Joshua Jay Reicher, Abstract Co-Author: Medical Advisory Board, Health Butler, LLC Investor, Health Butler, LLC

Matthew Borzage, Abstract Co-Author: Nothing to Disclose

Murray A. Reicher MD, Abstract Co-Author: Chairman, DR Systems, Inc Owner, Imaging Healthcare Specialists, Inc Shareholder, Health Beacons, Inc

Mark Daniel Kovacs MD, Presenter: Nothing to Disclose

Taylor Randall Jordan MD, Abstract Co-Author: Nothing to Disclose

Michael A. Trambert MD, Abstract Co-Author: Medical Advisor, DR Systems, Inc

For full-field digital mammography (FFDM), lossy compression is currently permitted by MQSA only for secondary viewing, if acceptable to the viewing clinician. Many benefits, including reduced storage costs, faster telemammography, and greater access to expert mammographer consultants, would result if a standard could be developed safely enabling significant lossy data compression effective for mammographic images created by a wide range of vendors. In order to develop such a lossy data compression standard, we qualitatively and quantitatively evaluated various compression techniques using digitized and digital exams and mammography phantoms.

Eighty breast and sixteen phantom images from five separate vendors were compressed with thirty different JPEG and JPEG2000 algorithms. Trends in the JPEG quality (Q) factor and the measured image compression ratios were analyzed by vendor. Peak signal-to-noise ratios (PSNR) at JPEG2000 80:1 compression were compared between vendors. In addition, three radiologists reviewed images on FDA-cleared five megapixel monitors to determine whether or not JPEG2000 80:1 compressed images were visually differentiable from lossless compressed images. 

In a two-alternative forced choice test, radiologists were found to be unable to differentiate JPEG2000 80:1 compressed from lossless compressed full-field digital mammography (FFDM) images. In separate quantitative image analyses, significant variability was observed in JPEG Q factor compression ratios and PSNRs by vendor.

Despite varying effects of compression on images from different vendors, uncompressed FFDM images from all tested vendors could be compressed with JPEG2000 80:1 without visually detectable degradation. This evidence suggests that a definitive study to establish a standard for lossy image compression in digital mammography is warranted and that JPEG2000 80:1 compression for FFDM is a promising standard compression scheme.

A new standard enabling significant lossy data compression for FFDM could result in many benefits, including reduced storage costs, faster telemammography, and greater access to expert mammographers.

Reicher, J, Borzage, M, Reicher, M, Kovacs, M, Jordan, T, Trambert, M, Compression Tolerance of Full-Field Digital Mammography Images: Pilot Findings with Multiple Vendors and Call for a Definitive Study to Establish a Standard.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034226.html