Pursuit of an Optimized Surgical Pathology Workflow with Digital Pathology Integration

J. Mark Tuthill, MDDivision of Pathology Informatics

Henry Ford HospitalDetroit, MI 48202mtuthil1@hfhs.org

Digital Pathology and AI WorkshopPittsburgh, PA

December 13-14, 2019

Objectives1. Understand digital integration prerequisites and

opportunities2. Present examples of operational systems used and

digitally integrated at HFHS3. Highlight common strategies the lead to success4. Describe the integration of WSI 5. Recognize challenges and points of failure in digitial

workflow and WSI

Philosophical Hypothesis• “We can integrate AP workflow so that it models

efficiency of the clinical laboratory”– What are the pre-requisites?– What is the low hanging fruit?

• Sequence– What are the technology gaps?

• Advanced robotics and pipeline development– Automation of manual processes

Operational Systems

Digital Integration

Lab Information

System

TECHNICAL & SUPPORT STAFF

PATHOLOGISTS

Information Technology

Sunquest :Lab and CoPath Histotrak HLA

MAS POCAqueduct Hematology

Lane Faxing

LAB PORTAL Atlas

APOLLO ARCC

Scantron

LAB USER’S GUIDE

pathology.hfhs.org/lug

Digital PathologyRoche

MikroScanDigital Cameras− Store AP Req Scans

− Store Clin Path Req Scans− Integrate External AP Results− Telepath Integration− Associate Imaging to Reports− Interface to acquisition devices

Cameras (1-2 M records/yr)- Gels-Microbiology

− Surg Path Reqs− Cytopath Reqs− Clin Path Reqs− Outreach Documents

SCANNING

TISSUE BIOREPOSITORY

DOCUMENT CONTROL

Master Control

HEALTHSTREAMTraining (HFHS U)

Competency

Wired

Molecular Pathology and NGS

EpicSyapse

Clinical Pathology SystemsFunctional Areas Applications• Chemistry• Hematology• Coagulation• Urinalysis• Blood Bank• Microbiology• Serology• Virology• HLA• Molecular• Genomic Pathology• Point of Care

– Glucometers >650– Coagulation clinics

• SunquestLab• SMART – barcoded specimen tracking,

management. and archiving

• >170 Instrument Interfaces via Sunquest Instrument Manager (SIM)

• Autovalidation• Macroscheduler automated jobs• Histotrac: HLA laboratory system• MSQL Report Writer w/ Crystal• Results transmitted to EPIC• Beckman Automation Line; Conexus;

Remisol

Anatomic Pathology SystemsFunctional Areas Applications• Histology• Molecular Pathology

– Illumina NGS

• Cytology• Frozen Room• Autopsy• CP Integration

– Special Hematology– Flow Cytometry– Immunohistochemistry– Microbiology

• Sunquest CoPath• Biomaterial tracking system (BTM)• Scantron: requisition scanning• mTuitive: synoptic checklists• JFCC reporting via ePath• Barcoded LEAN production• Apollo Digital photography• Telepathology: Mikroscan• WSI: Ventana Virtuoso, Mikroscan

– Ventana Connect

• Dako Connect• Tens of molecular pathology applications

Other LIS Applications and Operations Functions• Atlas Portal: Outreach EMR; order entry and result reporting• Scantron: Document scanning for regulatory retention• LUG: Laboratory Users Guide• Lane Faxing: supports all Health system laboratory faxing• BTM: Tissue Biorepository• IT Service Now: Helpdesk; PI Change Control database for compliance• Apollo ARCC: Image storage integrated with both

Sunquest systems (Lab & CoPath)

• Cytovision Cytogenetics imaging system for Karyotyping and FISH analysis

• Hologic Automated pap smear screening

Road Map for Digital WorkflowThe Big Picture

• Pre analytic– Prior to receiving or analyzing the sample– Preparing samples for analysis

• Analytic– The process of analyzing the tissue

• Post analytic– The reporting of diagnostic information– Preparing for additional analytic studies

Prerequisites for Anatomic Pathology Digital Workflow

• Sophisticated electronic medical records system– Electronic orders interface for Anatomic Pathology

• Bar code labeled assets with the laboratory– Assets with unique identifiers

• Development of robotic technologies– *Grossing*– Embedding, Sectioning– Tissue transport– Sampling– Storage systems: cassettes and slides

Prerequisites for Anatomic Pathology Digital Workflow

• Electronic orders interface to the Anatomic Pathology Laboratory Information System (AP-LIS)– Similar to the clinical laboratory, a flow of orders to the LIS

will enable:• Sample receipt• Tracking• Routing• Analytic Processing• Automation of several elements of case accessioning

– Decrease errors– Increased throughput

Prerequisites for Anatomic Pathology Digital Workflow

• Bar code labeled assets within the laboratory– This is most essential early prerequisite to achieve

automation within the laboratory– Bar coding of assets allows for:

• Bar code driven workflow• Identification error reduction due to mislabeling• Improved efficiency by reducing manual labeling• Automation of subsequent activities

– Integration whole slide imaging, interface devices– This is the key requirement for all automation

Prerequisites for Anatomic Pathology Digital Workflow

• Bar code labeled assets within the laboratory– Critical for each asset to have an unique ID embedded in the bar

code– This will allow each block and slide to be managed uniquely

supporting (Assets)• Sophisticated routing• Tracking of assets • Digital Pathology (unique ID on slides will be essential!)• Systems interfaces

• Without uniquely identified assets the clinical laboratory could not have achieved the level of automation currently experienced

Examples of Anatomic Pathology Digital Workflow

Real World Examples and Status Updates

Examples of AP Digital Workflow• Automation of histology orders (stain protocols)• Interfaced immunostain orders to automated immunostain

platform• Bar code labeling automation

– Automated production of cassettes at accessioning– Cassette driven generation of labeled slides

• Tracking, routing and storage• Automated tissue embedding, Sakura AutoTEC• Automated microtome's

Examples of AP Digital Workflow• Automated block sampling• Automated slide sampling

– Laser capture micro dissection• Conveyor belt systems, tubes, roving robots: routing• Slide collation robotics• Automatic diagnostics

– Whole slide imaging algorithms for immunostains quantification

– Automated pap smear readers

Examples of AP Digital WorkflowHistology Protocols

• Automated ordering histology protocols for different sample types at case accession– When a particular part is accession the appropriate blocks and

initial stain orders are generated– Initial billing fee codes are applied– Histology logs are electronically sent and printed providing early

notification of work• This has increases efficiency and allows for LEAN processes

– Work is standardized– Revenues were enhanced through better charge capture

• This is not easy and required iterative re-work and constant attention to defects to get the most satisfactory end result

Examples of AP Digital WorkflowReal Time Labeling

• Essential first step to widespread AP automation• As previously stated the implications of bar code labeled

assets drives all other processes• By themselves, the impact of automation of cassette

labeling following by slide label generation are profound

20

This case is submitted in 3 specimen containers consisting of:part A - sigmoid colon biopsy, part B - transverse colon biopsy and part C - stomach biopsy with standing preorder

for Helicobacter pylori immunostain.

Protocol driven information is reflected in the slide labels dictating 2 levels cut for each part.

The stomach biopsy protocol, part C, calls for an additional 2 blanks slides to be cut, one for the immunostain & a 4th left unstained.

1

4

3

2

Barcode SpecifiedWork Processes

Outcome• With 'real-time labeling' the batch slide label printing

process has now been entirely eliminated

• Specimen misidentification rates have been reduced

• Workflow efficacy in the histology lab has increased as cassette reading defects have been eliminated– Barcode reading defects required the histotechnologist to

manually type in cases numbers, leading to increased risk of patient misidentification

Results: Misidentification Rates

Baseline Linear BarCodes(Jan.2007)

2D BarCodes(June2012)

45

18

10

5

10

15

20

25

30

35

40

45

Number Mis-IDDefectsPercent of Cases

1.67%

0.62%

0.02%

Chart1

BaselineBaseline

Linear Bar Codes (Jan. 2007)Linear Bar Codes (Jan. 2007)

2D Bar Codes (June 2012)2D Bar Codes (June 2012)

Number Mis-ID Defects

Percent of Cases

45

1.67

18

0.62

1

0.02

Sheet1

Number Mis-ID DefectsPercent of Cases

Baseline451.67

Linear Bar Codes (Jan. 2007)180.62

2D Bar Codes (June 2012)10.02

To resize chart data range, drag lower right corner of range.

• Allows us to record the location and status of specimen assets (parts, blocks, slides, etc.) as they are processed and move through the pathology laboratory

• All components of a case can be tracked from when they are accessioned on through to storage or eventual disposal

• Tracking allows for us to locate case assets and identify who has handled them and where assets have been (i.e. history)

• I will not address routing and processing protocols as it applies to conveyance systems which we are not yet using

Examples of AP Digital WorkflowSpecimen Tracking and Routing

AP-LIS: CoPath Plus v6.3 (Sunquest Information Systems, Tuscon, AZ)

CoPath Specimen Management Routing and Tracking (SMART) module

InfoMaker reports were created using PowerBuilder software

(Sybase, Dublin, CA).

Specimen TrackingTechnology

• We defined specimen points of tracking (SPOTs) in the AP-LIS dictionary Each SPOT is linked to a specific workstation as defined on the

health system’s internal network

• As each asset is scanned and processed – location, scan time, status, and associated user data are

automatically recorded

• Standard tracking tools allowed us to monitor assets in real-time

DesignSpecimen Tracking

AP Workflow Design*

*

** *

*

*

* * *

*

*

*

*

PresenterPresentation NotesWorkflow from start to finish (Accessioning down to the Pathologists). Each red asterisk represents a SPOT in the workflow.

Design: SPOT Implementation

• Currently 145 SPOTs have been defined

Histology Lab

Pathologists

PresenterPresentation NotesArrows = movement of specimens from the grossing room through the histology lab.Can be very granular in your tracking capability.

Tracking Activities Tracking Reports

Asset Manager Asset Location Report

Specimen Discard Parts/Blocks/Slides at a SPOT

Specimen Tracking Update Parts/Blocks/Slides Aging Report

Slide Status Update Scan History Log

View/Update Specimen Tracking Specimen Tracking Scan Errors

Asset Reconciliation SPOT Turnaround Time Report

Unique ID Report

• Some functions are workflow specific while others can be utilized at multiple points in the process

SPOT Turnaround Time Report

• Displays the time between when assets were tracked at one SPOT (i.e. Accessioning) to when they were tracked at the next SPOT (i.e. Grossing)

• This can be used to find out how long it takes assets to be processed from one workstation to another

Workload Reports• Scan events can be

tracked at individual SPOTS and users

• Allows us to provide workload and productivity data for lab managers Example: Slide

counts by microtomyperson

Asset Reconciliation• Allows the user to reconcile the assets in hand with the

assets assigned to a case• Can be use iteratively

– Gross– Accession– Embedding– Foldering

• Perhaps the most commonly automated process in the current AP lab– Including automated cover slipping

• This saves hundreds of man hours per year• Has more consistent results versus manual staining and

cover slipping• Digital interfaces with AP LIS will further enhance

productivity and decrease errors

Examples of AP Digital WorkflowAutomated Staining Platforms

Automated ImmunostainingInterface Design

• An HL7 interface was created between CoPath and the Dako autostainer Link 48 platforms– Allowed IHC orders placed in CoPath to be directly transmitted

and received by the DakoLink instrument control software; used Dako labels

• Our CoPath LIS was upgraded to version 6.0 which provided the capability to uniquely identify and track each case assets – Assign unique identifiers to each and every case asset (i.e.

parts, blocks, and slides)– With this in place, unique slide IDs (linked to IHC orders in the

AP-LIS) were transmitted to Dako autostainer control software– Dako autostainer instruments could then read and utilize

native CoPath labels

New workflow after deployment of the automated stainer interface

Stains Ordered

Slide are cut & labeled

Creates HL7 message

Slides placed onto instruments

Slide labels scanned

Slides processed

CoPath Interface

IHC

Special stains

CoPath – Dako New Workflow1. Stains are ordered in CoPathPlus.2. Stain orders are released to the interface on demand or

scheduled.3. Interface creates an HL7 message and delivers to Dako 4. Message is routed to instrument. Slide labels print from CoPath. 5. Slides are cut in the laboratory and Slide labels print from

CoPath and labels are applied to slides.6. Slides are placed into the instrument. Label is scanned by

instrument which indicates the reagent stain workup.7. Slides are processed.

From CoPath Printer to Slides to DakoBypassing Dako Relabeling

Results• With elimination of relabeling the slides and dual order

entry through automation markedly decreases assay run time – This saves upward of ~700 hours of manual effort per

year while eliminating errors, improving patient safety and improving laboratory throughput

• Increases order accuracy by reducing keystroke errors.• Enhances operational efficiency by automating processes.• Enforces safe, consistent, efficient handling of specimen.

Examples of AP Digital WorkflowDigital Pathology

• Not just whole slide imaging! (WSI)• Distributed microscopic images• Distributed EM• Gross images• Scanned documents• Image analysis• Clinical lab images: gels, plates, hematology analyzers• Cytogenetics analysis• Digital Pathology is:

– digitally capturing, storing, moving, analyzing, interpreting tissues (and other assets) submitted to the laboratory for the purpose of diagnosis, clinical communication, documentation and quality assurance

Apollo ARCC Implementation• Apollo ARCC Image management provides:

– Expandable EMC SAN based storage– Thin-client, web services based deployment– Management for all existing image-generating devices – Disk and network monitoring– Security management for users

• Permission-based logon, very nuanced• Thick-client and Thin-client access

– LIS integration • Interfaces for ADT and images• Images to be send to the EMR via AP-LIS (CoPath)

– Sophisticated device integration including interfaces and file mover services

TECHNOLOGYFile mover services leveraged

Autopsy suite

Apollo ARCC

Specimen X-rays

Grossing stations

DocumentScanners

Autopsy suite

Robotic Telepathology

LIS Integration

Final Workflow• Cases accessioned into CoPath

– ADT sent to Apollo• Images acquired for case automatically stored in Apollo• Images “links” are sent back to CoPath (or not) as

“results”– Those returned are configurable:

• Associated with the case• Embed into the report

• Images viewable in Apollo or CoPath• Report send to EMR with or without images as desired

Examples of AP Digital WorkflowWSI: Drivers for HFHS

• Maturity of the WSI technology platforms – Improved performance– Decreasing costs

• The need for a distributed digital pathology workflow to support diagnostic studies– Move images not glass!

• Loss of support for our robotic telepathology platforms• The desire to adopt cISH and image analysis for diagnostic

panels

Decisions and Goals• Stand up technology to replace robotic telepathology

– Selected Mikroscan as a vendor• Stood up four D2 systems in each hospital to allow for

telepathology support of intraoperative consultation• Select WSI partner for cISH analysis and distributed diagnostic

imaging– Selected Ventana Roche as a partner

• Implement iScan Coreo and Coreo HT platforms• Interface these platforms with Sunquest CoPath• Validate the Ventana Ultra cISH platform• Harmonize bar code labeling symbologies so that CoPath

generate bar codes labels would be used• Integrate this technology into CoPath and Apollo workflow

TECHNOLOGYFile mover services leveraged

Autopsy suite

Apollo ARCC

Specimen X-rays

Grossing stations

DocumentScanners

Autopsy suite

Whole Slide Imaging

LIS Integration

Prerequisites for Interfacing WSI• Requisite hardware and software for WSI capable of

using HL7 messaging– Network attached storage solution– Network bandwidth

• We had WSI in place since January 2016• An LIS capable of communicating with WSI systems via

HL7• Electronic histology orders used for all histology

processes• Bar code labeled assets with unique ID’s

Value of Interfacing WSI• Immediate case access by pathologist

– No annotation• Significantly more data is populated

– Part, Block, stain– Patients details: age, sex, DOB; – Assigned pathologist

• Leveraging bar code technology• Case available to pathologist immediately upon

successful scan– No need to annotate and release

iScan HT Digital Workflow• Cases are created in CoPath at accession• HL7 message is sent to Ventana Connect and onto

Virtuoso• Slides cut, stained and placed on HT for scanning• Scanned images automatically associated with patient,

CoPath accession and pathologist• Scanned images viewable through CoPath interface• Select fields can be exported and integrated into the

CoPath report using Apollo EPMM

LIS Workflow Integration Examples

Launching WSI from your LIS

LIS Integration• All gross images and images captured from cameras mounted on

microscopes go into Apollo ARCC and can be sent directly via HL7 interfaces to pathology reports, then onto the EMR embedded in PDF’s– The include Scanned documents, EM, and X-ray (Faxitron) images

• WSI is a completely different imaging stream– WSI for a pathology cases can be launched from

• CoPath LIS• WSI image viewer (Roche Virtuoso)• Apollo Enterprise Viewer (soon to come)

• The LIS is the integrator– All images can be opened from CoPath on a case by case basis

• This is the basis of the true next generation pathology workflow– Examples….

Challenges and Points of Failure

• Pathologist workstation– Multi monitor solutions, form factor obsolescence– Desktop computing upgrades and rollover’s– Real estate in offices

• Bar code production stability– New devices– New printers– Bar quality monitors

Overall Challenges

• Form factors constantly changing and varying– Slides, coverslips, racks, trays, labeling, bar codes

• Slide Handling variations• Loading process impacts run times

– Continuous versus batch• Overall system durability

– Preventive maintenance– Downtimes, redundancy– System monitors

Overall Challenges

• Security and configuration– Multi user profile versus single account

• Interoperability and proprietary file formatting • Interfaces: Instruments, LIS, HER

– Maintenance, cost• Automation

– How to must effectively connect systems?– Conveyance– Storage of assetts

Overall Challenges

• Workflow– Launching of images

• Window targeting• Inside or outside of application

– Work list• Priority• Status: quality etc• Notifications

– Security hand off between systems• LIS versus image management system

– Image integration into pathology reports

Gaps in LIS-WSI Integration

• Image management– Storage– Retention– Purging– Updates to the cases: insertion of new images,

deletions of old– Case Corrections– Rescan requests– Acceptance criteria

Gaps in LIS-WSI Integration

• Scanning process– Thumbnailing vs. not– Identification of errors– Error management communication– Run management– Continuous flow– Impact of fatal errors

Gaps in LIS-WSI IntegrationInstrument Controller vs. LIS

Summary

Realizing Effective Digital-Enabled Anatomic Pathology Workflow

SummaryRealizing Effective Digital-Enabled Anatomic Pathology Workflow

• Many aspects of AP can be digitally integrated– More integration is available than meets the eye– We don’t count some of these that have a big impact

• Digital integration saves time, money and improves patient safety

• Very few gaps in robotic solutions remain– Gaps will close– Pricing will come down

• Build upon your foundations: sequence carefully!

Questions?J. Mark Tuthill, MD

Division of Pathology InformaticsHenry Ford Hospital

Detroit, MI 48202mtuthil1@hfhs.org

Digital Pathology and AI WorkshopPittsburgh, PA

December 13-14, 2019

Pursuit of an Optimized Surgical Pathology Workflow with Digital Pathology IntegrationObjectivesPhilosophical HypothesisOperational SystemsSlide Number 5Slide Number 6Clinical Pathology Systems��Functional AreasApplicationsAnatomic Pathology Systems��Functional AreasApplicationsOther LIS Applications and Operations FunctionsRoad Map for Digital Workflow�The Big PicturePrerequisites for Anatomic Pathology Digital WorkflowPrerequisites for Anatomic Pathology Digital WorkflowPrerequisites for Anatomic Pathology Digital WorkflowPrerequisites for Anatomic Pathology Digital WorkflowExamples of Anatomic Pathology Digital WorkflowExamples of AP Digital WorkflowExamples of AP Digital WorkflowExamples of AP Digital Workflow�Histology ProtocolsExamples of AP Digital Workflow�Real Time LabelingSlide Number 20OutcomeResults: Misidentification RatesExamples of AP Digital Workflow� Specimen Tracking and RoutingSpecimen Tracking�TechnologyDesignAP Workflow DesignDesign: SPOT ImplementationSlide Number 28SPOT Turnaround Time ReportWorkload ReportsAsset ReconciliationSlide Number 32Slide Number 33Slide Number 34Examples of AP Digital Workflow�Automated Staining PlatformsAutomated Immunostaining�Interface DesignNew workflow after deployment of the automated stainer interface CoPath – Dako New WorkflowFrom CoPath Printer to Slides to Dako�Bypassing Dako RelabelingResultsExamples of AP Digital Workflow� Digital PathologyApollo ARCC ImplementationSlide Number 43TECHNOLOGYFinal WorkflowExamples of AP Digital Workflow� WSI: Drivers for HFHSDecisions and GoalsTECHNOLOGYPrerequisites for Interfacing WSIValue of Interfacing WSIiScan HT Digital WorkflowLIS Workflow Integration ExamplesLIS IntegrationSlide Number 54Slide Number 55Slide Number 56Slide Number 57Slide Number 58Slide Number 59Slide Number 60Slide Number 61Slide Number 62Slide Number 63Slide Number 64Slide Number 65Challenges and Points of FailureOverall ChallengesOverall ChallengesOverall ChallengesGaps in LIS-WSI IntegrationGaps in LIS-WSI IntegrationGaps in LIS-WSI Integration�Instrument Controller vs. LISSummarySummary �Realizing Effective Digital-Enabled Anatomic Pathology WorkflowQuestions?