Digital Microscopy: New Paradigm's for Teaching Microscopic Anatomy and Pathology

Transcription

1 Digital Microscopy: New Paradigm's for Teaching Microscopic Anatomy and Pathology Michael Feldman, MD, PhD Assistant Dean IT Assistant Professor Pathology University of Pennsylvania Health System

2 Agenda Introduction (Feldman) Aperio (TBA) Microbrightfield (Jack Glaser) Break Aurora (John Stinson/Kemp Watson) Zoomify (Dave Urbanic) Panel Discussion

3 Objectives Define Digital Microscopy Hardware solutions for creating slides Software environment s for: File formats and storage Server options Viewer options Annotation tools Educational Pedagogy how to best use these tools?

4 Digital Imaging in Education How many use Digital images in Pathology or Histology Education? How are the images used? Have you been satisfied with the technology? Limitations? How does it fit your educational pedagogy for teaching histology/pathology?

5 Penn Answer s How are the images used? Static - Powerpoint lectures, web based small group teaching Have you been satisfied with the technology? Resolution of images at low magnification Variability of images from different sources No interactivity Limitations? Time to take all these images Static Pedagogy Static Passive Limited self exploration

6 Static Microscopic Imaging in Pathology CCD chip 6 um pixel Filter for RGB Histology Slide

7 2X 4X 10X 40X

8 Digital Slide Entire slide or large area of slide imaged at sub micron resolution ( um/pixel) to simulate the action of a microscope zoom from 1X 40X and slide movement XY directions Resulting images are HUGE! K pixel on edge for a cm 2 piece of tissue Uncompressed 5-10 Gb, compressed Mb Can t download these babies using standard Html!

9 Digital slide process Teaching Scan slide Aperio Microbrightfield Bacus Media Cybernetics Annotate Image Microbrightfield Aurora Zoomify Store Image Flashpix JPEG2000 Zoomify Mscope Image Server Iseemedia Aperio Zoomify Aurora Mscope Microbrightfield Display Image Java applet Active X Executable Html Quicktime

10 How are Slides captured? Tiling Manually take pictures and assemble them with software Panovue Image Assembler ($79, XY tiling, small slides only 5-7K/side) Mac Quicktime VR Robotic tiling Microbrightfield ( Bacus ( Striping Aperio Scanscope (

11 Manual Tiling Images captured with static digital camera and manual stage Works for small areas of a slide (up to 50 images for a 7x7 field of capture) Stitched with Panovue ImageAssembler Demonstration of small slides Images served using ZoomServer from Iseemedia

12 Digital Slide Stitching - A 1 cm 2 piece of tissue imaged at 40X (0.25 mm FOV at 40x) - Creates a 40x40 grid = 1600 Images (25k x 25k pixels) - Total Image Tiled together is 1.6 Gb uncompressed - With compression Mb

13 Striping Aperio - Instead of using individual images, scanner works like fax and scans a stripe (width of scanning element) at very high resolution - Linear array is 54,000 pixels/inch using 20X lens (0.4 um/pixel) - Using 40X, array achieves 108,000 pixels/inch (0.2 um/pixel) - Very fast scan times (10-15 minutes with 40X lens for a 1.5 cm portion tissue)

14 Image storage/format Flashpix pyramidal format JPEG Aperio Mscope Aurora Zoomify - pyramidal

15 Pyramidal format Flashpix Zoomify Thumbnail High resolution Webpage

16 Pixel on Demand Do not retrieve entire image, only data need to view image in a given window at a given magnification Return image as tiles (you can see the blocks come in as they are retrieved by the server) Makes for efficient viewing of HUGE files over reasonable bandwidth

17 Aperio and Aurora JPEG2000 Aperio and Mscope Aurora formats More compression than jpeg (wavelet) Rather than viewing images from a selected subfile, wavelet reconstructs the image at the desired resolution Image is not stored as pyramidal format. More compression without as much loss of image quality than jpeg

18 How are slides Served (Accessed) Central server houses and serves images (Image server) Images are retrieved from server by client using pixel on demand as image tiles: Servers vary How send data (tile size) How data is cached (server, client or both) How much CPU is used, how much disk is used How many concurrent processes are used by server Combine to define how many tiles of data can be pulled off disks and served out to concurrent clients

19 Image Server (du Jour ) So many flavors: Zoomserver ( - flashpix Zoomify ( Microbrightfield Zoomserver or zoomify Aurora Mscope proprietary, flashpix & jpeg2000 in near future Aperio jpeg2000 Bacus proprietary

20 Client side viewing What software is used to pull images from server Java Applet from web page (Aurora, Microbright, Bacus) ActiveX (Zoomify, Bacus) Stand alone application (Aperio) HTML (Aperio) Quicktime (Zoomify) Flash (Zoomify)

21 Advantages of Applet Whether Flash, Java, Quicktime Cross platform viewing Mac and PC Easier to support across enterprise for software updates Portable to devices including (PDA s) Lightweight (<100 Kb)

22 Annotation Serving up standardized microscope quality high resolution images is start Uniformity No bad slides, cracked, not representative No slide handling and management Slides always available Even better if we can author content on slides and with the slides into educational lessons

23 Annotation Cont d Ability to annotate image can be used in multiple manners Simply label slide simple answer Prior to lesson - Author material with questions and no answer s After lesson provide same images with answers and pointers Allow students to label images with questions independent of faculty annotations Questions Interactive like a threaded discussion that embeds digital slides Which fit s your pedagogy?

24 How do you teach with glass vs. digital? Glass: Search and identify with you assisting and facilitating in a small group Digital: Same Digital: Compare and contrast multiple slides at once Digital: Before and after scenario Digital: support s asynchronous learning, threaded discussion Student to student Student to faculty Digital: expert learning system with intelligent tutor (Rebecca Crowley, University Pittsburgh) Digital: distributed multiheaded microscope Digital: share content

25 Virtual Slide Examples Iowa Slides Penn Med Comparative slides Penn Penn MBR sight Aperio web Microbrightfield Aurora Mscope

26 Usage Data from Penn Pilot Usage stats from one course section (4 labs): 1.6 million hits (each time move slide registers as hit Only 12 manually tiled slides used Avg 22 minutes per visit More 50% viewed outside campus 150 students viewed slides from > 200 computers Pushed > 2 Gb data during these 4 small group sessions

27 Feedback from Pilot Uniformity of image Always there to study Supported group study Image quality excellent Student s liked it, more comfortable than search and find on microscope Wanted more of it

28 Future direction Integrate multi-role annotation tool to support different instances of annotation of the same image faculty, student Integrate threaded discussion Use interactivity to make education sessions dynamic (educational traces vs expert trace) Create a distributed multi-headed client-server for real time interaction Awaiting delivery of Aperio scanner Create archives for UME and GME (planned scope thousands of images over 3-5 year period)