APIII - Advancing Practice, Instruction & Innovation Through Informatics

Marriott City Center, Pittsburgh, PA | September 20 - 23, 2009

Presented at the 1999 APIII Conference                        Return to 1999 Abstract Index


EFFECT OF IMAGE COMPRESSION ON TELEPATHOLOGY: A RANDOMIZED, CONTROLLED TRIAL

National Library of Medicine
Bethesda, Maryland
Alvin B. Marcelo, MD

Alvin B. Marcelo, MD1; Paul Fontelo, MD, MPH1; Miguel Farolan, MD2; Hernani Cualing, MD3

1National Library of Medicine
2University of Illinois in Chicago
3University of Cincinnati

Telepathology is one of the fastest growing branches of telemedicine. This is mostly attributable to recent developments in technology that have made it easier to exchange compressed images over the Internet. A raw telepathology image is usually large and in order to save time, it is compressed prior to transmission. The Joint Photographic Expert Group (JPEG/JPG) algorithm, a lossy compression technique, is a very popular method for decreasing image size. Yet concerns about the loss of data with JPG compression and its effect on image quality and diagnosis have never been addressed. This study aims to gather evidence on this matter.

A randomized controlled design (double-blind) was used to determine the effect of JPEG compression on telepathology diagnosis. One hundred test cases were digitized into non-compressed (50 BMP) and compressed (50 JPG) formats and were randomly assigned to ten pathologists through a web interface. Each pathologist assessed five compressed and five non-compressed cases. File name extensions were masked server-side and cached client-side to minimize the bias of format identification through name extension and transmission time. Each respondent accomplished an accompanying questionnaire. Parameters measured were:

  1. acceptability of a case for diagnosis,
  2. accuracy of diagnosis,
  3. confidence in diagnosis, and
  4. over-all image quality.

Results show there are no statistically significant differences between non-compressed and compressed images in terms of acceptability for diagnosis, accuracy of diagnosis, confidence in diagnosis, and over-all image quality.

This randomized controlled trial provides a basis for the use of JPEG compression for formal telepathology diagnosis.

Search