S1403 Implementing Whole Slide Imaging f Cli i l U for Clinical Use

Where the experts are. Where you want to be.™
S1403 Implementing Whole Slide Imaging
f Clinical
for
Cli i l Use:
U
What to Do and What to Avoid
Andrew J. Evans, MD, PhD
Walter H.
H Henricks,
Henricks MD,
MD FCAP
Liron Pantanowitz, MD, FCAP
Mohamed E. Salama, MD
 2014 College of American Pathologists. Materials are used with the permission of the faculty.
Agenda
Topic
Presenter
Introductions and Course Overview
Dr. Evans
Applications, Where to Start, Implementation Team
Dr. Salama
Scanner Selection Strategy
Dr Pantanowitz
Dr.
Required Resources
Dr. Pantanowitz
Adoption Strategies and Validation
Dr. Henricks
What to Expect After Going Live
Dr. Evans
General Discussion
All
P
Pearls
l off Pathology
P th l
and
d Closing
Cl i
2
Course Objectives
• To review the major issues one needs to
consider when implementing WSI for
clinical purposes.
purposes
 not to provide the complete handbook on
how to do it
Institution-specific issues
3
Definitions
• Telepathology
 transmission of pathology images and
patient information various clinical
applications.
• Whole
Whole-slide
slide imaging (WSI)
 high-resolution digitized replica of glass
slide created by a scanner
 manipulated by software to simulate
microscope review (virtual slide)
4
WSI: Value Proposition
• Growing emphasis (and public expectation)on
improving quality in pathology by creating
healthcare networks
• WSI enables:
 enhanced access to sub-specialty expertise
 decreased need to transport glass slides
 reduce cost
 avoid
i lost or broken slides
i
 improved turnaround times for consultations
 work load levelling across sites
 remote/under
remote/under-serviced
serviced regions
 support for solo pathologists
 enhanced recruitment and retention of local surgeons
5
WSI: Barriers to Adoption
p
• Cost,, IT support/infrastructure
pp /
• Pathologist perception of inferior performance
• Lack of standards/best practice guidelines
• Regulatory issues - FDA
• Medicolegal
di l
l liability,
li bilit lilicensure and
d credentialing
d ti li
• Professional billing/business models
6
7
Digital Pathology Committee
• 2012 – Digital Pathology Working Group
• Dr. Eric F. Glassy, Chair
• To support the development by the various
College committees of CAP programs,
products and requirements that contain
digital pathology applications and to serve
as a resource for the CAP public position
on digital pathology applications and
practice tools.
p
8
Where the experts are. Where you want to be.™
Applications, Where to Start, Assembling an
Implementation Team and Work Flow
Considerations
Mohamed Salama MD
U i
University
it off Ut
Utah
h & ARUP Reference
R f
Lab
L b
 2014 College of American Pathologists. Materials are used with the permission of the faculty.
Outline
•
•
•
•
•
•
•
Terminology
Where to start
I
Implementation
l
t ti
phases
h
System components
Assembling an implementation team
Common applications
Selected examples
11
Digital technology Usage
Static images
for photomicroscopy
Static images
f gross photography
for
h t
h
78%
75%
Whole slide images (WSI)
Image analysis
32%
9%
30%
14%
Video microscopy
(live video feed)
Telemedicine/
telepathology
PACS (picture archiving
and communication system)
20%
Digital consultations
20%
Robotic microscope
2%
25%
Currently
y Use
56%
66%
14%
61%
7%
72%
15%
7%3%
23%
59%
9%
25%
0%
2%
20%
64%
90%
25%
50%
Plan to Use Within Next 12 Months
75%
100%
Do Not Use
© 2014 College of American Pathologists. All rights reserved. 12
Terminologies
g
Source: Williams S et al. Adv Anat Pathol. 2010;17(2):130‐49.
13
Pre‐impleementation
Implem
mentation
Post‐
impleementation
n
where to start
 Define your need, goals and objectives  Identify required infrastructure Identify required infrastructure
 Secure resources
 Acquire and build infrastructure
 Train implementation team and pathologists  Define workflow and configure system
g
y
 Write SOPs , validate system  Go live  execute process
 Assess efficiency  Expand to new applications
14
Pre-implementation
Pre
implementation phase
 Define your need, goals and objectives
• Define your needs  application
• Start Simple
• Consult with colleagues
g
and collaborators
• Assess what may or may not work in your institution
 Identify required infrastructure in you institution
• Hardware / scanners
• Software / IT infrastructure including storage
• Implementation team (personal)
 Secure resources
• Department / operation support
• Budget
15
Implementation phase
 Acquire and build infrastructure
• Vendor installation
• IT setup
 Assemble Train implementation team
• Operators
• Management team
• Pathologists
P th l i t
 Configure and Validate system
• Federal regulations
• CAP Guidelines & checklists
 Define workflow and Write SOPs
 Go live  execute process
16
Assembling an implementation team
Technicians
Slide management system
System administrator
Server
Working stations
Work flow and QA
Pathologist
IT support
Coordinator
Implemen
ntation Team
Syystem ccompon
nents
Scanner
Pathologist
17
Implementation Team task assignments
Scanning
E‐slide management
Work flow Server
management
Work station
Interface user modules
Technicians
System administrator
Coordinators
IT support
Pathologist: Super user
Pathologist: User 18
Implementation Team expansion is a
d
dynamic
i process
Scanning
E‐slide management
Work flow Server
management
Work station
Interface user modules
One Technician
Pathologist: Super user
19
Implementation Team expansion is a
d
dynamic
i process
Scanning
E‐slide management
Work flow Server
management
Work station
Interface user modules
Technicians (2)
System administrator (1)
Pathologist: Super user
Pathologist: User 20
Implementation Team expansion is a
d
dynamic
i process
Scanning
E‐slide management
Work flow Server
management
Work station
Interface user modules
Technicians (4+2)
System administrator (1+1)
Coordinators (2)
IT support (2)
Pathologist: Super user
Pathologists : User (40+) 21
Common WSI applications in clinical
l b
laboratory
t
•
•
•
•
•
•
•
Frozen section
Consultation
Primary diagnosis
Quality assurance
O it cytology
On-site
t l
assessmentt
Tumor boards
Teaching
• Trainees and lab personal
22
11%
Other
Multtiplex image a
analysis
11%
Primary diiagnosis
17%
Frozen ssection
interpre
etation
18%
Inte
ernational in-ssourcing
Quality Asssurance
Virrtual IHC
25%
PT from CAP or other
vende
er
Digittal image ana
alysis of
ER, PR, HER2, e
etc.
Sec
condary consu
ultations
54%
Re
esearch
Arrchiving
Clinical confe
erences
75%
Te
eaching
Profile of those using WSI
69%
63%
50%
51%
34%
29%
23%
15%
9%
3%
0%
© 2014 College of American Pathologists. All rights reserved. 23
Thinking about current state when
d fi i
defining
work
k flow
fl
24
24
Thinking about current state when
d fi i
defining
process
25
25
Workflow process
Cases listing by clinical treatment planning team
g y
p
g
Pathology coordinator arrange slides Slides selection and marking for oil scanning by pathologist
Slides selection and marking for oil scanning by pathologist
Slides scanned by technicians
Digital slide assigned to cases by technician in E‐slide manager
Pathologist notified once scanning is competed Pinning is completed by pathologist
Cases are presented in tumor board
Cases are presented in tumor board
26
“Pinning” function in eSlide Manager
27
Scrolling through pins during TB
28
Pre‐implementation
Implem
mentation
Post‐
impleementation
Next Step
 Define your need, goals and objectives  Identify required infrastructure Identify required infrastructure
 Secure resources
 Acquire and build infrastructure
 Train implementation team and pathologists  Define workflow and configure system
Define workflow and configure system
 Write SOPs , validate system  Go live  execute process
 Assess efficiency  Expand to new applications
29
Results
Mean time in minutes (range)
Prep time/presented slides
Microscope
projection (MP)
projection (MP) 1.39 (0.26 – 5.0) WSI
1.14 (0.80 – 1.19)
Prep time/total number of slides
0.32 (0.13 – 0.83)
0.51 (0.39 – 0.72)
P
Prep time/number
ti /
b of cases
f
2 64 (1 03 6.67)
2.64 (1.03 –
6 67)
3 85 (2 32 4.85)
3.85 (2.32 –
4 85)
Discussion time/presented slides (p = 0.03) 0.77 (0.37 – 1.31)
0.45 (0.29 – 0.53)
30
• 85% experienced reduced motion sickness with WSI
31
Advantages of WSI for TB
• Higher image quality/high power oil
immersion scanning
• More efficient use of TB time ((due to p
pinning
g
function)
• Ability to demonstrate all findings in real-time
• Increased clinical team satisfaction
• Archival of scanned slides
• Little
Littl incremental
i
t l costt for
f TB presentation
t ti
if
WSI already implemented
32
Challenges implementing WSI for TB
• Internet connectivity
• Last minute additions (<24 h)
– Inadequate
I d
t prep time
ti
by
b pathologist
th l i t or
technician
• SScanner ttechnical
h i l iissues (<5% off cases))
• WSI scan rate: 84% from cases requested
• Significant capital outlay to set up WSI
system
y
33
© 2014 College of American Pathologists. All rights reserved. 8%
Other
Insufficient educational
materrials
8%
Poor ima
age quality
10%
Unproven technology
26%
Fear or lack
k of
un
nderstanding of technology
25%
Lack of need
28%
Lack o
of demand
30%
Disruption iin workflow
31%
No FDA c
clearance
for primary
y diagnosis
31%
IT- spe
ecific issues
Complexity off purchase
and installlation
Cost (sc
canner,
storage, sup
pport, etc.)
Barriers to adoption of WSI technology
100%
77%
75%
50%
22%
8%
0%
1%
%
34
Pre‐impleementation
Implem
mentation
Post‐
impleementation
n
Expanding WSI
 Define your need, goals and objectives  Identify required infrastructure Identify required infrastructure
 Secure resources
 Acquire and build infrastructure
 Train implementation team and pathologists  Define workflow and configure system
g
y
 Write SOPs , validate system  Go live  execute process
 Assess efficiency  Expand to new applications
35
Where the experts are. Where you want to be.™
Scanner Selection Strategy
Liron Pantanowitz MD
UPMC
 2014 College of American Pathologists. Materials are used with the permission of the faculty.
39
Available WSI Scanners
40
Matching Attributes with Intended Use
Q
Questions
ti
tto Consider
C
id
• What is my
y intended use ((remote frozen sections,,
image analysis, fluorescence research, digital slide
teaching sets, etc.)?
• What
Wh t is
i my existing
i ti
IT iinfrastructure
f t
t
& support?
t?
• Occasional scans vs. high throughput work?
• Type of glass slides to be scanned?
• Type of cellular material to be digitized?
• Does my
y lab require
q
WSI-LIS integration?
g
• How much can I afford to spend?
41
WSI System Attributes
•
•
•
•
•
•
•
First understand WSI systems
Physical device features
Scanning (digitization) features
Images & image management
S ft
Software
(viewers,
( i
image
i
analysis)
l i )
Workflow features (collaboration)
Match these with your needs
42
43
Physical Features
• Instrument size (large vs. desktop)
• Cameras (proprietary, CCD vs. CMOS)
• Slide loading & handling capacity
– Racks, trays, cartridges, hotels
– Capacity (range 1 to 400 slides)
– Walk away (continuous autoloading)
i ((standard
t d d vs. whole
h l mount)
t)
– Slid
Slide size
44
49
WSI Scanner Review
Rojo MG. European Society of Pathology. 24th European Congress. Sept 2012. Prague
WSI scanners have improved in the last 5 years,
but they still have a way to go!
• Scanning speed:
– Average
g 1 cm2/min at 20x
– Some can scan slides <60 sec at 40x
• Scan failure rates:
– Can
C be
b as high
hi h as 5% for
f 20x
20 & 20% f
for 40
40x
• Image quality
– Dependent on objectives,
object ves, average NA = 0.75
• Image resolution
– Average 0.5 micrometers/pixel at 20x
51
Scanning Features
• Scan time (1-8+ minutes, per 15x15 mm2)
• Scan method (line vs. tile based scanning)
• Brightfield vs.
vs Fluorescent (spectrum range
may restrict fluorophores) (e.g. 380-680nm)
• Manual vs.
vs Automated (walk away capability,
capability
scan & view slides simultaneously)
• Tissue detection (automatic, manual override to
scan user defined areas)
• Color calibration (automatic white balance)
52
WSI Scanning
g Strategies
g
A ‐ Tile‐based acquisition mode (grid pattern)
q
(g p
)
B ‐ Line‐scan acquisition mode (linear pattern)
B ‐
Line‐scan acquisition mode (linear pattern)
Pathology Informatics: Theory & Practice. Pantanowitz et al. ASCP Press. 2013.
53
54
Image Quality
• Objective
j
lenses (numerical
(
aperture)
– E.g. 20x - 0.65 to 0.85 NA Plan Apo
– E.g.
E 40
40x – 0.65
0 65 tto 0
0.95
95 NA Pl
Plan A
Apo
• Available scanning magnification
(4x, 20x, 40x, 60x, 83x oil)
• Scanning resolution varies per
magnification
– E.g. 20x – 0.38 vs. 0.5 μm/pixel
– E.g. 40x – 0.25 μm/pixel
55
Digital Magnification & Resolution
Sellaro TL et al
al. J Pathol Informatics 2013; 4:21
• Traditional glass microscope image quality monikers
((e.g.×40)
g
) should be replaced
p
with vendor-neutral
descriptors (e.g. microns/pixel)
56
ZZ-Stacking
Stacking Features
• Z-Stacking capability:
– Yes (e.g. Hamamatusu, 3DHISTECH,
Huron, Leica, Ventana)
• Best for cytology slides
• User selectable zz-scans
scans
– Z spacing (typically 0.25, 0.5 or 1 micron)
– # planes (typically 3-15 offered)
57
Z Stacks
Z-Stacks
3D cell group
Z stack
Focal point interpolation
58
Multiplane Images
Take a long time to scan
Produce large files
59
Images & Image Management
• Available file formats
– Primary formats are proprietary (e.g. SVS,
MRXS, SWS, RTS, BIF, NDPi)
– Open file formats (e.g. TIFF, JPEG 2000, BMP)
1
Software: Image Viewers
• E.g. ImageScope, NDP.view, others
• Local install (server & workstation)
• Web-based (for online collaboration,
collaboration
may have different features)
• Mobile device (e.g.
(e g iPad) compatibility
• Multi image viewing (e.g. side-by-side,
overlay
l
off images)
i
)
• Ability to view other file formats
2
Image Applications
3
Omnyx Viewer Split Screen Display
4
Software: Image Analysis
• Various image analysis algorithms
• E.g. IHC quantification (some are FDA
approved
app
o ed for
o diagnostic
d ag os c use)
• Image analysis vendor relationships
– E.g.
E g Visiopharm,
Visiopharm Definiens,
Definiens Indica labs
5
Workflow Features
• Image management software (e.g. eSlide
manager, Omnyx
O
IDP)
• Image file sharing options (local vs. virtual
server cloud services,
server,
services enterprise sharing
e.g. eSlide share)
• Telepathology employing real-time remote
control with hybrid robotics (e.g. Sakura’s
VisionTek, Mikroscan))
• Barcode reader support (1D, 2D)
• LIS integration
g
((required?,
q
, vendor solutions))
6
Omnyx Pathologist Worklist
9
Omnyx “Slide
Slide Tray”
Tray View
10
Where the experts are. Where you want to be.™
Required Resources
Liron Pantanowitz MD
UPMC
 2014 College of American Pathologists. Materials are used with the permission of the faculty.
IT Resources: Technical
• WSI system (scanner,
(scanner software & PC)
• Server(s) ± back-up storage
• Network
N t
k infrastructure
i f t
t
(bandwidth,
(b d idth
security, 3rd party applications, firewalls)
• Monitors (workstations, mobile devices)
• Operations
p
p
plan ((to mirror clinical use))
12
IT Resources: Personnel
• IT staff for installation & networking
• Vendor consultation & maintenance
• Lab
L b staff
t ff for
f training
t i i
users, validation,
lid ti
imaging slides, troubleshooting &
compliance
li
(documents,
(d
t metrics
t i data)
d t )
• Pathologist champions (time investment)
• ± Tool development (programmers)
(admin, contract) support
• Business (admin
13
Data Storage
•
•
•
•
•
•
•
•
Robust file storage
g p
platforms
Scalable archival storage plan
Cost may be an impediment
Plans should include back-up
Retention (purge) policies
Image compression
i
& integrity
i
i
Database should support all searches
Maintain high-performance
high performance throughput (e.g.
(e g network
connectivity to upload/download)
14
Image
g File Sizes
Romero Lauro G et al. J Digit Imaging. 2013.
Image
Avg. Size (MB)
Avg. Size (MB)
Endoscopy image
7
MRI 100
CT scan
150
PET
370
WSI (core biopsy 20x compressed)
300
WSI (core biopsy 20x compressed)
750
15
LIS Integration
• Is it required?
– Simultaneous
Si
lt
access tto iimage & clinical
li i l metadata
t d t
– Single database (repository) to maintain
• What is currently available?
– Limited vendor solutions exist
– Open architecture (WSI interface with LIS vendor)
– Integral
g (within
(
LIS)) vs. Modular ((middleware)) options
p
• Challenges
–
–
–
–
–
Complexity & variability of pathology workflow
Slid (case)
Slide
(
) delivery
d li
drives
di
workflow
kfl
(push
(
h vs. pull)
ll)
Ergonomic workstation display (image/LIS “real estate”)
Availability of extra digital slides (recuts, stains)
Reporting options (transcription,
(transcription synchronous telepathology)
16
WSI-LIS INTEGRATION
AP‐LIS Digital Workstation AP LIS
17
Expected Costs
• Direct Costs
• WSI scanner
• Server(s) that meets specifications
• Other computer equipment (e.g.
(e g monitors)
• Indirect Costs
• Maintenance fee (vendor,
(vendor annual)
• Licensing fees
• Additional software (image management,
educational
d
ti
l package,
k
iimage analysis)
l i)
• LIS vendor integration (customization)
• IT costs (personnel, operations)
18
Where the experts are. Where you want to be.™
Getting Pathologists Comfortable with WSI:
Adoption Strategies and The Importance of
Validation
Walter H. Henricks, M.D.
Cleveland Clinic
 2014 College of American Pathologists. Materials are used with the permission of the faculty.
Understanding
U
d t di
pathologist
th l i t concerns
is crucial to implementing WSI.
Walter H. Henricks, M.D.
20
To get pathologists
pathologists’ buy
buy-in
in for WSI:
• Overcome resistance
• Eliminate
Eli i t b
barriers
i
21
Resistance to WSI for clinical use
stems
t
from:
f
FEAR
M ki
Making
a di
diagnostic
ti error
LOATHING
Being forced into using WSI
PRESSURE
Taking more time per case
vs glass
vs.
COSTS
Direct and indirect costs
22
What contributes to pathologist
discomfort with using WSI?
• Pathologists may have little to no training
or experience
p
in using
g WSI systems.
y
• Substantial mechanical and ergonomic
differences exist in WSI vs. microscopy
py
• WSI may introduce less-familiar “digital
artifacts” in addition to familiar histology
artifacts
artifacts
23
Walter H. Henricks, M.D.
Strategies to Improve Pathologists’
C
Comfort
f t with
ith WSI
• Provide literature that shows good
correlation between WSI and glass
• Perform internal validation of WSI for
intended uses
• Offer
Off less
l
stressful
t
f l ways for
f pathologists
th l i t
to gain experience with WSI
• Ensure adequate training
y valve”
• Provide “safety
Recent Publications with WSI-Glass Correlation
Reviews
Pantanowitz L, Sinard JH, Henricks WH, Fatheree LA, Carter AB,
Contis L,
L Beckwith BA,
BA Evans AJ
AJ, Lal A,
A Parwani AV.
AV Validating
whole slide imaging for diagnostic purposes in pathology:
Guideline from the College of American Pathologists Pathology
and Laboratory
yQ
Quality
y Center. Arch Pathol Lab Med. 2013
Dec;137(12):1710-22.
Bauer TW, Schoenfield L, Slaw RJ, Yerian L, Sun Z, Henricks WH.
Validation of whole slide imaging for primary diagnosis in
surgical pathology. Arch Pathol Lab Med. 2013 Apr;137(4):51824.
Pantanowitz
Pantano
it L,
L Valenstein PN,
PN Evans
E ans AJ,
AJ Kaplan KJ
KJ, Pfeifer JD,
JD
Wilbur DC, Collins LC, Colgan TJ. Review of the current state of
whole slide imaging in pathology. J Pathol Inform. 2011;2:36.
Cornish TC, Swapp RE, Kaplan KJ. Whole-slide imaging: routine
pathologic diagnosis. Adv Anat Pathol. 2012 May;19(3):152-9
25
Perform internal validation of WSI for
i t d d uses
intended
• Validation: demonstrating that a
method or system produces expected
results or outcome in a given setting.
setting
• Performance characteristics to be
assessed
• Acceptable levels of performance.
26
WSI Validation – Why?
• Diagnostic pathology is interpretive and
practice of medicine – how can it be
validated?
• WSI images depend on complex hardware
and software components
• Use of such a system for diagnostic result
merits performance verification
• Consensus: perform WSI validation for clinical
uses
27
Arch Pathol Lab Med. 2013;137:1710–1722
•
•
•
•
•
Expert panel experienced in WSI use
767 articles on “digital pathology” filtered
23 articles after rigorous inclusion criteria
Public comment p
period on draft g
guidelines
12 guideline statements with evidence
grades
28
Highlights from CAP WSI Validation
G id li
Guidelines
•
•
•
•
•
•
•
•
Intended clinical use(s)
Emulate real world clinical environment
Entire WSI system as whole
Pathologist involvement
At least 60 cases
Intraobserver diagnostic concurrence
At least 2 week “washout” period
Confirm scanning of all material on slide
29
Rationale for “Intra-observer” WSI
V lid ti
Validation
• Pathologist makes same diagnosis on same
slide, digital vs. glass
e e a (for
( o this)
s) whether
e e or
o not
o diagnosis
d ag os s iss
• Irrelevant
“correct” or agrees with another pathologist
g
expertise
p
or experience
p
affect
• Diagnostic
“correctness” of diagnosis
• Removes inter-observer variation such as
criteria, “thresholds”
• Expert consensus: intra-observer more
important for clinical purposes
30
Another Approach to WSI Validation
Arch Pathol Lab Med. 2013;137:518–524
•
•
•
•
•
•
•
•
a priori power study: need 450 cases to demonstrate noninferiority of WSI vs. microscope re-review of previously
dx’ed cases.
Hypothesis: WSI review non-inferior if no more than 4% major
discrepancies
Pathologist
a o og s review
e e o
of own
o
cases from
o 1 yea
year p
prior,
o , us
using
g
microscope or WSI
607 cases composed of 1025 individual parts
Major discrepancies: 1.65%
1 65% WSI and 0.99%
0 99% glass
Minor discrepancies: 2.31% WSI and 4.93% glass
WSI dx better than original dx in a few cases
C
Conclusion:
l i
di
diagnostic
ti review
i
b
by WSI was nott iinferior
f i tto
microscope slide review (p<0.001)
31
Ease Pathologists into Using WSI
•
•
•
•
•
Interesting case archives
Teaching files
T
Tumor
boards
b
d
Clinical conferences
Research collaborations
32
Ensure Adequate Training
• Training ≠ Validation
• How to use the WSI viewer software
• How to perform workflow for intended
use(s)
–
–
–
–
How to know WSI case is available
How to access the WSI case
What to do to defer to glass
How to generate and distribute report
• Checklists and sign off sheets helpful
33
Do Not Underestimate Training
Walter H. Henricks, M.D.
Provide “safety
safety valve”
valve
35
To get pathologists
pathologists’ buy
buy-in
in for WSI:
• Overcome resistance
• Eliminate
Eli i t b
barriers
i
36
Make WSI Life Easy(
Easy(-ier)
ier) for Pathologists
Lack of attention to WSI workflow can create
barriers for pathologists:
• Access to clinical information for cases
submitted
b itt d by
b WSI
• Notification that WSI case is waiting
• Mechanics/ergonomics of WSI case review
• Report generation and distribution
• Deferrals to glass slide review
• Obtaining additional studies
Accounting for WSI in Pathologists’
W kl d
Workload
• More time per slide/per case for WSI (20500% increase)
• Expectations of shorter TAT for WSI cases
• Scheduled and/or ad hoc availability for
real-time
l ti
consultation
lt ti
or discussion
di
i
on
WSI cases with remote sites
• Expectations of remote clients for more
informal consults
38
Potential “Side” Benefits of WSI for
P th l i t
Pathologists
New computer monitors and other equipment
Elimination of travel to cover remote sites
Exposure to new technologies
Innovative practice models
Archive of interesting cases for teaching,
teaching
research, etc.
• Opportunity for flexible work schedules
• Enhanced revenue through expanded
consultative practice
•
•
•
•
•
39
From Sten Thorstenson, M.D. presentation: www.telepathology2010.com/uploads/_fck/1_vilnius_digital_routine_histopathology_%5Bcompatibil
40
ity_mode%5D.pdf
Walter H. Henricks, M.D.
Where the experts are. Where you want to be.™
What to Expect After Going Live: UHN’s
Experience
Andrew Evans MD
University Health Network
Toronto, ON, Canada
 2014 College of American Pathologists. Materials are used with the permission of the faculty.
Regulatory Disclosure
• Health Canada has approved the use of
WSI for routine use in surgical pathology,
including primary diagnosis in Canada.
Canada
• WSI has
h nott been
b
approved
db
by the
th FDA
for primary diagnosis and such systems
are considered
id
d investigational
i
ti ti
l until
til such
h
approval is granted.
42
WSI: Clinical Applications at UHN
• Frozen sections
 Toronto Western Hospital – 1 mile
 Kingston General Hospital – 165 miles
 Timmins and District Hospital – 425 miles
• Consultations/Quality Assurance
 Hospitals in greater Toronto area, northern
Ontario and Kuwait
• Primary diagnosis
 Lakeridge
g Health Oshawa – 40 miles
43
Why
y Telepathology
p
gy at UHN?
• Full departmental consolidation at TGH in
early
l 2006
• No regular on-site pathologist at TWH as of
2004
TWH Frozen Sections: Challenges
• Single
Si l pathologist
th l i t traveling
t
li
from
f
TGH to
t TWH
– Inefficient - traveling and waiting
– Disruptive to regular workflow at TGH
o delays
d l
iin regular
l sign-out
i
t affecting
ff ti
other
th UHN patients
ti t
– No consultation on difficult cases
o potential to affect TWH surgical patients.
44
TWH Whole-Slide Imaging: October
2006-Present
• > 4000 frozen sections/3500 patients
• > 90% from neurosurgery
•14-16
14 16 minute total turnaround
time
• 0-2% discrepancy rate
• < 1-5% deferral rate
 difficult frozen sections
 not caused by use of WSI
- 2 pathologists review all deferrals
45
Intra-Operative Consultations: Work
Flow for Single Block Frozen Sections
10-12
minutes
1-4
minutes
Times Documented – tissue received, slide scanned, image received, surgeon called
46
Skin Ellipses: en face Margins
Range
g of separate
p
frozen
sections per case = 4-17
 Requires PA or assistance from
surgeon
g
if no on-onsite p
pathologist
g
47
Value of Daily
y System
y
Test
• Unexpected network failure.
• IP address change – no connectivity.
• Scanner issues:
– dirty objective
– background calibration
48
Episodes of Mid-Case System Failure:
• 9 episodes (0.2% of cases to date) requiring a
pathologist to go to TWH
–
–
–
–
Small pale pieces of tissue (x2)
E
Excess
mounting
ti
media
di ((x1)
1)
Burned out light bulb (x1)
Calibration errors (x5)
 dirty slides
 aging light bulb
www blog al com/spotnews/2009
www.blog.al.com/spotnews/2009
49
Why Has This Worked at UHN?
• St
Started
t d with
ith a clearly-defined
l
l d fi d
application
• Uncomplicated
U
li t d frozen
f
section
ti
workflow
• 18 month
th development
d
l
t period
i d
 time to build confidence and trust
• Team
T
approach
h
 trust-worthy histotechnolologists/PA’s to carry out
delegated tasks
50
Image Quality: The Importance of
G d Histology
Good
Hi t l
• Poor quality slides = poor image quality
• 20x scans –
20
ask for 40x when necessary k f 40 h
51
Poorly Focussed Areas ?
Initial scan
Re‐scan
Re‐scan selected areas – adds 1‐2 minutes to turnaround time (< 5%)
52
Primary
y Diagnosis
g
By
y WSI
• First diagnosis made on scanned slide images (H&E, special stains/immunohistochemistry)
• Diagnostic information becomes part of the Di
i i f
i b
f h
patient record
• Treatment decisions to be made based on this Treatment decisions to be made based on this
information
53
65 km/40 miles
Lakeridge Health-UHN
• Transition of Lakeridge to subspecialty reporting 2011-2012
including a common LIS
• 5-8 on-site pathologists
54
Routine scanning at 20x
UHN Data Center
Local
Server
Secure Private Network
Central Server
Pathologist
55
Phased Implementation Strategy
• Complete a validation study
• Start with most experienced users
– GU,
GU endocrine,
d
i
liver,
li
head
h d and
d neck
k
• Attempt to scan all cases for these
groups
• Review digital slides and sign out
– request glass slides whenever it is required
to sign out a given case
56
Transurethral Resection of Bladder Tumor
57
pT2
CIS
LVI
58
November 1, 2012-January 31, 2014
• Cases scanned for primary diagnosis
– 2990
• Total slides
– 15097 (cases ranging from 1-74 slides)
• Cases
C
signed
i
d outt without
ith t glass
l
review
i
– 2850 (95%)*
• Cases deferred to glass review before sign-out
– 140 (5%)
• Inaccurate/Incomplete Diagnoses as Pointed Out
By
y Clinicians or External Consultation ((Glass Slide
Review)
– 0
59
Reasons for Deferral to Glass Slides
• Learning curve issues/difficult cases –
pathologists still establishing comfort
level (80% of deferred cases)
• IT performance issues (15%) – too slow
• Sub-optimal image quality - soft focus
in area of importance (4%)
• Things you cannot do digitally the way
you would with a microscope
y
p ((1%)) *
60
IT Infrastructure
• 32 bit encryption operating system and
limited RAM (4 GB) – need at least 12 GB
• Unacceptable lag/pixelation – pan and
zoom on cases of > 40 slides
61
WSI – LIS Interface Issues
*Pathologists access all cases and scanned slides
through the LIS need for back‐up access!!!!
through the LIS –
need for back up access!!!!
• Interface security password expiry
– every 3 months
– shut system
y
down for 10 days
y on first occurrence
• Slides scanned but not crossing the interface
– incomplete cases
– shut system down for 2-3 weeks
• Bar codes not being read
– < 20 slides
– printing imperfections/stain on the label
62
Combined Glass Slide – WSI
Workflow
• Varying
y g levels of p
pathologist
g experience
p
with WSI diagnosis
– what slides to scan for which pathologist
– pathologist duty roster changes
• Culture change
– using an electronic work list
• Special stains
– when
h
are they
th ready
d for
f review?
i ?
– glass slides sent when pathologist is looking
for scanned slides
63
User Interface
•
Problematic for large cases and/or slides requiring a lot of panning and zooming
64
Pearls of WSI Implementation
• Applications, Where to Start, Implementation Team
 Define an initial application and assemble a team
 Identify critical infrastructure and work flow requirements
 Implement,
Implement learn lessons and expand
e pand to new
ne applications
• Scanner Selection and Required Resources
 Match scanner attributes with intended use and budget
 Think carefully about image storage/retention and the need for LIS
integration
• Adoption Strategies and Validation




Understand
d t d pathologist
th l i t concerns over adopting
d ti
WSI
S
Ensure adequate training
Perform internal validation according to intended use(s)
Provide a safety valve (defer to glass when required)
• What to Expect After Going Live




WSI can be used safely and reliably for diagnostic work
Deferral to glass slides will occur (should be < 5% of cases)*
You need a ‘Plan B’ if the system fails
Some initial growing pains associated with mixed work flows
65
References
• List to be provided during the course.
course
66
Where the experts are. Where you want to be.™
Thank you for participating!
Please complete the course
evaluation
l ti
b
before
f
you lleave.