “Global Digital Pathology Systems Industry Report” (2022). Available at: https://bit.ly/3WeU87I.
- BJ Williams et al., “Future-proofing pathology: The case for clinical adoption of digital pathology,” J Clin Pathol, 70, 1010 (2017). PMID: 28780514.
- EIT Health and McKinsey & Company, “Transforming healthcare with AI: The impact on the workforce and organisations” (2020). Available at: https://bit.ly/3Mj7E5X.
- K Paranjape et al., “The Value of Artificial Intelligence in Laboratory Medicine,” Am J Clin Pathol, 155, 823 (2021). PMID: 33313667.
- MG Hanna et al., “Digital Pathology Clinical & Operational Metrics,” Arch Pathol Lab Med, 143 (2019). PMID: 31173528.