Digital Imaging Behind the Scenes Stacey Yates
Dec 17, 2015
Digital ImagingBehind the Scenes
Stacey Yates
•Implemented Workflow
•Challenges/Solutions
•Added Value
•Future Plans
Overview
Telepathology
Trestle robotic microscope installed for across-campus frozen section consults
Problems:▫Too Cumbersome for everyday use▫Scanning too slow for real-time review
Approximately 1 minute per slide▫Scan not retained from slide to slide
Where are we coming from?
Needs:▫Capability to view multiple slides at once
from anywhere▫Quickly and easily find images to be used
for teaching and research▫Ability to reference images from past cases
without waiting for requested slides.▫Enhance the learning experience for
trainees.
TelepathologyWhere are we coming from?
Starting the Process
• Digital Imaging Committee▫ Pathologists, Residents and
IT Staff
• Evaluated multiple scanners and database management packages
• Important Requirements:▫ Simple Database Management▫ Ease of Use▫ Speed and Quality of images
Equipment/
Infrastructure
Aperio ScanScope
• ScanScope XT: 120 slide capacity.
• All clinical scanning• 20x and 40x Magnification
BioImagene
• iScan
• 20X
• 160 slide capacity
• Currently research only
• Not linked to clinical image database
Storage
• Dell MD3000 PowerVault▫15 - 300gb Harddrives
• Dell MD1000 PowerVault▫15 – 300gb HarddrivesTotal Clinical Image Storage = 7.5 Terabyte
• Dell PowerEdge 2970▫Runs server software and services
• Dell PowerVault TL2000▫Backup Tape Library
Network
gigabyte
100 mb
100 m
b
100 mb
100 mb
3 blocks apart
Lets Get Started….
•Scanned only research and conference slides
•Became comfortable with the scanner and database
•Began instructing users on how to use Spectrum
•Small group of “Power Users” utilized Spectrum for conferences
February 2008
Oops!
• Problems found:▫Scanning process long and tedious
▫Manually entering patient/case information into
Image Database from LIS
▫Equipment Downtime and Storage Space
▫Learning curve for finding images Residents embraced the technology Pathologists did not want to use two separate
systems
▫ Image interface not good for educational purposes.
•The Plan:▫Group and individual training sessions
•The Result:▫Confusion▫Lack of interest
•Solution: ▫Viewing images directly from LIS
Learning CurveYou want me to do what?
Scanning ProcessMarriage of the LIS and Image Database
Before the marriage could take place Barcoding had to be implemented.
Barcode contains:Inside Accession #Block #Slide #
1. Histology is entered to create a label with barcode for each slide.
2. The barcode is temporarily attached to the slide.
Scanning ProcessMarriage of the LIS and Image Database
3. Each slide must be cleaned and free from dust and fingerprints
4. Once the slides are loaded a snapshot is taken of each of the slides.
Area of Interes
t
Scanning ProcessMarriage of the LIS and Image Database
5. Once the image is scanned an HL7 message is sent to our LIS containing the link to the image on the server and the barcode id.
6. The LIS then sends a message back to the Image database with identifying information:
▫ Medical Record Number▫ Patient Name▫ Birth date▫ Hospital▫ Part Type▫ Accession Number▫ Stain
Scanning ProcessMarriage of the LIS and Image Database
7. Image Database uses the information to file the slide under a new patient or attaches it to an already existing patient.
▫ The MRN number, patient name, date of birth and hospital must match.
8. Each slide is reviewed for quality, the label is removed and sent back for filing or returned to referring institution.
9. Verify each slide is linked to LIS and the information sent to Spectrum is correct.
Scanning ProcessMarriage of the LIS and Image
Database
Slides Received
Case Signed Out
Diagnostic Slides Selected
Barcode Printed
Slides Scanned
Slides Returned to Referring Institution
Slides Reviewed
Scanning ProcessThe Workflow
Timeline: Start to Finish
• 20X ▫ Label Slide with barcode = 1 minute▫ Prep slide for scanning = 1 minute▫ Scan slide = 6 minutes
Total = 8 minutes per slide• 40X
▫ Label Slide with barcode = 1 minute▫ Prep slide for scanning = 1 minute▫ Scan slide = 33 minutes
Total = 35 minutes per slide
• Most scans are between 500,000 mb and 1 gb• Turn around time usually 48 hours
Scanning Process
▫LIS integration increased productivity and took away some of the manual data entry
▫Ensured slides where filed with the correct patient information
▫Pathologists only used one program to access slides
▫Over 12,000 digital images linked to LIS
Problems Solved by LIS Integration
Learning CurveViewing Images through LIS
Learning CurveViewing Images through LIS
Equipment Downtime
•Over the past 2 years▫5 days of complete scanner downtime▫6 days of manual scanning due to scanner
hardware malfunction▫2 -3 days to catch up after downtime
•Back-up Plan?▫Bring 2nd Scanner on line
Image Storage
•April 2010▫Original 7.5 tb full▫18,078 images
•Today▫Added 25 tb ▫Currently 23,663 images
•Will we need more storage?
Education
▫Faculty can create sets of digital images▫Used as teaching tool for residents and
fellows
Study Sets
▫Quizzes are created to evaluate progress
EducationQuiz
▫Weekly case is presented▫Trainee’s get a chance to review▫Diagnosis is posted at the end of the week
EducationCase of the Week
Added Value
•Availability: immediate access to images for increased efficiency in patient care, education and research
•Permanent: no lost or broken slides•Portable: send images instead of slides for
consults, legal requests•Conferences: Ability to take images to
conferences instead of actual slides•Teaching: Create teaching sets and
quizzes using digital images
Future Challenges
•Viewing images remotely through Citrix▫Citrix will not work for viewing images
through LIS▫Citrix will allow the searching of the
Spectrum database and opening of images through Spectrum using ImageScope but is extremely slow
•Continue to build teaching sets to aid in the education of residents and students.▫Evaluate software to manage the educational
and clinical aspect of digital imaging
Future Challenges
•Currently requires one full-time position•Continue to build an image repository.•Image Analysis•Ability to search for images based on
diagnosis without opening each case in LIS
•De-identified searches for education and research purposes
Digital ImagingBehind the Scenes
Stacey Yates
Thank you!