Volume 49 Issue 5, October 2022, pp. 547-555

The advantages of digital pathology (DP) have been recognized as early as 1963, but only within the last decade or so have the advancements of slide scanners and viewing software made the use and implementation of DP feasible in the classroom and in research. Several factors must be considered prior to undertaking the project of implementing the DP workflow in any setting, but particularly in an academic environment. Sustained and open dialogue with information technology (IT) is critical to the success of this enterprise. In addition to IT, there is a multitude of criteria to consider when determining the best hardware and software to purchase to support the project. The goals and limitations of the laboratory and the requirements of its users (students, instructors, and researchers) will ultimately direct these decisions. The objectives of this article are to provide an overview of the opportunities and challenges associated with the integration of DP in education and research, to highlight some important IT considerations, and to discuss some of the requirements and functionalities of some hardware and software options.

Light microscopy has played a vital role in many scientific discoveries and has provided major advances in all biological fields. Traditionally, pathologists have used the light microscope to assess tissue architecture (normal and diseased)—the major advantage being the ability to study cells in situ. However, with the integration of digital pathology into the workflow of veterinary laboratories, curricula, and research, we are able to move beyond the light microscope to the use of the virtual microscope (VM). Digital pathology (DP), at its most basic level, only requires Internet and computer access to visualize microscopic images after glass slides have been scanned. Software to manage and analyze the data that are generated from the digitized glass slides can also be added to the workflow, increasing the capability of the digital platform. However, the first step in both the traditional and DP workflow is to properly process a tissue sample and make a high-quality glass slide (Figure 1).1,2 The next step in the DP workflow requires a slide scanner to make a digital version of the glass slide, referred to as a whole slide image (WSI). Software (most often provided as an accompaniment to the scanner) is then used to view and/or analyze the digital slide.3 The viewing of digital images has progressed to the point that resolution of a slide on a computer screen is a remarkably close equivalent to that viewed with a light microscope. This is illustrated in Figure 2. Note that there are variations in color perceptions, which can be adjusted to the user’s liking by modifying the computer monitor’s settings and/or the color parameters in the viewing software. A major advantage of VM over a traditional microscope is also depicted in Figure 2. As shown in the first comparison, the lowest magnification on VM can capture the whole slide, essentially providing a sub-gross image, whereas using a traditional microscope the lowest magnification (usually 2.5 ×, as shown here) still only captures a fragment on the tissue section that is present on the slide. Ultimately, DP encompasses the use of both hardware and software to form a WSI that can then be used for histologic assessment in a research, academic, or diagnostic environment. The WSI also facilitates distant collaborations, remote learning, and off-site consultations.

Figure 1: Comparison of the traditional pathology versus the digital pathology workflow

Note: The traditional (top pathway) and digital (bottom pathway) workflows both start with making a high-quality glass slide. In the traditional flow, after the glass slide is made, the paraffin blocks are stored, and the slide(s) are physically delivered (by hand or mail) to the students or instructor to be used in the classroom or to the researchers. The key feature here is that the users must have access to a physical microscope that is generally tethered to one location. The DP workflow differs from the tradition flow in that the glass slide is digitized and the digital image is archived in a data storage site. The digital image is delivered electronically to the students, pathologists, or researchers. It can then be managed, viewed, and analyzed at a computer workstation.

Figure 2: Comparison of images captured with a light microscope versus a virtual microscope

Note: A slide of a mouse’s liver was photographed with the light microscope’s attached camera at 2.5 ×, 20 ×, and 40 × magnification (top panel). The same slide was scanned in at 40 × magnification and the digital images were then photographed with a virtual microscope in approximately the same location for the 20x and 40x images at the lowest magnification (0.2 ×), 20 ×, and 40 × for comparison (circled section was photographed).

This article aims to give an overview of major points to consider and available options when undertaking the integration of the DP workflow into a veterinary setting—specifically focusing on education and research. The requirements and recommendations for the transition to a DP workflow for primary diagnoses in a clinical diagnostic veterinary laboratory are outside the scope of this article and generally entail different requirements (hardware, software, security) to meet the goals of the lab. Additionally, the field of DP is quickly and ever evolving, and as such, manufacturers continue to develop products, and options may have changed from the time this article was written.

Deciding to integrate DP into the workflow can be an essential but potentially expensive and time-/labor-intensive project. It requires buy-in and support from pathologists, histotechnologists, information technology (IT), and executive leadership. Scanners can cost $100,000 USD or more and require annual and ongoing service and maintenance contracts as well as technical support, which can also be costly.4 As well, continued education and training of the DP team, data management, and delivery will be additional expenses. One of the first, and arguably most crucial, steps in the implementation process is to communicate with the IT team at one’s institution. Full support and open, continuous communication between the IT team, users, and vendors is paramount to the success of this process. Some of the institutional requirements and regulations under the purview of the IT department that are required to institute a computer-based platform will be discussed below. The Digital Pathology Association (DPA) recently provided a thorough and practical guide that highlights some of the requirements associated with using WSI in pathology, education, research, and collaborations.5 Members of the DPA have also published a white paper that provides computational pathology definitions, best practices, and regulatory guidance.6

COVID-19

In general, adopting WSI for teaching and research, as well as diagnostics, presents many challenges and opportunities. Institutions’ lack of readiness and abilities to transition to a DP workflow in the face of necessity were particularly highlighted during the COVID-19 pandemic. While research performed by pathologists tends to be a solitary endeavor, the use of a microscope for teaching of and training in pathology (human and veterinary) has been the cornerstone of pathology education and is an interactive, group endeavor. The social distancing safety requirements and recommendations that were instituted during the pandemic essentially restricted face-to-face learning, forcing academic institutions to rapidly adapt in order to continue teaching microscopy and pathology. Many of the technology and hardware software requirements are discussed in detail below; however, here we will highlight a few lessons learned during the pandemic.

In a special collection article published in 2021 by Academic Pathology, Christian and VanSandt discuss some of their perspectives on pathology education subsequent to the pandemic.7 In a cross-sectional study of pathology faculty and trainees, they found that dynamic virtual microscopy offered a cost-conscious alternative for pathology education to be delivered remotely while allowing both educators and trainees/students to actively participate. The virtual microscopy platform was essentially unchanged from the regular workflow (with the obvious exception of the lack of teaching in tandem at the microscope). Glass slides were digitized and delivered to the residents, who then evaluate and annotated the slides using their VM. Videoconferencing software, which had always been available but was rarely used prior to the pandemic, was used to allow for simultaneous review between the trainee and pathologist. The faculty and trainees were surveyed about various aspects of the virtual and optical platforms 6 months after implementation. Both advantages and disadvantages were discovered when compared with teaching via traditional microscopy. The advantages and disadvantages were similar to what has been found in recent studies, and anecdotally.810 Advantages include the ease of use of the VM and videoconferencing software; ability to maintain physical distance; that annotations can be performed in real time by instructor and trainee and is more detailed; increased attendance, participation, and enthusiasm at conferences/webinars by those within and outside of the departments; that one instructor can reach/teach more people at one time; convenience; that trainees are more easily assigned responsibility; easier requests of consultations; and easier onboarding of new faculty/trainees. Disadvantages include loss of visual cues and other social aspects of learning; that technology needs sometimes were greater than what was individually available (i.e., bandwidth, quality of monitors and cameras, video cards); lack of optimization of the virtual system; and the cost of some components (primarily scanners).710

As previously stated, this article will primarily focus on DP integration in veterinary academia for the purposes of serving research and educational needs. Also important to highlight is the fact that we discuss solutions that will accommodate anatomic pathology and basic histology. Clinical pathology (i.e., cytopathology and hematopathology) will likely have different requirements based on the need for higher resolutions and magnifications and, potentially, scanning in multiple layers. In some instances, purchasing a separate scanner to meet these needs, rather than trying to have one scanner perform all functions, would be more efficient. With these goals in mind and after determining the budget for the project, our opinion is that the main functionality considerations for a scanner are slide capacity needs, image capture resolution/magnification, and imaging modes (brightfield [BF] for education and research with the likely addition of fluorescence [FL] capability for research). Institutional IT infrastructure and compliance requirements should be considered as well. Some vendors do not focus on IT security and/or compliance and do not actively develop their applications to adhere to the latest standards. Selecting a vendor that has a well-rounded team that actively considers IT security and compliance is important for long-term success of the project.

Security

Many vendors that support science do not necessarily focus on IT security. IT security is responsible for identifying, assessing, and controlling the threats and liabilities that may be associated with the electronic transmission of potentially sensitive and/or confidential material. This vulnerable material may include digital slides, privileged information that may be contained in the distribution lists that receive the images, and the metadata associated with the digital file. To be able to use the data that are both acquired and stored, the scanning computer, the users’ computers, and the storage servers need to communicate through a network. In addition, federal, state, and organizational regulations often require frequent repairs to address security vulnerabilities (patches) or operating system (OS) upgrades. At the initiation of discussions with the potential vendors that will be used within one’s DP system, vendors should be made aware of any requirements imposed by IT, as well as the scheduling requirements for patches and upgrades. Therefore, selecting a vendor whose platform does not or cannot support an institution’s IT required changes should be avoided, and having the IT team integrated from the outset is the best way to be aware of institutional requirements and product vulnerabilities and limitations. Evaluation of the product’s life cycle is also important. If the product is in the early part of its life cycle, all the features that are required by IT may not be included in the current version. The opposite is true of a product that is in the declining phase of its life cycle. These older products may not receive scheduled patches or feature updates. Considering the monetary and human capital required for these projects, investing in a product that is not being regularly evaluated and developed by the vendor or one that is at the end of its product life cycle would be a poor choice.

Data Classification and Business Impact Plan

Storage of the acquired images will need to be planned prior to implementation. Determining the data classification and business impact will determine how the stored data are managed or categorized. Data classification informs IT and the entity responsible for the DP service of things such as where the data are located, what type of data is being stored, what access levels are available for each data type, the length of time that images should be stored, if the images should be backed up (and if so, for how long), and, if needed, whether there is sufficient redundancy to ensure that if components fail, service will be maintained or downtime will be minimal. The business impact plan will determine the answers to such questions as the following: (a) How are the data classified (i.e., is it public; subject to the Health Insurance Portability and Accountability Act [HIPAA], the Family Educational Rights and Privacy Act [FERPA], or Department of Defense [DOD] restrictions and regulations; etc.)? (b) If all the data that have been acquired are lost, how would that impact the organization’s ability to operate? (c) Can the data be easily re-created, if needed?

Storage Size

Storage that will be required over a time span of at least 5 years should be estimated to ensure that it is of sufficient size to sustain the project. To obtain a good approximation of the amount of storage a project will need, first evaluate representative samples of the images that will be generated to assess their average size. Next, determine the expected quantity of these images that will be collected and retained in a given year, and finally extrapolate this amount over 5 years, allowing for an excess of at least 10% unfilled space. The unused space should be reserved for IT maintenance operations.

Retention Plans and Governance Frameworks

Retention plans and governance frameworks should be documented and detailed in order to prevent data sprawl (the amount of data produced by the organization for daily business that may or may not have historical value) and duplication on the server. For example, slides scanned for research might be held for only a short time in a facility’s storage before being transferred to the investigator who owns them. On the other hand, one may be required to store images created from scanned teaching slides indefinitely. Determining prior to implementation what users’ storage limits and locations are, as well how the data will be managed, will allow user governance of their images and storage space.

Whole Slide Scanners

In general, whole slide scanners are fast and reliable and produce high-quality digital images. A non-exhaustive list of some commonly used whole slide scanners and their vendors is provided in Table 1. A multitude of options is offered; thus, scanner selection is based on the needs and goals of the users and the lab. For example, slides can be scanned and the resulting WSI can be viewed in BF or FL, in two or three dimensions, and at a variety of native or digitally zoomed magnifications; with some scanners, slides can also can be finely focused with extended focus or z-stacking. However, keep in mind that adding scan layers, increasing magnification, and scanning in FL all dramatically increase the digital file size and the scan time.

Table

Table 1: Summary of selected hardware vendors and their available scanner models

Table 1: Summary of selected hardware vendors and their available scanner models

Vendor Models
Morphe Index-1, Optimus-6, Continuum-48
Hamamatsu NanoZoomer S360, S210, S60, SQ
3D Histech Pannoramic 250 Flash III, Scan II, Midi II, Desk II DW, 1000, Confocal
Olympus VS200
Leica Biosystems (Aperio) Aperio AT2 DX, GT 450, AT2, CS2, LV1, Versa
Philips IntelliSite UltraFast Scanner
Zeiss Axio Scan.Z1

A whole slide scanner is similar in its composition to a traditional light microscope; however, it is under robotic control and must be connected to a specialized camera and a computer. For example, all scanners contain a microscope that has lens objectives, a stage/tray to hold and move slides around a light source (BF or FL), and software that allows for varying degrees of viewing during scanning but is also required to manage and view the final digital slide. As with a light microscope, the lens objectives range from low magnification (2.5 ×) up to high magnification of 40 × (or even 60 ×). Similar to a light microscope, lens power, number of lenses available, and lens quality all correlate to a machine’s cost. As mentioned above, one must also keep in mind that scanning at a higher magnification takes a longer scan time and produces a larger digital file to be archived and shared. The image resolution is determined by the scanning objective used, the numerical aperture of this objective, and the camera’s quality and also affects the final size of the file.

The total number of slides that can be loaded at one time (slide capacity) varies greatly among scanners. Scanners range from being able to scan a single slide to scanning hundreds of slides per load. Scanners may support continuous loading (slides can be uploaded while another is being scanned) and manual and/or automated scanning. Correspondingly, the footprint of the scanner increases with the slide capacity. Some scanners can only accommodate the traditional glass slide size (3” × 1”) and about 1 mm thickness, whereas other scanners have trays that can hold slides that are up to 6” × 8” (e.g, whole mouse brain slices). Very few scanners have oil or water mount capability. Most require a dry mount or a vertically positioned slide that has a glass or film coverslip. These constraints may impact scanning some slides for clinical pathologists and should be considered if this will be the primary usage area.

Scan speed varies by manufacturer and ranges from scanning a single slide (at 20 ×, BF) in less than a minute to up to 3 minutes. Again, scanning at higher magnifications, in FL, and/or with additional layers all increase scan time. In an education or research environment, the volume and turnaround time are typically low and longer, respectively. Although initially there may be many slides that need to be scanned in order to build a repository for teaching or to complete a large research project, for example, this can generally be accomplished at a slower pace, thereby alleviating the need for scan speed (generally reported out as the time it takes a scanner to scan one 15 × 15 mm section of tissue on a glass slide at 20 × magnification in BF mode) to be a major consideration. Similarly, research projects can take months to years to complete, and histopathology is but one component of the whole story. Thus, it is not unreasonable or unexpected to have labs with scanning turnaround times of 2–3 weeks. Contrast this with what would be required to support WSI in a diagnostic laboratory workflow where the turnaround times must be short (1–2 days), the daily volume can be consistently high, and at least two machines are needed to contend with the potential for failure or malfunction of one of the machines.

The computer workstation offers a distinct advantage in the realm of DP. Having access to a digitized version of a glass slide allows the pathologist, instructor, and/or student to view the slides from any location with a computer and an appropriate Internet connection. Routine use of smaller mobile devices such as an iPad to evaluate digital slides has even been shown to have almost perfect concordance with evaluation of the whole glass slide in side-by-side comparisons.11 Using smaller devices may be appropriate for telepathology (defined as using technology to facilitate the transfer of pathology data between distant locations); however, it doesn’t usually outperform using a larger computer workstation, given the cumbersomeness of navigating a touchscreen and the often inferior quality and small screen size.1 A robust DP workstation usually consists of a computer and one or more high-quality monitors and other input devices. The computer speed, monitor(s) size and resolution, and workspace ergonomics are some key features to consider when designing the workspace.12 Network speeds of at least 20–100 Mbps should allow for smooth downloads with continuous viewing of images. Most rural home networks do not yet have this download speed available and also have monthly data caps, which limits how much data can be received per month. Unlimited data plans are preferred due to digital learning’s high variability of requirements. We recommend providing minimum data and connection requirements needed for digital learning as a part of student orientation. This is especially important for individuals who live in more rural environments and likely have a a slower Internet connection to consider. Congress has instituted programs that include both loans and grants designed to promote broadband deployment in rural areas where 90% of households do not have sufficient access to broadband at 10 Mbps download and 1 Mbps upload speeds.13 Latency (the time a network packet takes to get from machine to machine) is also a factor in speed requirements but is less important since real-time viewing is usually not a necessity in an academic environment. Keep in mind that file servers (e.g., Dropbox, Windows shares, email) require that the whole file be downloaded, which requires more bandwidth. Contrast this to more modern image servers (e.g., Hamamatsu, Prosica, Aperio) that are specifically designed to manage and distribute WSI and will only transmit the part of the file required for current viewing (utilizing a smaller thumbnail view to display the whole slide). Larger monitors allow for a wider view field, and higher resolutions improve the quality of the images viewed. It is suggested that the highest-resolution medical-grade color display monitor available on the market be used so that the display matrix size is as close to the raw image data as possible.12 The need to increase computer and laptop use also necessitates ensuring an ergonomically sound working space. Sahu et al.14 have thoroughly reviewed ergonomic interventions to prevent musculoskeletal disorders due to awkward postures adopted during prolonged use of computers and laptops.

In the realm of medical education, pathologists recognized as early as 1963 that digitized light microscopic images coupled with computer-assisted teaching were conducive to accelerated learning of pathology, but only within the last decade or so has actual use and implementation become technically feasible, practical, and affordable.15,16 Recent reviews that discuss the use of digital images in education and diagnostics have predicted that it will be inevitable that their use will become the norm in the medical curriculum and the whole of diagnostic pathology.17,18 Indeed, digital microscopy has been so successfully implemented by some that in 2018, members of the German Medical Education Society proposed that digital teaching and digital medicine become a national initiative.1922 Additionally, others have shown that integrating histology and histopathology using virtual slides increased teaching efficiency.23 Medical students have easily adapted to the digital format, and in a comparative analysis of learning outcomes, better student learning and higher self-efficacy were confirmed with digital slides over conventional slides for learning hematopoietic cellular differentiation.24 A comprehensive systematic review found that medical students’ performance using digital microscopy was enhanced, and this was the preferred learning modality, even though pathology residents favored conventional optical microscopy.25 Clearly, the use of digital images for teaching pathology should be a standard component of human and veterinary medical education.2628

In 2006, Zwingenberger and Ward29 developed an open-source software and templates to create functional digital imaging files linked to case history and standardized nomenclature for teaching veterinary radiology, which in turn has facilitated the teaching of veterinary pathology in the veterinary medical curriculum and residency programs in anatomic and clinical pathology. Further evaluation of the usage of digital images for teaching histology to veterinary students found that teaching and personalized learning were improved in the context of practical classes.21

In 2016, a direct comparison of teaching histopathology to third-year veterinary students using digitized slides versus glass slides demonstrated that while neither format consistently outperformed the other, students identified the digital format as having many advantages.30 The authors noted that despite general acceptance of VM, students still prefer a traditional microscope and glass slides for examination (particularly for cytopathology).30 Irrespective of this stated preference, it should be noted though, that no significant difference was found when the results of histopathologic and cytopathologic examinations of certain specimens using VM versus traditional microscopy were compared. The authors of that manuscript also highlighted that in order to support traditional teaching, a dedicated laboratory equipped with microscopes and class slide sets (which, as a notable disadvantage, would undoubtable have to be produced in sufficient quantity to support class sizes and often have to be reproduced due to fading of color or breaking of slides) is needed.30 DP, on the other hand, can be taught using student computer facilities and, importantly, only one slide set need be produced. Additionally, it can be retained at the original quality level indefinitely.30

Bertram et al.31 extensively reviewed DP technologies as applied to the fields of veterinary pathology, concluding that DP will eventually become part of the workflow of veterinary pathologists and that trainees should be taught DP as digitization becomes the normal practice. They also provided succinct tables outlining the pros and cons of DP in a diagnostic service and the advantages/disadvantages of VM in veterinary education. Next, veterinary students’ change of perception toward VM use in their educational program was analyzed. Overall, VM was found to be acceptable—notwithstanding technical challenges, such as slow transmission speeds.31 However, the skill of using light microscopy was still well appreciated and considered to be important for the students’ professional education.1,32

The use of whole slide imaging received unanimous endorsement by trainees for distance teaching of pathology, and a randomized controlled trial found that using virtual microscopy was much more effective than using conventional microscopy was for teaching cytology to veterinary students.32 Lin and Johnson, in their special review, focused on relevance to online and digital learning, deployment, impact factor, scope, and appeal to a diverse population.28 They provided valuable insights into the future of digital teaching and learning, clearly illustrating the strengths, challenges, and future suggestions aimed at enhancing the digital age of education at all levels.28

An additional and valid concern raised by veterinary professionals and veterinary students is that with the transition to DP, students will not develop sufficient microscopic skills to be practice-ready. As determined by Stewart et al.33 in a survey published in 2014, the most common use for a microscope in practice is the evaluation of cytological specimens, and currently, sufficient time in the curriculum devoted to developing the microscope skills is needed to perform this function. However, as schools integrate technology into the curriculum, less and less time is being devoted to the traditional microscope. Stewart et al.33 urged academic institutions to consider ways to preserve this training in the curriculum. However, a relatively new service provided by Scopio™ (ScopioVet™)34 could offer an alternative for practitioners that need rapid analysis of cytology and hematology samples and could fill the gap left by the loss of microscopic instruction in veterinary school. In lieu of analyzing samples themselves, practicing veterinarians that have the ScopioVet digital cytology system in their clinic can scan prepared slides with this machine and then upload them to the network of clinical pathologists that are available to provide analytical services. Scopio Labs reports that slides are evaluated and interpretations provided within 1 hour. The scanning machine integrates with a cloud-based image management system that can also be used to manage slide images that will be used in education (or research).34,35

Nevertheless, if WSI and VM will constitute a major portion of the curriculum, prior to enrolling for classes, veterinary students should be informed of the minimal and preferred equipment and technical IT requirements that will enable them to successfully engage the digital learning environment. To assure an efficient and effective learning environment for teaching veterinary pathology using digital technologies, veterinary students need reliable, high-speed, cloud-based web access over time and distance, with high capacity to download high-quality digital images on multiple devices (e.g., laptops, pads, smartphones) and platforms (e.g., PC and Mac)—unless a universal device requirement is instituted—under secure password-restricted and protected conditions. The students will need access to cloud-based websites or applications installed on their devices to view, fully manipulate, change magnification of, measure, and annotate WSIs. Access to online videoconferencing software for communications between the instructor and students in a virtual classroom-like setting facilitates and enhances the learning experience. The digital environment will also need to be designed for compliance with the requirements of the 1990 Americans with Disabilities Act (requests for allowable exceptions may be required). Students seeking financial aid should also be aware that the cost of a laptop computer, up to certain limits, can be added once while they are in the DVM program to their budget for purposes of computing financial aid under the guidelines of the US Office of Education.

Table 2 displays recommended configurations and capabilities that will satisfy most requirements for students that will be taught utilizing WSI. It is essential for students’ preparation and for the successful integration of WSI into a curriculum that these recommendations be known to the students prior to them entering the professional veterinary program (i.e., before they even begin year 1). Having these configurations be a requirement and/or requiring all students to purchase the same device, which is managed by the college’s IT department, has been shown to be a highly effective standard operating practice. Another important consideration is that some components of the DP system (viewers, analytical software, etc.) may be incompatible with the Mac platform or may require additional programs and requirements to function properly on a Mac-based machine. Additionally, students are strongly recommended to ensure their personal devices, specifically laptops, tablets, and phones, have disk encryption toggled on. This will keep data safe should the student’s device be stolen. All software and OS should be set to automatically update, which will ensure that the latest patches are installed and the most recent OS is being used—particularly for mobile devices. It is also suggested that students set their screen savers to the 10–15-minute setting on all their devices and that they consider purchasing a privacy screen for laptops and or tablets. It’s also advisable to make sure that passcodes and/or biometrics are enabled for devices to ensure data security. The method of creating and maintaining student user accounts should be evaluated. It is important to determine whether the product can take advantage of common directory services or is able to integrate with an organization’s identity management office. This office regulates electronic identities and as such can make it possible to automate user account creation and deletion in order to decrease maintenance time by the lab that controls the DP platform or by the instructor. The automated management of user accounts greatly diminishes overhead management while still maintaining sufficient security of identities.

Table

Table 2: Example of configuration recommendations that would facilitate students’ use of DP in their curriculum

Table 2: Example of configuration recommendations that would facilitate students’ use of DP in their curriculum

Hardware and software Recommended for PC Recommended for Mac
Hard drive ≥ 256 GB M.2 solid-state drive ≥ 256 GB solid-state drive
Memory 16–32 GB of RAM 16–32 GB of RAM
Operating system Windows 10 Professional—current version OSX 10.15
Processor Intel 10th generation i7 Intel 10th generation i7
Screen 13–15" 13–16"
Other requirements PC and Mac
Internet speed/bandwidth 2.5 Mbps or better download speed (at least 10 Mbps is preferred)
Software PC and Mac
Office suite Microsoft Office
Antivirus protection TrendMicro Antivirus

DP = digital pathology; PC = personal computer; RAM = random access memory

Note: Table modified from Jones-Hall.48

Digitizing slides for VM viewing allows users on-site (e.g., the classroom) or off-site (e.g., the user’s home or a remote work location) to view, photograph, navigate, annotate, and analyze the WSI. Many WSI systems include image-viewing software that can be installed locally on user computers or cloud-based websites. However, in research applications where quantitative analytics have been automated, the ability to bypass the image viewer and directly access the images is often more efficient in the workflow. Additionally, image management systems that can organize and access images using image metadata, patient information, or other features that can collate images into groups may also be a useful addition.5

As previously mentioned, appropriate tool selection is generally determined by the users’ goals. For research, advanced viewing tools and software that support the ability to apply algorithms to WSI, annotate images and regions of interest (ROI), photograph, and export images are typically needed. The DPA recently published a review that nicely outlines some considerations for integrating image analysis in a research and diagnostic/clinical setting.2

In a research (and diagnostic) setting, image analyses need to be accurate and reproducible. The key elements in making this a successful endeavor include quality slides (sectioning, staining, and scanning) and a resulting high-quality image, ROI selection, algorithm selection, and the expertise of the pathologist who will be able to correlate the image analysis result with the relevant clinical information to make sure it is appropriate for a particular patient or case. Specifically, in the veterinary research setting, animal models are often used in early studies and play a pivotal role in drug discovery and development, not only for proof of the concept studies of efficacy but also for drug safety assessment. Image analysis, deep learning, and machine learning algorithms can be used to acquire objective measurements (quantitative and qualitative analyses) by automatically identifying histomorphological features and/or ROI in tissue, categorizing or segmenting individual cells or anatomical features, or quantifying the expression levels or location of specific biomarkers (histochemical or immunohistochemical) within a tissue or cell.2,36

The image formats that are supported by the product should also be considered. Many vendors offer products that support their own image format well but may not support other file types. Using a product that is as agnostic to the image source type as possible will prepare an organization to take advantage of the image management and analysis software that is best suited for the users’ needs.

For sophisticated and complex image analyses, several software options are available on the market. Many can be seamlessly integrated into the systems that the laboratory uses for scanning, management, and viewing. Ultimately, the goal of computer-assisted image analysis is to efficiently and accurately provide quantitative and objective measurements of the digitized slide’s features. This allows the pathologist to move beyond the traditional, descriptive, and invariably subjective assessment to automated and faster evaluations that produce objective (and reproducible) data. In addition, image analysis of WSI addresses other critical issues facing veterinary pathologists in research—namely the following: (a) the shortage of pathologists available, (b) the biases noticed when pathologists operate in an isolated environment, and (c) the natural intra- and inter-tissue variability seen within studies.37

Image analysis software vendors offer services in an à la carte fashion and/or in tiered packages with different pricing options. Free alternatives are also available. A popular free option made available by the National Institutes of Health is ImageJ,38 which offers many common tools that are used for histology images, and the Fiji package can be added to access more advanced algorithms for analysis.39 Some of the more popular software companies that offer products with varied pricing structure include, but are not limited to, Leica (Aperio), Visiopharm, Halo (Indica Labs), and Aiforia. The analytical abilities among the commercial packages are similar; however, the level of supervision that is allowable by the user, based on the adaptability of the platform, is often a vital factor that differentiates the packages and guides purchasing decisions. For example, Aperio and Halo offer unsupervised algorithm packages, meaning the algorithms’ functions are built into the software and only minor adjustments are allowed/required by the user. These software packages require minimal user training and are more useful for performing repetitive analyses. On the other hand, supervised packages such as those offered by Visiopharm and Aiforia require more upfront training to use and tend to be more expensive; however, users can program algorithms themselves, perform more complex analyses, and develop their own analyses.36,40

Aperio provides cloud-based and/or on-site use of software that allows for automated and quantitative evaluation of both BF and FL slides (the cloud hosting service will reportedly be discontinued in December 2021). Aperio has several applications that can be tailored for the user’s needs and include analysis of hematoxylin and eosin, cytochemical (including simple color deconvolution algorithms), and immunohistochemical stains (nuclear, cytoplasmic, and membrane staining); morphometry; pattern recognition; rare event detection; and microvessel detection. Analytics can be performed on tissue sections and on tissue microarrays.

Halo, as an unsupervised platform, offers predesigned modules and add-ons alleviating (and disallowing) the option to build modules from scratch. The available analytics are similar to Aperio’s offerings, but Halo reports having the fastest processing speeds available and unmatched ease of use and scalability. Halo appears to be ideal for high throughput workflows.

Visiopharm integrates deep learning into its image analysis tools, which allows algorithms to identify structures based on underlying patterns, using examples instead of code. As it relates to computer images, which come in an array of pixel values, deep learning allows computers to mimic the process of human visual perception. Deep learning discovers intricate structural detail in large data sets and has dramatically improved visual object recognition and detection.41,42 Visiopharm has all of Aperio’s capabilities, mentioned earlier, in addition to a library of ready-to-use applications that can be tailored to the user’s specifications; additionally, new algorithms can be designed from scratch. Visiopharm also offers optional add-ons, including programs for stereology, virtual double staining, and phenotyping.

Aiforia is a supervised, cloud-based platform that allows the most control of the packages mentioned and to design algorithms. The user is able (i.e., required) to develop the artificial intelligence modules that will be used; however, the company reports that no coding ability or even prior extensive experience with WSI is required. Additionally, in contrast to most supervised packages, the cost of Aiforia is highly competitive by comparison.

Digital pathology allows users to acquire, view, manage, and interpret pathology from a digitized glass slide using VM. Many instructors and students appreciate students’ development of their technical skills by using conventional glass slides and light microscopy. Subsequent to this acquisition of basic skills, however, a shift to digital teaching of veterinary pathology requires very little instruction and allows for continuity of learning in any environment with access to a sufficient Internet connection and computer workstation. Given the current trend of teaching histopathology using digitally scanned images of tissue sections, the inherent advantages over space and time, and the broad acceptance of a virtual platform by students, virtual microscopy is becoming the norm for enhancing the learning experience. The transition to automated image analysis in research presents a new role for the veterinary pathologist and allows for the production of quantitative, rather than descriptive, data. The veterinary pathologist will continue to serve as a valued team member (or leader) in the experimental design, including assisting with selecting the most relevant animal model, ensuring that analyses are correctly performed, and determining the associated pathologies that manifest in the animal.4345 However, to fully realize the advantages of DP, and prior to making DP available to instructors, students, and researchers, the importance of initial and sustained integration of IT in the process of purchasing products (hardware and software) and in determining the business model (from storage to management to delivery of images) that will be deployed cannot be stressed enough.

Finally, while image analysis is an invaluable tool in research (as well as education and diagnostics), several steps are needed to validate its use for each new project, and the limitations of this technology should not be overlooked. Some challenges and requirements to keep in mind include the following: the standardization of the algorithms used is needed to improve reproducibility; inherent limitations of most hardware (computer, camera, scanner, etc.) for processing and storing large files remain; the costs can be significant for implementation and maintenance of the DP workflow, including the costs of the software used for image analysis; questions related to the regulation of the analytical processes and components are yet to be answered; and algorithms that have been validated in one lab, using a particular software program, may not be reproducible in another lab for a variety of reasons, including basic differences in specimen handling, staining protocols, and glass slide production.46,47

We acknowledge and thank Drs. Gillian Beamer and Ramesh Vemulapalli for their critical review of this manuscript and constructive comments.

1. Bertram CA, Klopfleisch R. The pathologist 2.0: an update on digital pathology in veterinary medicine. Vet Pathol. 2017;54(5):75666. https://doi.org/10.1177/0300985817709888. MedlineGoogle Scholar
2. Aeffner F, Zarella MD, Buchbinder N, Bui MM, Goodman MR, Hartman DJ, et al. Introduction to digital image analysis in whole-slide imaging: a white paper from the Digital Pathology Association. J Pathol Inform. 2019;10:9. https://doi.org/10.4103/jpi.jpi_82_18. MedlineGoogle Scholar
3. Farahani N, Parwani AV, Pantanowitz L. Whole slide imaging in pathology: advantages, limitations, and emerging perspectives. Pathol Lab Med Int. 2015;7:2333. https://doi.org/10.2147/PLMI.S59826. Google Scholar
4. Cornish TC, Swapp RE, Kaplan KJ. Whole-slide imaging: routine pathologic diagnosis. Adv Anat Pathol. 2012;19(3):1529. https://doi.org/10.1097/PAP.0b013e318253459e. MedlineGoogle Scholar
5. Zarella MD, Bowman D, Aeffner F, Farahani N, Xthona A, Absar SF, et al. A practical guide to whole slide imaging: a white paper from the Digital Pathology Association. Arch Pathol Lab Med. 2018;143(2):22234. https://doi.org/10.5858/arpa.2018-0343-RA. MedlineGoogle Scholar
6. Abels E, Pantanowitz L, Aeffner F, Zarella MD, van der Laak J, Bui MM, et al. Computational pathology definitions, best practices, and recommendations for regulatory guidance: a white paper from the Digital Pathology Association. J Pathol. 2019;249(3):28694. https://doi.org/10.1002/path.5331. MedlineGoogle Scholar
7. Christian RJ, VanSandt M. Using dynamic virtual microscopy to train pathology residents during the pandemic: perspectives on pathology education in the age of COVID-19. Acad Pathol. 2021;8:23742895211006819. https://doi.org/10.1177/23742895211006819. Google Scholar
8. Sibel Sensu OT, Demirci M, Kutlu S, Genc BN, Ayatali S, Gurbuz YS, Erdogan N. Telepathology in medical education: integration of digital microscopy in distance pathology education during COVID-19 pandemic. J Med Sci Clin Res. 2020;8(11):53643. https://doi.org/10.18535/jmscr/v8i11.93. Google Scholar
9. Lujan GM, Savage J, Shana’ah A, Yearsley M, Thomas D, Allenby P, et al. Digital pathology initiatives and experience of a large academic institution during the Coronavirus Disease 2019 (COVID-19) pandemic. Arch Pathol Lab Med. Epub 2021 May 4. https://doi.org/10.5858/arpa.2020-0715-SA. MedlineGoogle Scholar
10. Hassell LA, Peterson J, Pantanowitz L. Pushed across the digital divide: COVID-19 accelerated pathology training onto a new digital learning curve. Acad Pathol. 2021;8:2374289521994240. https://doi.org/10.1177/2374289521994240. Google Scholar
11. Brunelli M, Beccari S, Colombari R, Gobbo S, Giobelli L, Pellegrini A, et al. iPathology cockpit diagnostic station: validation according to College of American Pathologists Pathology and Laboratory Quality Center recommendation at the Hospital Trust and University of Verona. Diagn Pathol. 2014;9(Suppl 1): S12. https://doi.org/10.1186/1746-1596-9-S1-S12. MedlineGoogle Scholar
12. Krupinski EA. Virtual slide telepathology workstation of the future: lessons learned from teleradiology. Hum Pathol. 2009;40(8):110011. https://doi.org/10.1016/j.humpath.2009. 04.011. MedlineGoogle Scholar
13. Humphreys BE. Demand for broadband in rural areas: implications for universal access [Internet]. Washington (DC): Congressional Research Service; 2019 Dec 9 [cited 2021 Aug 13]. Available from: https://fas.org/sgp/crs/misc/R46108.pdf. Google Scholar
14. Sahu M, Gnana Sundari K, David A. Recent ergonomic interventions and evaluations on laptop, smartphones and desktop computer users. In: Arockiarajan A, Duraiselvam M, Raju R, editors. Advances in industrial automation and smart manufacturing. Lecture notes in mechanical engineering. Springer, Singapore; 2021. p. 207–24. https://doi.org/10.1007/978-981-15-4739-3_18. Google Scholar
15. Entwisle G, Entwisle DR. The use of a digital computer as a teaching machine. J Med Educ. 1963;38:80312. MedlineGoogle Scholar
16. Durcan TM, Hinchcliffe EH. Digital image files in light microscopy. Methods Cell Biol. 2007;81:31533. https://doi.org/10.1016/S0091-679X(06)81015-0. MedlineGoogle Scholar
17. Furness PN. The use of digital images in pathology. J Pathol. 1997;183(3):25363. https://doi.org/10.1002/(SICI) 1096-9896(199711)183:3<253::AID-PATH927>3.0.CO;2-P. MedlineGoogle Scholar
18. Pantanowitz L. Digital images and the future of digital pathology. J Pathol Inform. 2010;1:15. https://doi.org/10.4103/2153-3539.68332. Google Scholar
19. Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F. Comparison of a virtual microscope laboratory to a regular microscope laboratory for teaching histology. Anat Rec. 2001;265(1): 104. https://doi.org/10.1002/ar.1036. MedlineGoogle Scholar
20. MacMillan FM, Barnes SD, Hamilton JC, Steane IS, Langtpm PD. The virtual microscope in histology teaching: evaluation of effectiveness and student preference. Proceedings Physiological Society; 2009; University College Dublin, Dublin, UK. Google Scholar
21. Gatumu MK, MacMillan FM, Langton PD, Headley PM, Harris JR. Evaluation of usage of virtual microscopy for the study of histology in the medical, dental, and veterinary undergraduate programs of a UK University. Anat Sci Educ. 2014;7(5):38998. https://doi.org/10.1002/ase.1425. MedlineGoogle Scholar
22. Haag M, Igel C, Fischer MR. Digital teaching and digital medicine: a national initiative is needed. GMS J Med Educ. 2018;35(3):Doc43. https://doi.org/10.3205/zma001189. MedlineGoogle Scholar
23. Kumar RK, Freeman B, Velan GM, De Permentier PJ. Integrating histology and histopathology teaching in practical classes using virtual slides. Anat Rec B New Anat. 2006;289(4):12833. https://doi.org/10.1002/ar.b.20105. MedlineGoogle Scholar
24. Solberg BL. Digital and traditional slides for teaching cellular morphology: a comparative analysis of learning outcomes. Clin Lab Sci. 2012;25(4 Suppl):412-18. Google Scholar
25. Kuo KH, Leo JM. Optical versus virtual microscope for medical education: a systematic review. Anat Sci Educ. 2019;12(6):67885. https://doi.org/10.1002/ase.1844. MedlineGoogle Scholar
26. Mea VD, Carbone A, Di Loreto C, Bueno G, De Paoli P, García-Rojo M, et al. Teaching digital pathology: the international school of digital pathology and proposed syllabus. J Pathol Inform. 2017;8:27. https://doi.org/10.4103/jpi.jpi_17_17. MedlineGoogle Scholar
27. Rodrigues-Fernandes CI, Speight PM, Khurram SA, Araújo ALD, da Cruz Perez DE, Fonseca FP, et al. The use of digital microscopy as a teaching method for human pathology: a systematic review. Virchows Arch. 2020;477(4):47586. https://doi.org/10.1007/s00428-020-02908-3. MedlineGoogle Scholar
28. Lin L, Johnson T. Shifting to digital: informing the rapid development, deployment, and future of teaching and learning: introduction. Educ Technol Res Dev. Epub 2021 Feb 2:15. https://doi.org/10.1007/s11423-021-09960-z. MedlineGoogle Scholar
29. Zwingenberger AL, Ward PR. Medical Imaging Resource Center (MIRC) for veterinary medicine: a digital image teaching file. J Vet Med Educ. 2006;33(4):61821. https://doi.org/10.3138/jvme.33.4.618. LinkGoogle Scholar
30. Brown PJ, Fews D, Bell NJ. Teaching veterinary histopathology: a comparison of microscopy and digital slides. J Vet Med Educ. 2016;43(1):1320. https://doi.org/10.3138/jvme.0315-035R1. LinkGoogle Scholar
31. Bertram CA, Firsching T, Klopfleisch R. Virtual microscopy in histopathology training: changing student attitudes in 3 successive academic years. J Vet Med Educ. 2018;45(2):2419. https://doi.org/10.3138/jvme.1216-194r1. LinkGoogle Scholar
32. Evans AJ, Depeiza N, Allen SG, Fraser K, Shirley S, Chetty R. Use of whole slide imaging (WSI) for distance teaching. J Clin Pathol. 2021;74(7):4258. https://doi.org/10.1136/jclinpath- 2020-206763. MedlineGoogle Scholar
33. Stewart SM, Dowers KL, Cerda JR, Schoenfeld-Tacher RM, Kogan LR. Microscope use in clinical veterinary practice and potential implications for veterinary school curricula. J Vet Med Educ. 2014;41(4):3316. https://doi.org/10.3138/jvme.0614-063R. LinkGoogle Scholar
34. Scopio [Internet]. Tel Aviv: Scopio Labs; [cited 2021 Jun 15]. Available from: https://scopiolabs.com/. Google Scholar
35. Rajewski G. Using advances in digital pathology to transform teaching and research. Tufts Now [Internet]. 2021 Jan 19 [cited 2021 Jun 15]. Available from: https://now.tufts.edu/articles/ using-advances-digital-pathology-transform-teaching-and-research. Google Scholar
36. Webster JD, Dunstan RW. Whole-slide imaging and automated image analysis: considerations and opportunities in the practice of pathology. Vet Pathol. 2014;51(1):21123. https://doi.org/10.1177/0300985813503570. MedlineGoogle Scholar
37. Potts SJ, Young GD, Voelker FA. The role and impact of quantitative discovery pathology. Drug Discov Today. 2010;15(21–2):94350. https://doi.org/10.1016/j.drudis.2010.09.001. MedlineGoogle Scholar
38. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9(7):6715. https://doi.org/10.1038/nmeth.2089. MedlineGoogle Scholar
39. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9(7):67682. https://doi.org/10.1038/nmeth.2019. MedlineGoogle Scholar
40. Saravanan C, Schumacher V, Brown D, Dunstan R, Galarneau J-R, Odin M, Mishra S. Meeting report: tissue-based image analysis. Toxicol Pathol. 2017;45(7):9831003. https://doi.org/10.1177/0192623317737468. MedlineGoogle Scholar
41. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521(7553):43644. https://doi.org/10.1038/nature14539. MedlineGoogle Scholar
42. Niazi MKK, Parwani AV, Gurcan MN. Digital pathology and artificial intelligence. Lancet Oncol. 2019;20(5):E25361. https://doi.org/10.1016/S1470-2045(19)30154-8. MedlineGoogle Scholar
43. Ferreira GS, Veening-Griffioen DH, Boon WPC, Moors EHM, Gispen-de Wied CC, Schellekens H, van Meer PJK. A standardised framework to identify optimal animal models for efficacy assessment in drug development. Plos One. 2019;14(7):e0218014. https://doi.org/10.1371/journal.pone.0218014. Google Scholar
44. Denayer T, Stöhr T, Van Roy M. Animal models in translational medicine: validation and prediction. New Hor Transl Med. 2014;2(1):511. https://doi.org/10.1016/j.nhtm.2014.08.001. Google Scholar
45. Knoblaugh SE, Hohl TM, La Perle KMD. Pathology principles and practices for analysis of animal models. ILAR J. 2018;59(1):4050. https://doi.org/10.1093/ilar/ilz001. MedlineGoogle Scholar
46. Acs B, Hartman J. Next generation pathology: artificial intelligence enhances histopathology practice. J Pathol. 2020;250(1): 78. https://doi.org/10.1002/path.5343. MedlineGoogle Scholar
47. Colling R, Pitman H, Oien K, et al. Artificial intelligence in digital pathology: a roadmap to routine use in clinical practice. J Pathol. 2019;249(2):14350. https://doi.org/10.1002/path.5310. MedlineGoogle Scholar
48. Jones-Hall Y. Digital pathology in academia: implementation and impact. Lab Anim. Epub 2021 Aug 4. https://doi.org/10.1038/s41684-021-00828-6. Google Scholar