Presented at the 1999 APIII Conference                        Return to 1999 Abstract Index


WEB-BASED "VIRTUAL MICROSCOPY" FOR CYTOPATHOLOGY EDUCATION

Carnegie Mellon University
Pittsburgh, Pennsylvania
Adam Landman

Cedars Sinai Medical Center
Los Angeles, California
Alberto M. Marchevsky, MD

UPMC Health System Department of Pathology
Pittsburgh, Pennsylvania
Yukako Yagi

Yukako Yagi1, Adam Landman2, Alberto M. Marchevsky, MD3
1University of Pittsburgh, Department of Pathology
2Carnegie-Mellon University, H. John Heinz III School of Public Policy and Management
3Cedars Sinai Medical Center, Department of Pathology and Laboratory Medicine


We have developed a Web-Based Virtual Microscopy system that allows the user to examine a cytology "virtual slide" in a manner that closely simulates light microscopy. The image can be scrolled and selected areas can be zoomed, allowing for scanning of the "slide" at various magnifications. The user can view the images using a standard web browser (Internet Explorer 4.0 or higher). Currently the technology does not allow for changing depth of focus.

This system uses pathology images in a FlashPix format. This format is a multi-resolution image format in which the image is stored as a series of independent arrays, each representing the image at a different spatial resolution. This allows the image to be displayed on different output devices at different resolutions with minimal resizing of the image. With an intelligent software layer on top of the format, this multi-resolution ability is achieved transparently to the user.

Our system currently includes twenty "virtual slides" from Pap smears. Twelve images were photographed from each case at 100x optical magnification, at 3200 x 2400 pixel, 24 bit resolution. They represent approximately 1 x 1 cm of the slide area.

The photos were arranged in an image tile. The user can examine the entire image at low magnification and enlarge digitally each area of interest, up to 800x.

The software also provides nagivation tools to move the image in all directions. A questionnaire that allows the user to select a diagnosis is provided next to the image. The user gets instantaneous feedback from the system, providing the correct answers and selected comments. The system is now available in the Internet and has been tested with several users who reported excellent satisfaction with the computer interface.