The Leeds Guide to Digital Pathology

The Leeds Guide

to Digital Pathology

TABLE OF CONTENTS

Introduction ...................................................................................................................................................... 1

Deﬁning Digital Pathology ............................................................................................................................... 2

The Evidence Base and Case for Adoption ...................................................................................................... 2

The Beneﬁts of Digital Pathology .................................................................................................................... 3

A Business Case for Digital Pathology ............................................................................................................ 4

Mapping Your Path to Successful Deployment ............................................................................................... 5

Scanner Procurement and Operation ............................................................................................................... 8

Engagement with Colleagues ........................................................................................................................ 11

Validation and Training .................................................................................................................................. 13

IT Considerations and Systems Integration ................................................................................................... 14

The Pathologist’s Workstation ....................................................................................................................... 16

Continuing the Journey .................................................................................................................................. 17

Final Thoughts ................................................................................................................................................ 18

INTRODUCTION

The digital pathology team at Leeds Teaching Hospitals

NHS Trust and the University of Leeds has 15 years’

experience and a proven track record in delivering world

class digital pathology innovation and research.

The path to clinical digital pathology deployment is a challenging

but rewarding one. Leeds Teaching Hospitals NHS Trust

continues the journey and has gained valuable insight that will

help others navigate their way. The purpose of this guide is to

provide balanced and practical information on clinical deployment based on our experience and knowledge.

Leeds fully subspecialised diagnostic department of 45 consultant pathologists generates more than 290,000 H&E

slides per annum. We have a multi-award winning digital pathology team comprised of clinicians, scientists, and IT/

informatics professionals. Following successful pilot work in two key subspecialty areas, we are transitioning to 100%

digital scanning of glass slides; creating standard work for training and validating our staff.

Healthcare systems face a global shortage of pathologists, while diagnostic workloads and complexities continue to

increase. Hence, there is a pressing need to innovate and modernise the way pathology services are delivered.

The time is right for mainstream clinical adoption of digital pathology, which has the potential to add value and

completely transform pathology services. Technological maturity of digital pathology hardware and software, coupled

with growing acceptance by pathologists, as well as regulatory, and governing bodies, have aided this progression.

In addition, the evidence base for the validity of digital diagnosis has evolved, and we understand more about how to

use digital slides efﬁciently and safely.

At Leeds, we are developing best practice guidance to share with other organizations interested in implementing

digital pathology for routine diagnosis. We are champions of evidence-based medicine and have sought to apply

a rational approach to a rapidly evolving medical discipline, while focusing on patient safety and the maintenance

of professional standards. We hope you ﬁnd this guide a useful starting point when embarking on your own digital

pathology journey!

– The Leeds Digital Pathology Team

“ We want to see digital

pathology go mainstream

for the beneﬁt of patients.”

Darren Treanor, Consultant Pathologist and

Lead for the Digital Pathology Project at Leeds

Teaching Hospitals NHS Trust, UK

Dr Bethany Williams, Digital Pathology Fellow, using the Powerwall

Deﬁning Digital Pathology

Digital pathology (DP) includes three imaging technologies:

1. Gross imaging

The acquisition and storage of macroscopic pathological images to aid histological diagnosis and cancer dataset

completion.

2. Telepathology

The predecessor of modern whole slide imaging systems, telepathology involves the live remote control of a modiﬁed

microscope, and the image is not stored on a computer system.

3. Whole Slide Imaging

When people talk about DP today, they are usually referring to whole slide imaging (WSI). Also known as “virtual

microscopy,” WSI is normally performed with a highly

sophisticated, dedicated scanner that is used to capture

a digital image of the whole slide for later review at

a remote location, which now, thanks to DP, can be

anywhere in the world.

A WSI or “virtual slide” is an image created by the

complete digitisation of the glass slide at up to 200,000

dots per inch (DPI). The resulting images are up to 10

gigapixels in size. If printed at a standard 300 DPI, they

would be the same size as a tennis court. Fortunately,

specialised software is available to compress the image so it is reduced to a negotiable and optimal size for viewing

and analysis. Modern WSI scanners can scan from 1 to 400 slides at a time, typically using a microscope lens with

magniﬁcation of 20x or 40x. For some applications, there are systems available that can offer higher magniﬁcation

than 40x, if needed.

The Evidence Base and Case for Adoption

The Leeds diagnostic pathology department is a typical NHS department. We face similar issues to those encountered

by pathology services worldwide. We believe DP can provide a ﬂexible platform for safety, quality, and efﬁciency

improvement. We are equally enticed by the prospect of future-prooﬁng our diagnostic capabilities, whilst allowing us

to work more ﬂexibly and creatively to meet targets and maintain standards.

Key Facts: The Strategic Context of Digital Pathology

• In the UK, cellular pathology requests are increasing by 4.5% year on year. (3)

• Specimens require ever more complex assessment to meet the requirements of national datasets.

• In the UK, we face increasingly challenging cancer turnaround time targets.

• 32% of UK cellular pathologists are expected to retire within the next ﬁve years. (3)

Note: These trends are reﬂected globally. Clearly, pathology departments face an international recruitment and

retention shortfall.

At the onset of this project, we completed a systematic review of existing literature on DP accuracy. We identiﬁed

and synthesised the data from 38 peer-reviewed publications spanning the last 20 years. We found a mean diagnostic

concordance of glass slide microscopy diagnosis and digital microscopy diagnosis of 92.4%. (1), compared with a

concordance rate of 93.7% for repeat microscope review of a case using glass slides. We went on to analyse, in

detail, the relatively small number of discordant cases, to identify the key areas of difﬁculty that pathologists can

experience when they make the transition from glass slide to digital slide reporting. (2) Knowledge of these potential

difﬁculties allowed us to plan our training and validation procedures, whilst maintaining a strong focus on patient safety.

The Beneﬁts of Digital Pathology

The principal beneﬁts can be divided into four categories:

1. Improvements in Patient Safety

• Use of an integrated DP system allows paperless transmission of digital slides directly to the pathologist,

decreasing the possibility of misidentiﬁcation or transposition errors at multiple points along the diagnostic

workﬂow.

• Digital slides offer a readily portable, instantaneously transmissible diagnostic image, which is not subject to the

physical limitations and fragility of glass slide transport.

2. Advances in Workﬂow

• Flexibility– WSI presents the ability to manipulate workload allocations by pushing and pulling of cases. This allows

departments to proactively respond to ﬂuctuations in workload or case mix allowing for maximized use of resources.

• Rapid case tracking, archiving, and retrieval of slide images and diagnostic information.

• Fast case transfer times between the laboratory and assigned pathologists, resulting in streamlined turnaround

times and deﬁned diagnostic pathways.

3. Positive Impact on the Laboratory Workforce

• The innate ﬂexibility of the digital slide process offers the potential for diverse and appealing working

conditions. WSI also offers the potential for a better balance in individual workloads. Pressure can

be reduced and accuracy increased with the ease and assurance of remote corroboration of any diagnosis.

• Potential to customise and optimise contributions by the workforce will allow those that work less

than full time to maximise their on-the-job productivity. Those considering retirement can choose to offer

their services on more ﬂexible terms, which may beneﬁt individuals and organizations.

• Making “work-life balance” more attainable is likely to appeal to the next generation of pathologists

and drive recruitment of medical graduates to the specialty.

4. Service Quality

• WSI offers the potential for faster access to diagnosis and decreased turnaround times, offering patients access to

their results sooner.

• Improved information sharing and collaboration (e.g. streamlined double-reporting and rapid access to second opin-

ion and expert review) may lead to enhanced diagnostic conﬁdence with a higher quality of diagnosis.

• Convenient recording of cancer staging parameters, including measurements, to improve the outcomes and repro-

ducibility of cancer dataset reporting.

A Business Case for

Digital Pathology

Developing an effective business case for

deployment of DP can be one of the most daunting

tasks a department can face. Your ﬁ rst priority

should be to deﬁ ne your “use cases” – how users

will interact with the technology to achieve speciﬁ c

goals. We recommend a phased deployment to

allow for a gradual integration and accrual of skills

in both the diagnostic laboratory and the reporting

desk. Our phased approach began with primary

diagnosis of breast histology, neurohistology, and

immunohistochemistry assessment.

You may want to consider the relative merits of the

following use cases within your institution:

Use Case Description

Use of digital pathology for primary

diagnosis of pathological specimens

Replacement of light microscopic examination of glass slides with

examination of whole slide images as part of the standard diagnostic

workﬂ ow. You may initially want to explore primary diagnosis for a limited

subset of cases ,e.g. all gynecological pathology cases, or for the work-

load of an individual pathologist or pathologist groups.

Assessment of immunohistochemistry

(IHC)

Replacement of light microscopic examination of glass slides with digital

slides to assess immunohistochemical stains.

Multi-disciplinary team (MDT)/

tumour board

The selection, collection, review and presentation of whole slide images

or annotated regions of interest of cases for discussion at multidisciplinary

meetings or tumour boards. If this use case is taken in isolation, slides can

be scanned after initial conventional glass slide diagnosis, then reviewed

and presented digitally.

Frozen section diagnosis The use of whole slide images to provide rapid, intraoperative

histopathological opinion. You may be considering this option if you need

to provide remote frozen section support to a partnered institution or need

to utilise specialised pathologists working off site.

Receiving and requesting second

opinions/review cases

The use of a DP system to render or request a second opinion on a

previously examined case, e.g. a difﬁ cult skin case from a general

pathologist to a dermatopathologist.

Remote diagnosis The use of DP to allow pathologists to view and report slides from off-site

locations, including other networked hospitals and their home ofﬁ ce.

Insourcing/outsourcing of

diagnostic work

The use of DP to allow the movement and exchange of units of diagnostic

work among different pathology service providers.

Improving

Patient

Safety

Improving

Efﬁ ciency

Optimising

the

Workforce

Improving

Service

Quality

Best Practice for Preparing a Business Case

• Begin with a clear, concise introduction, in which you describe your current laboratory and diagnostic practices and

your desire to transition to DP. Use simple terms that do not require expert knowledge of pathology processes. You

will also need to summarize the pressing reasons for why your department should invest in DP. Make use of the data

and references from the “Case for Adoption” paper referenced in the “Further Resources” section to add detail on

the strategic context of DP. (4)

• Describe the scope of investment by listing all the components and itemising the estimated cost of your proposed

DP systems (scanners, servers, slide management software, interfaces, storage)

• Outline how DP links to regional, national, and international strategy:

» Regional/local – reinforce your business case with demographics, data on workforce characteristics, workload

volume and composition, as well as imminent shortfalls in stafﬁng or expertise.

» National – what initiatives, policies or guidelines would your DP implementation help support?

» International – is the prospect of worldwide access and/or rolling time zone access of potential beneﬁt to your

department? If so, build this into your strategic context section.

• Create the economic case for DP adoption – you might want to include an options assessment here, evaluating the

proposition that you “do nothing” versus investing in DP. These options could be scored on the criteria of service

sustainability, affordability, quality, beneﬁts, and risks.

• Your ﬁnancial case needs to summarise potential cash-releasing savings and cost-avoidance beneﬁts of DP, in

addition to any proposed income to be derived from adoption of whole slide digital imaging. For more on this

subject, refer to our paper and business case template. (6)

Mapping Your Path to Successful Deployment

Once the decision has been made to go digital, the emphasis will be on creating a detailed plan for effective

deployment. We believed that implementing DP reporting in a phased-approach, starting with the self-declared,

enthusiastic subspecialties, would result in success from the start. We conducted a survey of our pathologists to assess

the level of buy-in across the department. The results showed differing levels of conﬁdence in the potential usage of

DP for routine task, multidisciplinary team (MDT) meetings, Immunohistochemistry (IHC) review, second opinions, and

primary diagnosis. Pathologists from the renal, GI/liver, neuro, skin, and breast subspecialties indicated the strongest

level of support.

Not at all interested Uninterested Neutral Somewhat interested Very interested

0 1 4 11 25

0% 2% 10% 27% 61%

Not at

all likely

Unlikely Neutral Likely

Very

likely

MDT review 0 2 3 5 21

IHC review 0 2 4 5 17

Second opinion 2 3 4 7 15

Primary diagnosis 3 6 6 7 9

Effective Project Management

For implementation to proceed smoothly and effectively, it is essential that multiple contributing components are

managed and well-planned, and we believed that we would minimize our risk of project delays by hiring a dedicated,

professional project manager.

Our project manager has responsibility for DP deployment including management of many subprojects that

synergistically contribute to total success. Our subprojects were led by various members of the DP team, and the

project manager played a pivotal role in coordinating efforts, mapping out the steps of each phase, prioritising tasks,

and, overall, being responsible for keeping deployment on track and on time. Your project manager will be equally

charged with these responsibilities and opportunities. It’s important to note that all of the DP team members have

fulltime jobs outside of the DP project.

Required Resources

To gain support and initiate change both inside and outside of the lab, we created a cross-functional team to drive our

DP deployment.

Our DP project team members:

• DP champion at the executive level (at Leeds, we chose the Medical Director)

• Clinical lead for DP

• Project manager

• Training and validation lead

• Laboratory lead

• Pathology IT lead

• Business analyst/lean engineer

• Informatics representation (network/storage/support)

• Corporate service representation (ﬁnance/HR/communications)

• Vendor lead/representation

At Leeds, we have successfully employed a combination of regularly scheduled meetings and emails, plus posters,

newsletters and presentations, to provide transparency on rollout progress across the whole department. Maintaining

good relationships through ongoing communications with DP vendors and laboratory information system service

providers play a pivotal role in positive project outcomes.

Workﬂow Analysis and Process Mapping Techniques

To create a baseline of our workﬂow, we started with the end-to-end mapping of our process. Our analysis showed

signiﬁcant opportunity to improve overall turnaround time for diagnostic cases, as well as a reduction in the number

of touchpoints (and sources of error) along the way.

Below are our top tips for process mapping:

• Document your “current” lab workﬂow to create a value stream map.

• Understand and plot laboratory pressure points, bottlenecks, and waiting time.

• Create your “future” value stream map (VSM):

» This map should be your “idealised process” with as much waste removed as possible, and a focus on reducing

turnaround time by removing delays and bottlenecks.

» Your future state should include any additional steps required for slide scanning and should be created when

there is a reasonably clear understanding of the IT architecture and technical solutions.

• Develop an incremental improvement plan that will help you move to the future state.

• Using the concepts of “Lean Engineering” and a “Culture of Continuous Improvement”, the project manager will

create a detailed project plan to drive this change. This will facilitate concurrent execution of tasks in different work

areas, and the chance to identify dependencies needing sequential execution of deliverables.

Leeds VSM Comparison

Tips for an Effective Deployment

• Some changes may require capital investment, but application of appropriate improvement methodologies (e.g.

Lean and Six Sigma) can produce signiﬁcant improvements at a low cost.

• T o ensure input on the current state and buy-in on the future state, we recommend holding a two-day meeting

with key personnel involved in each step of the workﬂow process. Spend time observing the current workﬂow,

measuring both hands-on time and waiting time.

• T rust in analytics. Remember, if it’s not measured, it can’t be improved!

Scanner Procurement and Operation

Choosing the right scanner is a key consideration when implementing DP. Attention should be given to your sample

type, size, volume and throughput requirements and, of course, the capability of the scanner to meet or exceed the

requirements.

What type of scanners do we need?

We based our scanner decision on our total volume of work including the total amount of time it would take to

produce a slide in terms of effort from both staff and machinery.

How many scanners do we need?

To procure the correct number of scanners, we calculated the total scanning time required, total scanning time

available, as well as the expected utilisation of the scanners. For the total required scan time, we multiplied our

volume of slides by the average time taken to scan a slide. It is important to remember that scanners cannot

scan without interruption; time needs to be allowed for loading and unloading slides, maintenance (planned and

unplanned), as well as for ﬂuctuations in workﬂow in the rest of the lab.

Based on these factors, we estimated that our scanners would be utilised 70% of the available time, allowing us to

calculate the actual required scan time:

Total Required Scan Time = Actual Required Scan Time

% Utilisation

In our department, we decided that the scanners could run unattended overnight during the week but would not be

used during the weekends, offering us 120 operating hours per week. Other options would have been to let scanners

run only when staff were available to attend to them or to let them run 24/7.

Actual Required Scan Time = Number of scanners required

Operating Hours Required

We calculated the following as our scanner requirements:

Project phase Slides to scan per day (approx.) Scanners

Breast pilot 150 One AT2 and One CS2

Breast plus IHC 300 Two AT2 and One CS2

Full digital 1200 Six AT2 and Three CS2

To aid turnaround times, a good rule of thumb is to scan urgent work during the day and save less, time-critical

work for overnight runs. We determined that we needed two types of scanners; a high and a low capacity scanner

each for different use cases. We chose the Aperio-branded scanners from Leica Biosystems. The high capacity

Aperio AT2 (400 slides) was fed with small batches of urgent breast biopsies throughout the day, with lower priority

resection work collected throughout the day for an overnight run. The lower capacity Aperio CS2 (5 standard slides

or 2 large slides), shared the burden of urgent biopsies throughout the day but also proved invaluable for the

scanning of large format “mega-block” (2”x3”) slides, which our breast pathologists use to demonstrate multiple

margins and tumour dimensions.

How many staff members will we need to operate the scanners?

There were two distinct workﬂow steps to consider when we were calculating how many personnel hours would be

required to run the scanners: 1) the slide scanning process and 2) slide quality control.

Once again, we used the average weekly volumes and mapped the overall process. This process can vary from lab to

lab, but for us, it broadly consisted of loading and unloading the scanners, taking snapshots to ensure that all tissue

from a slide was captured and then quality- checking the scanned images, (plus any required data input). Multiplying

the total hands-on time of this process by the volume gave us a ﬁgure for the total required human resources. This

information allowed us to consider how to staff the scanners, bearing in mind operating hours, likely ﬂow of work

throughout the day, and whether we would have a dedicated scanning team or not. Note that the required FTE (the

hours worked by one employee on a full-time basis) will be reduced as the process matures.

Our calculations were as follows:

Project phase Full time equivalent staff required to operate scanners

Breast pilot 0.4

Breast plus IHC 0.8

Breast plus IHC plus all other specialties

(i.e. full lab digitisation)

2.2

Personnel assigned to work on the scanners need to be fully trained in the tasks that they are asked to do, with

appropriate standard operating procedures, supervision and sign off processes. In this way, scanner operations should

be treated like all other lab processes.

Where should the scanners be located?

Take time to consider where you are going to locate your

scanners. One of the key tenets of the Leeds approach to

laboratory digitisation was that the scanners should be an

integral part of normal laboratory function, and as such,

should be sited in the laboratory. Unlike other technology

systems, the Aperio scanners from Leica Biosystems did

not require specialised benches or necessitate location in a

separate room due to noise or vibration, making it possible to

install them virtually anywhere in the laboratory. We looked

at all potential areas within the laboratory and ranked them

on criteria we thought were important using a trafﬁc light

system: green for “go”, yellow/amber for “caution”, and red

for “stop” or “no go”. Option 4 (see below) was chosen as

the optimal site. And fortunately, given the relatively small

footprint of the scanners, we could accommodate them

within our existing lab space.

Adam Stocks, Biomedical Support Worker

Option

Centralised

Approach

Walk Around

Time

Existing

Infrastucture

Suitable

Bench/

Building Work

Supports

Existing

Lab Workﬂow

Scanner

Maintenance

Option 1 – Seniors Room

Option 2 – Wash Room

Option 3 – Temporary Slide Storage

Option 4 – Back Wall

Option 5 – Equipment Storage Room

Option 6 – Slide Storage Room

Option 7 – Spare Bench

Option 8 – IHC Area

Option 9 – IHC QA Area

Option 10 – Addiitional Work Area

How to Save Time if You’re Running Standard and Digital Reporting Simultaneously

If you decide to implement DP in a phased approach, you will inevitably encounter a stage where you have both

digital and glass-based processes running in parallel. Using both processes can potentially add steps and time to your

laboratory workﬂow. One of the keys to keeping wasted time to a minimum is to maximise your use of the scanners

during the working day.

Tips for Better Deployment

• You should not design your complete system to handle a single, high-volume day; you need to design for reliable

performance day in and day out. The best way to do this is to calculate your capacity requirements using weekly

volumes. This will allow you to balance workloads across the entire week.

• Per lean principles, scanners need to be loaded throughout the day to maximise utilisation. You may want to change

“feeding times” – the timing of how slides are produced throughout the day. Rather than producing a large batch of

slides at the end of the day, you may need to move to a staggered production of slides, or even better, an even ﬂow

of production. To do this, you will need to make changes and improvements to your process.

• High quality images require high quality glass slides. If you have persistent issues with folding of tissue or thick

sections, these will need to be resolved or the requirement for rescans will be high.

• Treat scanning as an integral part of the slide production process. Staff need to be rostered onto it as much as any

other step in the process. If scanning is treated as an “add on”, it will most likely lead to delays in images being

produced, slowing down the whole diagnostic pathway.

• Consider whether a staggered or “big bang” roll out is most appropriate for your laboratory. In larger laboratories,

gradually scaling up the process on a specialty-by-specialty basis can lower operational risk by allowing unforeseen

issues to be ironed out, whilst volumes are low. However, these risks should be balanced against the difﬁculty of

running parallel processes.

• Engage your lab staff in the change process as they are a valuable source of innovative and constructive ideas.

• Make sure your feedback loops are working well. Listen to pathologist and lab staff input and speak to the

departments that your lab supports. Keep all stakeholders informed of changes and welcome suggestions. Do not

be reluctant to ask others to make changes; it may be that a small change in their process will help you to provide

better service.

• Stay positive! This is a big change that initially requires an investment of effort when ﬁ rst introduced, whilst glass

slides continue to be shipped to pathologists. The demand on your resources will start to ease up once pathologists

are validated, and your lab gets used to the new way of working. Processes such as ﬁ nding slides and cases will

become more straightforward than before, and your lab will have an easier job ﬁ ling cases.

Engagement with Colleagues

Generating and maintaining enthusiasm and cooperation in the laboratory and the wider department is essential to

a successful DP deployment. You will be asking your colleagues to change some of the most fundamental aspects of

their working practice. It is likely that you will confront a wide range of feelings and opinions, particularly in the early

stages of your project. You will encounter those who greet the prospect of change enthusiastically, counterbalanced

by the more reticent and ﬁ nally, those who are outright resistant. We have identiﬁ ed strategies to help motivate your

entire team – department by department – throughout the deployment process:

A. Improving Pathologist Engagement

1. Your department is likely to contain three groups of pathologists: the enthusiasts, the uncertains and the sceptics.

Recognising and identifying the key personnel in your department, who fall into each of the categories, will help

you target and create engagement, communication, educational materials, and activities that can help shift

attitudes. During the earliest stages of deployment, circulate a survey to your pathologists regarding current

attitudes and beliefs surrounding DP. A well-designed survey will capture data on readiness to go digital and allow

pathologists to express any underlying fears or uncertainties regarding digital reporting.

2. If you are planning a phased deployment, or an initial

pilot project, it may be beneﬁ cial to trial your DP system

on a smaller group of participants from each subspeciality.

Some of the key characteristics of the three groups are

described as follows:

• The “enthusiasts” or “cheerleaders” are the early adopters;

their enthusiasm and positivity can help motivate those who

are uncertain. Since cheerleaders can turn those sitting in the

stands into fans, their positive feedback on the technology

should be shared with the rest of the department.

• At the other end of the spectrum, your “sceptics” deserve

equal if not more attention. The sceptics will present

you and your colleagues with a long list of reasons why

your deployment will not work; some of these may be

genuine concerns that require addressing. It’s important to

view pathologist sceptics not as adversaries but as vital

contributors who can help you plan for a better deployment. Conversely, enthusiasts are appreciated and valued,

but often they are so in favour of deployment that they will overlook weak points and potential problems.

Enthusiasts

ScepticsUncertains

• “Uncertains” are the hardest to deal with; many will resist attempts to pull them into deployment plans. Arranging

informal visits and one-on-one discussions with enthusiastic pathologists who already work successfully with DP can

help assuage those who are apprehensive and help advance their appreciation and comfort level with the technology.

3. Have a representative of the digital deployment team attend all relevant departmental and pathologist meetings to

provide updates on the deployment planning and progress.

4. Provide email updates to the pathologists, even on the small successes, as this can lend credence and conﬁdence

to the DP deployment (e.g. “This week, Dr. X reported 100 skin biopsies digitally and presented his/her ﬁrst digital

MDT. Clinicians were impressed with the quality of the images and look forward to receiving more digital histology

updates and other examples of the technology at work.” We found that positive messages from “normal” users

were more powerful than those from the project team.

5. Log any workﬂow, performance, or diagnostic issues and respond to them quickly. Identify key personnel in the

department who can troubleshoot issues such as image quality, workstation set up difﬁculties, workﬂow issues,

etc. Ensure that everyone knows who to contact, the method of contact, and the best time to communicate to get

issues resolved.

B. Engaging Laboratory Personnel

1. In the early stages of deployment, seek opportunities to discuss your vision of DP to laboratory personnel and

allow staff to provide feedback, express concerns, and make suggestions for laboratory workﬂow improvements.

We found that this is best achieved with distribution of a brief survey, followed by targeted focus groups and/or

discussions with key individuals – the movers, shakers, and shapers of opinion in the lab.

2. DP should be presented as an integral part of the laboratory process and not an optional “add-on.” The laboratory’s

role in the diagnostic workﬂow is not complete until the glass slides have been scanned and sent out to the

pathologist diagnosticians. Cementing the idea that DP is going to be part of standard departmental practice

is important in encouraging staff to make the effort necessary to learn new skills and ultimately embrace the

advantages that digital reporting offers. Laboratory managers are vital here.

3. Seek every available opportunity to educate and explain the technology and plans for deployment. Take advantage

of informal lunchtime or coffee break meetings. Promote the idea of discussions as educational content and

continuing professional development activity.

4. Identify enthusiastic supporters who might be willing to share their knowledge and belief in DP with others who are

less on board. With their help, you can implement and sustain change in the laboratory.

C. Engaging with IT and Informatics

1. It is critical to engage with your IT and informatics departments and secure their involvement, at as high a level

as possible, from the earliest stages of your deployment. The buy-in and expert advice of these professionals is

fundamental to a successful DP deployment.

2. It is important that all stakeholders, whether pathologists, laboratory personnel, or members of the IT/Informatics

group, can communicate about the project in clear language. Setting up early sessions with key team members

to exchange and explain information from diagnostic, laboratory and IT perspectives should be one of your ﬁrst

priorities. Try to come to a group understanding of what is expected and what is potentially achievable from your

deployment, and what each professional group will be expected to contribute in terms of time and staff.

3. Early opportunities for engagement may include laboratory ofﬁce tours for your IT and informatics teams. Let

them see for themselves the pressure points of your current conventional glass slide workﬂows and grasp the

opportunities that digital can present. Explain ideas that you have for future digital workﬂows and see what

potential problems and solutions your IT colleagues can generate.

Validation and Training

A well developed and appropriately resourced training and validation programme for your pathologists leads to

numerous beneﬁ ts for your DP deployment. Giving your pathologists a risk-controlled period of digital diagnosis,

when they can familiarise themselves with the digital system and with the appearance of digital slides, helps them

gain conﬁ dence in the system and their individual ability to use it. It also provides valuable material for continuing

professional development targets, validation, and inspection cycles. We have a full sub-specialised pathology

department, and as such, we decided to validate, sequentially, by subspecialty. Insights from this approach have

shown that our overall approach to validation can be used across all histopathology subspecialties, but each has its

own speciﬁ c nuances that need time and attention to train and validate safely.

We developed an innovative validation and training protocol for primary diagnosis of histological specimens using DP.

Our protocol is designed to make pragmatic use of available resources, maintain a strong focus on patient safety, and

champion professional engagement and education.

Unlike previously published validations, our validation protocol is an individual validation for a single pathologist

and is evidence-based where possible. The protocol enables

self-identiﬁ cation of digital diagnostic pitfalls and allows

the pathologist to gain competence and conﬁ dence in a risk-

mitigated environment with early exposure to live digital

reporting. It avoids the need to do a full crossover study

comparing light microscope and digital diagnoses, as the

pathologist does the digital read from the outset (and checks

it on the microscope until they are fully conﬁ dent with digital).

The full validation protocol, and sample training materials are

available in the Royal College of Pathologists’ guideline for DP

as an example of best practice. (7)

Our protocol incorporates formalised training in use of the

digital microscope and observed practice with feedback. The

pathologist is presented with a test set of challenging and informative digital cases with immediate digital glass

reconciliation. We then scan all of a

consultant’s cases prospectively and

ask our pathologists to make their

diagnoses digitally, with immediate

glass reconciliation before sign out,

and adjustment of diagnosis,

if necessary. At the end of the process,

we produce a validation document

for each pathologist, documenting

training and concordance levels with

recommendations for scope of their

digital practice.

By providing an individual validation

for each pathologist, we allow them to decide which cases they are conﬁ dent to diagnose digitally and which need

more practice or workﬂ ow modiﬁ cations to ensure a conﬁ dent and safe diagnosis.

Pragmatic

Professional

engagement and

education

Patient

Safety Focused

After approximately two months of whole-time-equivalent reporting on digital slides (with glass slide reconciliation

prior to sign out), the concordance rates of the pathologist are reviewed and any problematic areas discussed. At this

point, a mutual decision is made between the pathologist and their trainer on whether the pathologist is either:

a) validated for full digital practice in the speciﬁed area of pathology

b) validated for digital practice in the speciﬁed area with some exceptions (requiring glass slide checks for certain

diagnostic scenarios)

c) not currently validated for digital diagnosis in this area of pathology

Tips for Validation

• Validate on an individual rather than a departmental level; let every pathologist have the opportunity to evaluate

their use of the technology and make their personal journey to conﬁdent and safe digital reporting.

• Make your validation procedure relevant for your own department. Tailor your training to ensure that your

pathologists view specimens, stains, and diagnoses that are relevant to their regular workloads, including some

harder cases and cases that are potentially challenging on digital, to direct training.

IT Considerations and Systems Integration

For any pathology lab to fully realise the efﬁciencies of going digital, a number of systems need to work in concert.

This includes the laboratory information system (LIS), specimen tracking system, slide management software, as

well as the slide viewer. Where required, uni-directional or bi-directional interfaces need to be put in place to ensure

a smooth transition of data between systems. The seamless integration of these systems will enhance the digital

experience and ease the workﬂow for laboratory staff and pathologists. Orchestrating system harmony will also

improve engagement of personnel and, perhaps most importantly, the speed at which a diagnosis is received by

anxious patients and their families.

At Leeds, we have chosen to make the LIS the master system with all other systems (i.e. tracking, slide archive and

viewing platform) linked back to the LIS via coded interfaces or links, which launch other speciﬁc applications.

The Beneﬁts of a Fully Functional Integration:

• Decreased need for pathologists and laboratory staff to manually enter data into various systems numerous times

• Reduced time to open several different software packages

• Better tracking of any case at any given time

• Automatic creation of audit trail for cases

• Faster notiﬁcation of case availability

• Simpliﬁed workﬂow for pathologists

Key Facts Regarding Leeds Breast Pathology Validation (8)

• Three consultant breast pathologists viewed 694 complete breast histology cases consisting of 3500 slides.

• All standard size slides were scanned at 40x equivalent magniﬁcation, with large slides scanned at 20x.

All slides were viewed on 6MP medical grade displays.

• Complete clinical concordance between digital and glass slide readings of cases was observed in 98.8%

of cases.

It’s important to include funding for system integration when creating your business case. Also, ensure that adequate

time is built into the project plan to allow for full testing of integration interfaces. Testing will help eliminate the

discovery of costly and time-consuming issues once the system is live.

The Value of a Digital Barcode Tracking System

In laboratories where a high quantity of slides are being scanned, it is essential that slides are barcoded and a

suitable slide tracking system is in place. Barcodes enable case and slide information to be held on a slide label. This

allows slides to be scanned in any order, on any scanning instrument, and for cases to be automatically organized

and managed, without the need for human intervention. Barcodes help to improve patient safety by reducing the

possibility of sample mismatching and alerts when slides go missing – something that is much more difﬁcult to track

with manual systems.

Additional Advantages of a Barcode System:

• Reduced need for manual data entry during scanning process

• Enhanced safety for the whole laboratory with reduced likelihood of misidentiﬁcation errors and transposition

• Real-time tracking of each specimen in the process

• Automated input of demographic information and digital images into a case

• Access to valuable operational and management information that can be used to optimize process performance and

efﬁciency.

Our experience has reinforced the belief that having a barcoding/asset tracking system in place is a prerequisite to

going 100% digital at scale.

IT and Storage Capacity

A fundamental aspect of a successful deployment is ensuring that an adequate IT infrastructure is in place to support

networking demands and the need for large digital slide storage capacity.

Your Information Technology department will be responsible for:

• Understanding the number of slides generated at 20x and 40x, including large slides (2”x3”)

• Image compression used and ﬁle size

• Overall growth year over year

Our experience demonstrates that, on average, a slide scanned at 40x will produce between 1GB – 2GB of data,

subject to the size of tissue with a 20x image typically generating a WSI of between 500MB – 1GB. For larger 2”x3”

slides the amount of storage per slide is signiﬁcantly greater and proportional to the size of the tissue sample.

The other key impact on storage volumes is the retention time of the scanned image. To ensure a complete diagnostic

record and to allow revisiting of slides in the future, we store all digital images. According to our calculations, a full

digital workﬂow requires 100 TB of storage per year. Fortunately, we have our own bulk storage capabilities. We

will consider archiving older cases onto cheaper storage, though this can mean a short delay when the image is

requested by the pathologist. There are two main aspects to consider when looking at your network requirements:

ﬁrst, connectivity between the scanner and image server, and second, network performance for your total number of

reviewers when running at full capacity. We recommend a dedicated connection between scanner(s) and image server

since you can anticipate continuous high trafﬁc as your WSIs are being generated.

Tips for IT, Networking and Data Storage:

• Identify a local IT lead (systems administrator) and a corporate IT lead.

• Identify and allocate key IT contacts for PC Support / Networks / Infrastructure.

• Set up regular IT meetings for internal issues and utilize video conferencing facilities for regular contact with third-

party suppliers.

• Stress-test network and infrastructure capabilities and identify where enhancements are required.

• Calculate storage requirements and develop a longer-term storage strategy.

• Plan sufﬁcient time for upgrades and system changes.

Allow adequate budget for IT requirements such as increased storage, upgraded network capabilities, as

well as procurement of displays screens, and systems integration.

The Pathologist’s Workstation

The overall feel and usability of your DP system will partly depend on the key hardware components you select for

your diagnosticians, particularly when it comes to display screens and input device selection.

Rebecca A Millican-Slater MBChB MSc FRCPath, Department of Histopathology, Leeds

Teaching Hospitals NHS Trust Leeds, UK

Display Screen Selection

We have chosen to use medical grade, high resolution (6-8 megapixel) display screens for primary diagnostic work.

This decision was inﬂuenced by a screen evaluation experiment in which we invited our pathologists to “road test”

an array of different screens to view a range of standard slide sets and rank them in order of preference.

Tips for Selecting Display Screens:

• It is likely that the majority of work can be accomplished successfully on any modern desktop display screen, with

a minimum resolution of 3-4 megapixels. However, there will be a minority of speciﬁc cases that will beneﬁt from a

display screen with higher resolution, contrast ratio, and luminance.

• Medical grade display screens provide guaranteed levels of consistency over time. If your department will not

permit the purchase of high resolution display screens for all pathologists, you may want to consider purchasing

a smaller number of these display screens, which can be located in a shared space and reserved for difﬁcult or

problematic cases.

• Determine how you will position display screens in your reporting rooms including the physical space that will be

allocated for each workstation. Larger display screens (e.g. 30 inches) allow for easy low-power views of slides.

These substantial display screens can create a more immersive visual experience; however, they can be more

expensive and require more head movement if placed too close to the user.

• Consider the effect of natural light on your display screens. Medical grade display screens tend to be less

susceptible to the effects of natural light on the display screen due to increased luminance. If you opt for display

screens with lower luminance, it is important to ensure that natural lighting can be controlled by using a black-out

blind and alternative artiﬁcial lighting.

• The variation in cost of display screens is vast (from ~£200 to £30,000) and, therefore, it is important to give display

screens due consideration. Note that as the technical speciﬁcations rise, the cost of the display screens increase.

Input Devices

One of the key advantages of DP over conventional light microscopy is the increased ﬂexibility in terms of

your desktop environment. Digital pathologists can use a range of input devices to navigate through their slide

management and slide viewing software. Our pathologists were given scope to trial different devices before choosing

the most comfortable ﬁt for them. All of them now use a combination of keyboard shortcuts and high-performance

gaming mice, which allow ﬁner control of mouse movement with less effort. Some make occasional use of trackball

devices.

Careful selection of a device not only makes slide navigation easier and more efﬁcient, it can also improve ergonomics

and help pathologists with pre-existing work-related strain injuries.

Continuing the Journey

After a DP system has been successfully deployed and integrated, the possibilities for DP applications and utility are

very great. DP offers a ﬂexible foundation allowing institutions to pursue additional service improvement projects and

investigate novel and innovative diagnostic workﬂows. Two of the most talked about applications of digital slides are

remote reporting and the use of artiﬁcial intelligence for automated or partially automated diagnosis.

Remote Reporting

Many pathologists have dreamed of reporting cases from the comfort of their own home (or better still, a luxury villa

in the south of France!). Digital slides certainly put this goal within reach. Given the growing shortage of pathologists

and other factors affecting a decreasing pathologist population, the need for a more ﬂexible and favourable pathology

reporting environment has never been greater.

Career beneﬁts including convenient and comfortable hours and locations are a potential solution to local, regional,

and national specialty-reporting shortages.

If pathology reporting is liberated from temporal and locational constraints, there is the potential for 24-hour access to

diagnostic opinion.

We believe, that as experience and research in this area mounts, professional and regulatory bodies will be able

to provide more detailed guidelines to nurture best practices and ensure the ability to reap the rewards of ﬂexible,

remote working without compromising patient safety or professional standards.

Artiﬁcial Intelligence (AI)

The use of AI to make or augment pathological diagnosis is a hot topic. While nothing can replace the all-round

experience and judgement of a trained professional pathologist, computer-aided diagnosis will help with some of

the more tedious and demanding tasks involved in pathology diagnostics. Algorithmic systems can provide accurate

and repeatable quantiﬁcation metrics for immunohistochemical stains. These systems may be able to assist the

pathologists in screening large volumes of tissue for rare events e.g. metastasis/micrometastasis in lymph nodes or

detecting and counting mitotic ﬁgures for cancer grading systems.

As the experience of pathologists in making primary digital diagnosis develops, AI is likely to expand, with the

accumulating archives of digital diagnostic images providing the big data necessary for the development of more

software tailored to the needs and demands of cutting edge histological diagnosis.

Final Thoughts

Following successful departmental implementation and integration, a digital pathology system offers a ﬂexible

foundation to allow institutions to pursue further service improvement projects, and investigate novel and innovative

diagnostic workﬂows. The possibilities are seemingly endless.

We believe that digital pathology will fundamentally affect the way every pathologist works for the next 10, 20, 30

years. We remain committed to helping other institutions navigate the path in their pursuit of digitization, as we pur-

sue the next generation of innovations that will make the future of digital pathology even brighter than it is today.

Dr Darren Treanor and Jerome Clavel, Head of Digital Pathology for Leica Biosystems, view a digital image

at the Leeds Powerwall.

Further Resources

1. Goacher E, Randell R, Williams BJ,Treanor D (2017) The Diagnostic Concordance of Whole Slide Imaging and Light

Microscopy: A Systematic Review. Archives of Pathology & Laboratory Medicine: January 2017, Vol. 141, No. 1,

p. 151-161.

2. Williams BJ, DaCosta P, Goacher E, Treanor D (2017) A Systematic Analysis of Discordant Diagnoses in Digital

Pathology Compared With Light Microscopy. Archives of Pathology & Laboratory Medicine: December 2017,

Vol. 141, No. 12, pp. 1712-1718.

3. Cancer Research UK. Testing times to come? An evaluation of pathology capacity across the UK. 2016

https://www.cancerresearchuk.org/sites/default/ﬁles/testing_times_to_come_nov_16_cruk.pdf

4. Williams BJ, Bottoms D, Treanor D (2017) Future-prooﬁng pathology: the case for clinical adoption of digital

pathology Journal of Clinical Pathology;70:1010-1018.

5. Ahlers HJ, Stratman C, et al. Can digital pathology result in cost savings? A ﬁnancial projection for digital pathology

implementation of a large integrated health care organization. J Pathol Inform. 2014; 5:33

6. Williams BJ, Bottoms D, Clark D, et al Future-prooﬁng pathology part 2: building a business case for digital

pathology Journal of Clinical Pathology Published Online First: 16 March 2018. doi: 10.1136/ jclinpath-2017-204926

7. Royal College of Pathologists. Best practice recommendations for digital pathology. 2018. https://www.rcpath.org/

resourceLibrary/best-practicerecommendations- for-implementing-digital-pathology-pdf. html

8. Williams BJ, Hanby A, Millican-Slater R, Nijhawan A, Verghese E & Treanor D (2018) Digital pathology for the

primary diagnosis of breast histopathological specimens: an innovative validation and concordance study on digital

pathology validation and training Histopathology 72,662–671.

Dr Darren Treanor hosting an educational workshop on DP at Leeds

Authors:

Dr Darren Treanor [email protected]

Dr Bethany Williams Bethany[email protected]

With contributions from:

Basharat Hussain

Dharshana Jayewardene

Dr Alex Wright

Chloe Lockwood

Dr Emily Clarke

For further information, please contact the authors.

The Leeds Teaching Hospitals NHS Trust and the University of Leeds have a collaborative partnership with

Leica Biosystems for Digital Pathology research driven deployment.

The clinical use claims described for the Leica Biosystems Aperio products in the information supplied have

not been cleared or approved by the U.S. FDA or are not available in the United States.

18778 Rev A ∙ 07/2018