Tuesday, 29 September 2009

CAD and Screening Mammography: ready for prime time?

In this months RADIOLOGY, Robyn Birdwell and Liane Philpotts share their different viewpoints on CAD and screening mammography. Both are well known mammographers in the North East, with Robyn being a guest of the RCR Breast Group at November's Annual Scientific Meeting

In the first of the two articles, Robyn Birdwell talks about the various CAD studies, and how the proponderance of data supports the use of CAD in screening mammography, and that "Having a system to aid the human eye that does nottake vacations, is not vulnerable to fatigue or environmental distractions, is without, emotion, and is designed specifically to assist the very human eye to "look over here" seems like a good idea"
The Preponderance of Evidence Supports Computer-aided Detection for Screening Mammography
Robyn L. Birdwell
Radiology 2009;253 9-16

In the second editorial article, Liane Philpotts points out that recalling patients from screening is sometimes more of an art than a science, and I am sure that many experienced radiologists would share her view.

One of the main issues with CAD is the necessary high false-positive rate of CAD prompts which subsequently means that the specificity is low, and we have the distracting factor of many false calls, while also knowing that not all cancers are picked up by CAD

Understanding of the limitations of computer-aided detection is important for those interpreting mammograms; this cautious approach to the use of computer-aided detection should help optimize this presently imperfect system and minimize the possible detrimental effects

Can Computer-aided Detection Be Detrimental to Mammographic Interpretation?
Liane E. Philpotts
Radiology 2009;253 17-22