2006 Scientific Session Abstracts
Scientific Session Presentations & Schedule
E-Poster Presentations
Submission Compliance and Guidelines
Copyright Transfer Form (PDF)
Abstract Archive
A Rapid H&E Digital Microscope for Viewing, Archiving, and Analysis of Tissue Sections
Robert Filkins, PhD, (filkins@research.ge.com) Siavash Yazdanfar, PhD, Kevin Kenny, PhD; Eugene Barash, PhD, Mark Meyers, MS; Elizabeth Dixon, MS, Michael Gerdes, PhD, Max Seel, BS, Anup Sood, PhD, Evi Loghin PhD, Michael Rishel, PhD, Harvey Cline, PhD, Ali Can, PhD, Fiona Ginty, PhD, Michael Montalto, PhD, Molecular Imaging and Diagnostics Pathology Program, GE Global Research, Niskayuna, NY
Context: The growing trend toward digital imaging in the pathology laboratory is hampered by the comparatively long time required to digitize a sample with the necessary optical resolution. Making a significant advancement in digital imaging performance requires innovation along the entire imaging pipeline including: lens design, digital capture, auto-focusing, image processing, stitching and compression, and finally storage and retrieval. In many regards, it is the image latency due to retrieval and viewing that has the most detrimental effect on the performance of digital imaging systems.
Design: We have created a complete pathology specimen digital imaging system, optimized for the acquisition of bright field images, e.g. H&E and conventional IHC stained tissue sections. The system includes a large field of view microscope with 0.5 micron resolution, rapid scanning opto-mechanics and high speed data acquisition system. The subsequent image-processing pipeline acquires images, corrects for lens distortions, color balances, stitches contiguous frames into a macro-image and performs a user-defined level of compression. The images are stored directly on a large RAID system and are compatible with DICOM image environments.
Results: Our poster presentation includes sample images from colon, prostate and breast cancer sections acquired using the GE proprietary digital imaging systems. Each acquisition subsystem is analyzed in order to provide an overall digital pipeline capable of acquiring whole slide sections with 0.5 micron resolution at a rate much less than one minute, from photons to disk storage.
A novel image viewing environment is demonstrated for the retrieval and on-screen reading of the ‘digital sections.’ The viewer utilizes JPEG2000 compression and custom rendering algorithms to provide seemless pan and zoom features.
Conclusions: We present a rapid digital scanning microscope with an order of magnitude improvement in data rate over typical automated microscopes. The complete system will potentially accelerate acceptance of digital imaging within the pathology laboratory.
